/CT/CochraneCMS/TexRendering5/4666669.dvi

Transcript

1 Cochrane Database of Systematic Reviews Decision aids for people facing health treatment or screening decisions (Review) Stacey D, Légaré F, Lewis K, Barry MJ, Bennett CL, Eden KB, Holm es-Rovner M, Llewellyn-Thomas H, Lyddiatt A, Thomson R, Trevena L Stacey D, Légaré F, Lewis K, Barry MJ, Bennett CL, Eden KB, Holmes- Rovner M, Llewellyn-Thomas H, Lyddiatt A, Thomson R, Trevena L. Decision aids for people facing health treatment or screening decis ions. Cochrane Database of Systematic Reviews 2017, Issue 4. Art. No.: CD001431. DOI: 10.1002/14651858.CD001431.pub5. www.cochranelibrary.com Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by Jo hn Wiley & Sons, Ltd.

2 T A B L E O F C O N T E N T S 1 HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 SUMMARY OF FINDINGS FOR THE MAIN COMPARISON . . . . . . . . . . . . . . BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 8 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 12 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Figure 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 32 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 206 DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 1.1. Comparison 1 Knowledge, Outcome 1 Knowledge - a ll studies. . . . . . . . . . . . . . . 210 Analysis 1.2. Comparison 1 Knowledge, Outcome 2 Knowledge - s ubgroup by timing of intervention (in consultation versus in preparation for consultation). . . . . . . . . . . . . . . . . . 212 . . . . . . . . tudies without high risk of bias. . . . . . . 215 Analysis 1.3. Comparison 1 Knowledge, Outcome 3 Knowledge - s Accurate risk perceptions - all studies. . . . . 217 Analysis 2.1. Comparison 2 Accurate risk perceptions, Outcome 1 Analysis 2.2. Comparison 2 Accurate risk perceptions, Outcome 2 Accurate risk perceptions - subgroup by timing of intervention (in consultation versus in preparation for consu ltation). . . . . . . . . . . . . . . . 218 Analysis 2.3. Comparison 2 Accurate risk perceptions, Outcome 3 Accurate risk perceptions - studies without high risk of 219 bias. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 3.1. Comparison 3 Informed values-choice congruence, O utcome 1 Informed values-choice congruence - all 220 studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 3.2. Comparison 3 Informed values-choice congruence, O utcome 2 Informed values-choice congruence - actual 221 choice only. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 3.3. Comparison 3 Informed values-choice congruence, O utcome 3 Informed values-chose congruence -using 222 MMIC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . utcome 4 Informed values-chose congruence - Analysis 3.4. Comparison 3 Informed values-choice congruence, O . heterogeneous measures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Analysis 3.5. Comparison 3 Informed values-choice congruence, O utcome 5 Informed values-choice congruence - without studies of high risk of bias. . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 . . Analysis 4.1. Comparison 4 Decisional conflict, Outcome 1 Decisio nal conflict - all studies. . . . . . . . . . 225 Analysis 4.2. Comparison 4 Decisional conflict, Outcome 2 Decisio nal conflict - in consultation. . . . . . . . 231 Analysis 4.3. Comparison 4 Decisional conflict, Outcome 3 Decisio nal conflict - in preparation for consultation. . . 233 nal conflict - without studies having high risk of Analysis 4.4. Comparison 4 Decisional conflict, Outcome 4 Decisio 239 bias. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Outcome 1 Participation in decision making - all studies. Analysis 5.1. Comparison 5 Participation in decision making, 245 Analysis 5.2. Comparison 5 Participation in decision making, Outcome 2 Participation in decision making - in consultation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 Analysis 5.3. Comparison 5 Participation in decision making, Outcome 3 Participation in decision making - in preparation for consultation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 Analysis 6.1. Comparison 6 Proportion undecided, Outcome 1 Pr oportion undecided - all studies. . . . . . . . 251 Analysis 7.1. Comparison 7 Satisfaction, Outcome 1 Satisfacti on with the choice - all studies. . . . . . . . . 252 i Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

3 Analysis 7.2. Comparison 7 Satisfaction, Outcome 2 Satisfacti on with the choice - in consultation. . . . . . . 253 on with the choice - in preparation for consultation. . Analysis 7.3. Comparison 7 Satisfaction, Outcome 3 Satisfacti 253 Analysis 7.4. Comparison 7 Satisfaction, Outcome 4 Satisfacti on with the decision making process - all studies. . . 254 Analysis 7.5. Comparison 7 Satisfaction, Outcome 5 Satisfacti 255 on with the decision making process - in consultation. Analysis 7.6. Comparison 7 Satisfaction, Outcome 6 Satisfacti on with the decision making process - in preparation for consultation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 Analysis 8.1. Comparison 8 Choice, Outcome 1 Choice: surgery ov er conservative option. . . . . . . . . . 256 Analysis 8.2. Comparison 8 Choice, Outcome 2 Choice for screenin 259 g. . . . . . . . . . . . . . . . . Analysis 8.3. Comparison 8 Choice, Outcome 3 Choice: diabetes m edication (uptake new medication). . . . . . 261 261 ADDITIONAL TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

4 [Intervention Review] Decision aids for people facing health treatment or screeni ng decisions 7 2 , 3 1 1 5 4 6 , Carol L Bennett , Karen B Eden , Krystina Lewis , Margaret Holmes-Rovner , France Légaré , Dawn Stacey , Michael J Barry 8 9 10 11 , Richard Thomson , Lyndal Trevena Hilary Llewellyn-Thomas , Anne Lyddiatt 1 2 School of Nursing, University of Ottawa, Ottawa, Canada. ch Centre for Practice Changing Research, Ottawa Hospital Resear 3 Institute, Ottawa, Canada. Population Health and Optimal Health Practices Research Axis, CHU de Québec Research Center, 5 4 InformedMedical DecisionsFoundation, Boston, MA,USA. Clinical EpidemiologyProgram, Université Laval,QuébecCity, Canada. 6 Department of Medical Informatics and Clinical Epidemiology, Oregon Health Ottawa Hospital Research Institute, Ottawa, Canada. 7 Sciences University, Portland, Oregon, USA. tate University College Center for Ethics and Humanities in the Life Sciences, Michigan S 8 eisel School of Human Medicine, East Lansing, Michigan, USA. The Dartmouth Center for Health Policy & Clinical Practice, The G 10 9 No affiliation, Ingersoll, Canada. mpshire, USA. Institute of of Medicine at Dartmouth, Dartmouth College, Hanover, New Ha 11 , UK. Health and Society, Newcastle University, Newcastle upon Tyne The University of Sydney, Sydney, Australia . [email protected] Contact address: Dawn Stacey, School of Nursing, University of Ottawa, 451 Smyth Road, Ottawa, ON, Canada. Cochrane Consumers and Communication Group. Editorial group: ublished in Issue 4, 2017. Publication status and date: New search for studies and content updated (conclusions changed), p Stacey D, Légaré F, Lewis K, Barry MJ, Bennett CL, Eden KB, Holm Citation: es-Rovner M, Llewellyn-Thomas H, Lyddiatt A, Thomson R, Trevena L. Decision aids for people facing health tr eatment or screening decisions. Cochrane Database of Systematic Reviews D001431.pub5. 2017, Issue 4. Art. No.: CD001431. DOI: 10.1002/14651858.C hn Wiley & Sons, Ltd. Copyright © 2017 The Cochrane Collaboration. Published by Jo A B S T R A C T Background Decision aids are interventions that support patients by mak ing their decisions explicit, providing information about op tions and een decisions and personal values. associated benefits/harms, and helping clarify congruence betw Objectives To assess the effects of decision aids in people facing treatmen t or screening decisions. Search methods Updated search (2012 to April 2015) in CENTRAL; MEDLINE; Embas e; PsycINFO; and grey literature; includes CINAHL to September 2008. Selection criteria decision aids to usual care and/or alternative interventions . For this We included published randomized controlled trials comparing update, we excluded studies comparing detailed versus simple decision aids. Data collection and analysis Two reviewers independently screened citations for inclusion , extracted data, and assessed risk of bias. Primary outcomes, based on the International Patient Decision Aid Standards (IPDAS), were at tributes related to the choice made and the decision-making proce ss. Secondary outcomes were behavioural, health, and health syst em effects. We pooled results using mean differences (MDs) and risk ratios (RRs), applying a random-effects model. We conducted a subgroup analysis of studies that used the patient decision aid to prep are for the consultation and of those that used it in the consult ation. We used GRADE to assess the strength of the evidence. 1 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

5 Main results We included 105 studies involving 31,043 participants. This u dies pdate added 18 studies and removed 28 previously included stu isk of bias’ assessment, we rated two items (selective reporti ng and blinding comparing detailed versus simple decision aids. During the ’R of participants/personnel) as mostly unclear due to inadequa te reporting. Twelve of 105 studies were at high risk of bias. With regard to the attributes of the choice made, decision aids i ncreased participants’ knowledge (MD 13.27/100; 95% confidence interval (CI) 11.32 to 15.23; 52 studies; N = 13,316; high-qual ity evidence), accuracy of risk perceptions (RR 2.10; 95% CI 1.66 to 2.66; 17 studies; N = 5096; moderate-quality evidence), and cong ruency between informed values and care choices (RR 2.06; 95% CI 1.46 to 2.91; 10 studies; N = 4626; low-quality evidence) compar ed to usual care. Regarding attributes related to the decision-making process a nd compared to usual care, decision aids decreased decisional confl ict − 9.28/100; 95% CI − 12.20 to − 6.36; 27 studies; N = 5707; high-quality evidence), indecision related to feeling uninformed (MD − 8.81/100; 95% CI about personal values (MD 11.99 to − 5.63; 23 studies; N = 5068; high-quality evidence), and the prop ortion − of people who were passive in decision making (RR 0.68; 95% CI 0. 55 to 0.83; 16 studies; N = 3180; moderate-quality evidence). and appeared to have a positive effect on patient-clinician commu Decision aids reduced the proportion of undecided participants ni- cation. Moreover, those exposed to a decision aid were either e qually or more satisfied with their decision, the decision-maki ng process, and/or the preparation for decision making compared to usual ca re. lective invasive surgery in favour of more conservative optio Decision aids also reduced the number of people choosing major e ns (RR 0.86; 95% CI 0.75 to 1.00; 18 studies; N = 3844), but this reducti on reached statistical significance only after removing the stud y on prophylactic mastectomy for breast cancer gene carriers (RR 0.84; 95% CI 0.73 to 0.97; 17 studies; N = 3108). Compared to usual care, decision aids reduced the number of people choosing prosta te-specific antigen screening (RR 0.88; 95% CI 0.80 to 0.98; 10 studies; N = 3996) and increased those choosing to start new med ications for diabetes (RR 1.65; 95% CI 1.06 to 2.56; 4 studies; N = 447). For other testing and screening choices, mostly there were no differences between decision aids and usual care. 2.6 minutes longer (24 versus 21; 7.5% increase). The costs of the The median effect of decision aids on length of consultation was al care in four studies. People receiving decision aids do not ap pear to decision aid group were lower in two studies and similar to usu differ from those receiving usual care in terms of anxiety, gen eral health outcomes, and condition-specific health outcomes. St udies did not report adverse events associated with the use of decision a ids. ed in preparation for the consultation versus during the consu ltation, In subgroup analysis, we compared results for decision aids us ge and accurate risk perception. For other outcomes, we could not co finding similar improvements in pooled analysis for knowled nduct n each subgroup. formal subgroup analyses because there were too few studies i Authors’ conclusions Compared to usual care across a wide variety of decision contexts tter , people exposed to decision aids feel more knowledgeable, be informed, andclearerabouttheirvalues, andtheyprobablyh ave amore active role in decision making and more accurate risk pe rceptions. There is growing evidence that decision aids may improve value s-congruent choices. There are no adverse effects on health outcom es or satisfaction. New for this updated is evidence indicating im proved knowledge and accurate risk perceptions when decision ai ds are rther research is needed on the effects on adherence with the chos en option, used either within or in preparation for the consultation. Fu cost-effectiveness, and use with lower literacy populations. P L A I N L A N G U A G E S U M M A R Y Decision aids to help people who are facing health treatment or screening decisions Review question We reviewed the effects of decision aids on people facing health treatment or screening decisions. In this update, we added 18 n ew studies for a total of 105. Background Making a decision about the best treatment or screening option can be hard. People can use decision aids when there is more than one option and neither is clearly better, or when options have benefits and harms that people value differently. Decision ai ds may be 2 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

6 pamphlets, videos, or web-based tools. They state the decisio n, describe the options, and help people think about the optio ns from a rms). personal view (e.g. how important are possible benefits and ha Study characteristics For research published up to April 2015, there were 105 studie s involving 31,043 people. The decision aids focused on 50 diff erent decisions. The common decisions were about: surgery, screening (e.g. prostate cancer, colon cancer, prenatal), genetic testing, a nd he decision aids were compared to usual care that may have include medication treatments (e.g. diabetes, atrial fibrillation).T d general information or no intervention. In the 105 studies, 89 evalu ated a patient decision aid used by people in preparation for t he visit with the clinician, and 16 evaluated its use during the visit with th e clinician. Key results with quality of the evidence When people use decision aids, they improve their knowledge of the options (high-quality evidence) and feel better informed a nd more clear about what matters most to them (high-quality evidence). Th ey probably have more accurate expectations of benefits and harm s of options (moderate-quality evidence) and probably participa ho te more in decision making (moderate-quality evidence). People w use decision aids may achieve decisions that are consistent with their informed values (evidence is not as strong; more research could change results). People and their clinicians were more likely to talk about the decision when using a decision aid. Decision aids have a variable effect on the option chosen, depending on the choice be ing considered. Decision aids do not worsen health outcomes, an d people using them are not less satisfied. More research is need ed to assess if people continue with the option they chose and al so to assess what impact decision aids have on healthcare systems. 3 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

7 cate better knowledge. - - edge provement in knowl- tically significant im- ies showed a statis- 46 out of 52 stud- Higher scores indi- Comments a,d © a,b a,d,e,f ©© High (GRADE) Moderate Low ⊕⊕⊕ Quality of the evidence ⊕⊕ ⊕⊕⊕⊕ ] Explanation [ (17 studies) (52 studies) (studies) 13,316 No of participants (10 studies) 4626 5096 (1.46 to 2.91) (1.66 to 2.66) (95% CI) - RR 2.10 Relative effect RR 2.06 (447 to (422 to nsidering treatment or screening decisions tion groups was 13.27 Patient decision aid 565 per 1000 716 per 1000) 841 per 1000) 595 per 1000 higher) higher (11.32 to 15.23 The mean knowledge score in the interven- Corresponding benefit c c The mean knowl- Illustrative comparative benefits* (95% CI) 269 per 1000 Assumed benefit 289 per 1000 85.2% ranging from 27.0% to across control groups, edge score was 56.9% Usual care : adults considering treatment or screening decisions : patient decision aid : usual care : all settings Comparison tions - all studies Accurate risk percep- Congruence between the chosen option and informed values - all studies Assessed soon after exposure to the deci- sion aid Patient decision aids compared with usual care for adults co exposure to the deci- sion aid Assessed soon after sure to the decision aid edge), soon after expo- to 100 (perfect knowl- from 0 (no knowledge) Standardized on score Knowledge - all studies Outcomes Patient or population Intervention Settings S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N 4 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

8 trolled decision making is better decisions; lower pro- involvement in making aim to increase patient Patient decision aids Lower scores indicate feeling clearer about values feeling more informed Lower scores indicate portion of clinician-con- a,e © a,b a,b High High Moderate ⊕⊕⊕⊕ ⊕⊕⊕ ⊕⊕⊕⊕ (16 studies) 5068 3180 5707 (23 studies) (27 studies) tion, and no other adverse effects reported (0.55 to 0.83) - - RR 0.68 28 lower (12.20 to 6.36 clear values in the inter- The mean feeling un- informed in the inter- vention groups was 9. 155 per 1000 (125 to 189 per 1000) vention groups was 8. 81 lower (11.99 to 5.63 The mean feeling un- lower) lower) 25 as- c ≤ 25 associated ≤ Scores > 38 associated on decisions. with following through making with delay in decision making 1 to 61.1 Scores sociated with follow- come ’feeling unin- The mean for outcome ’feeling unclear about personal values’ ranged across control groups from 15.5 to 53.2 Scores formed’ ranged across control groups from 11. through with decisions. Scores > 38 associated 228 per 1000 The mean for out- There were no adverse effects on health outcomes or satisfac with delay in decision ician-controlled deci- sion making - all stud- ies Assessed soon after consultation with clini- cian informed subscale - all Decisional conflict: un- clear about personal values subscale - all studies Standardized on score from 0 (not unclear) to 100 (unclear) Assessed soon after exposure to the deci- sion aid Standardized on score from 0 (not unin- formed) to 100 (un- informed) Assessed soon after exposure to the decision aid Decisional conflict: un- studies Adverse events Participation in de- cision making: clin- 5 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

9 mate. (and its 95% confidence interval) is ange the estimate. corresponding risk ded in footnotes. The r confidence in the estimate of effect and may change the esti of the intervention (and its 95% CI). in the estimate of effect. on our confidence in the estimate of effect and is likely to ch . As well, the outcome was measured using various heterogeneity given the generally consistent direction of t rate. y. at high risk of bias. relative effect l care groups. (e.g. the median control group risk across studies) is provi : risk ratio RR assumed risk : further research is likely to have an important impact on ou : we are very uncertain about the estimate. : further research is very unlikely to change our confidence : further research is very likely to have an important impact : confidence interval; The GRADE rating was downgraded given the lack of directness The GRADE rating was downgraded given the lack of precision. The vast majority of studies measuring this outcome were not The GRADE ratings for these outcomes were not downgraded for The GRADE rating was downgraded given the lack of consistenc The data source for the assumed risk was the mean control even CI Moderate quality High quality Very low quality Low quality *The basis for the GRADE Working Group grades of evidence based on the assumed risk in the comparison group and the approaches with no gold standard approach. f d e c effects across studies for the decision aid compared to usua b a 6 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

10 ceptual model or theoretical framework ( Mulley ; Durand 2008 B A C K G R O U N D 1995 O’Connor 1998b ). Rothert 1987 ; ; Many health treatment and screening decisions have no single In response to concerns about variability in the quality of pat ient nce- ’best’ choice. These types of decisions are considered ’prefere decision aids, the IPDAS Collaboration reached agreement on cr i- sensitive’ because there is insufficient evidence about outcome s or ). More than 100 re- Elwyn 2006 teria for judging their quality ( there is a need to trade off known benefits and harms. Clinical searchers, clinicians, patients, and policymakers from 14 coun- Evidence analyzed 3000 treatments, classifying 50% as having in- quality: tries participated. Participants addressed three domains of ng sufficient evidence, 24% as likely to be beneficial, 7% as requiri clinical content, development process, and evaluation of a pati ent - trade-offs between benefits and harms, 5% as unlikely to be ben decision aid’s effectiveness. A series of background papers in form- s eficial, 3% as likely to be ineffective or harmful, and only 11% a IPDAS ing the original IPDAS criteria were updated in 2013 ( being clearly beneficial ( ). Not only does Clinical Evidence 2013 2013 ). Subsequently, an international team of researchers reached one have to take into account the strength of the evidence, but consensus on a shorter set of qualifying and certifying criteri a op- even for the 11% of treatments that show beneficial effects for p ( Joseph-Williams 2013 ). Informed by IPDAS, the Washington ulations, physicians need to translate the probabilistic na ture of State Health Authority launched the first programme for certif y- on the evidence for individual patients to help them reach a decisi ing patient decision aids in 2016 ( ). Washington State 2016 based on informed values. Patient decision aids are an interv ention that can be used to present such evidence ( Brouwers 2010 ). This review is an update of the review last published in 2014 of the How the intervention might work Stacey comparisons between patient decision aids and usual care ( 2014b ). To provide a more focused review, we removed 28 studies Decision aids can be used before, during, or after a clinical en- that compared detailed versus simple decision aids. counter to enable patients to become active, informed partici- the pants. Providing the patient decision aid in preparation for consultation allows people more time to digest the informati on and be ready to discuss the decision, but this may not be feasi- Description of the intervention ble in some health decisions (e.g. antibiotics for upper respir atory According to the International Patient Decision Aids Standard s k- infections). Decision aids can also facilitate shared decision ma (IPDAS) Collaboration ( IPDAS 2005a Elwyn 2006 ; ; Joseph- ing. Shared decision making is defined as a process through which Williams 2013 ), decision aids are evidence-based tools designed to clinicians and patients make healthcare choices together ( Charles help patients make specific and deliberated choices among healt h- 1997 ), representing the crux of people-centred care Makoul 2006 ; care options. Patient decision aids supplement (rather than re - ( Weston 2001 ). However, the way the clinician provides infor- f place) clinicians’ counselling about options. The specific aims o mation may strongly affect people’s preferences ( ), Hibbard 1997 decision aids and the type of decision support they provide may prompting the need for standardized information such as pati ent vary slightly, but in general they: decision aids. Patients who are more active in making decisions 1. explicitly state the decision that needs to be considered; about their health have better health outcomes and healthcare ex- 2. provide evidence-based information about a health ; ). In summary, patient de- Kiesler 2006 Hibbard 2013 periences ( ities, condition, the options, associated benefits, harms, probabil cision aids may help clinicians and patients come to quality de- and scientific uncertainties; cisions, grounded in patients’ values and taking into account t he 3. help patients to recognize the values-sensitive nature of t he potential trade-offs in benefits and risks of different optio ns. decision and to clarify, either implicitly or explicitly, the va lue they place on the benefits and harms. (To accomplish this, l patient decision aids may describe the options in enough detai Why it is important to do this review l, that clients can imagine what it is like to experience the physica emotional, and social effects, or they may guide clients to deci- Given the broad range of stakeholders interested in patient consider which benefits and harms are most important to them.) s a sion aids and the rapidly expanding field of research, there wa Decision aids differ from usual health education materials. D eci- s need to update this review to identify studies on new decision sion aids make the decision being considered explicit, providi ng or conducted in new countries and to strengthen the synthesized a detailed, specific, and personalized focus on options and out- hat evidence supporting use of patient decision aids for outcomes t comes for the purpose of preparing people for decision making. do not yet have high-quality evidence. In fact, the 2014 publica- In contrast, health education materials help people to unders tand tion was the most cited Cochrane Review in 2015 based on 1888 their diagnosis, treatment, and management in general term s, but reviews published in 2013 and 2014. With growing developmen t given their broader perspective, these materials are not focu sed on of patient decision aids for use in the consultation, we wanted decision points and thus do not necessarily help them to partici - to conduct a subgroup analysis of patient decision aids used in pate in decision making. Many decision aids are based on a con- preparation for versus within the consultation. 7 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

11 or coaching in the steps of making and communicating decisions Results from previous reviews were used to inform clinical pra ctice with others. guidelines such as Patient Experience in Adult NHS Services ( ut We excluded studies if interventions focused on: decisions abo NCGC/NICE2012 )andDecisionSupportforAdultsLiving with ectives lifestyle changes, clinical trial entry, or general advance dir Chronic Kidney Disease ( ). Subgroup analyses of RNAO 2009 (e.g. do not resuscitate); education programmes not geared to a includedstudieshave focusedonanxiety( Bekker 2003 ), adherence r- specific decision; and interventions designed to promote adhe ( ), participant Munro 2016 Trenaman 2016 ), values congruence ( ence or elicit informed consent regarding a recommended option. Brown 2015 trial identity ( ), and heterogeneity ( Gentles 2013 ). not We also excluded studies when the relevant decision aid(s) were Other systematic reviews have been conducted on the use of pa- e- available to us and not adequately described in the article(s), b hared tient decision aids as one type of intervention to facilitate s cause we could not determine the aids’ characteristics and whethe r ; Duncan 2010 ; Coyne 2013 decision making in clinical practice ( ion or not they met the minimum criteria to qualify as patient decis Elwyn 2013 ). ; ; Legare 2010 Legare 2014 aids. Types of comparisons O B J E C T I V E S We included studies that compared patients exposed to a patien t decision aid to patients in comparison groups that were expose d to To assess the effects of decision aids in people facing treatmen t or usual care, general information, clinical practice guideline, p lacebo screening decisions. eview, intervention, or no intervention. For the purposes of this r we refer to all such control comparisons as ’usual care’. We excluded studies that compared two different types of patie nt M E T H O D S decision aids. Types of outcome measures Criteria for considering studies for this review To ascertain whether the decision aids achieved their objective s, we examined a broad range of outcomes. Although the decision aids Types of studies es focused on diverse clinical decisions, many had similar objectiv such as improving knowledge scores, the accuracy of risk percep- We included all published studies that used a randomized con- tions, and participation in decision making. Many of these eva l- trolled trial (RCT) design evaluating patient decision aids. uation criteria mapped onto the International Patient Decisi on Aids Standards (IPDAS) criteria for evaluating the effectiven ess of Types of participants decision aids ( ; ). The Sepucha 2013 ; IPDAS 2005b Elwyn 2006 o We included studies involving adults aged 18 years or older wh IPDAS criteria were attributes related to the choice (e.g. match were making decisions about screening or treatment options fo r between the chosen option and the features that matter most to the - themselves, a child, or an incapacitated significant other. We ex informed patient) and to the decision-making process (e.g. help s cluded studies in which participants were making hypothetical patients to recognize that a decision needs to be made; know the choices. e deci- options and their features; understand that values affect th sion; be clear about the features that matter most; discuss val ues with their clinician; and become involved in their preferred wa ys). Types of interventions A complete list of outcomes, specified in advance of the review, rt We included studies that evaluated a patient decision aid as pa included primary and secondary outcomes. of the intervention. Decision aids were defined as interventi ons designed to help people make specific and deliberated choices Primary outcomes among options (including the status quo), by making the deci- sion explicit and by providing (at the minimum) information on as the options and outcomes relevant to a person’s health status Evaluation criteria that map onto the IPDAS criteria well as implicit methods to clarify values. The aid also may hav e included: information on the disease/condition; costs associa ted • Attributes of the choice made: does the patient decision aid l health with options; probabilities of outcomes tailored to persona improve the match between the chosen option and the features risk factors; an explicit values clarification exercise; informa tion that matter most to the informed patient (demonstrated by on others’ opinions; a personalized recommendation on the ba sis outcomes such as knowledge, accurate risk perceptions, values- of clinical characteristics and expressed preferences; and guida nce choice congruence)? 8 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

12 Electronic searches • Attributes of the decision-making process: does the patient be decision aid help patients to recognize that a decision needs to For this update, we used the same search strategy that was revi sed made, feel informed about the options and their features, be by the Trials Search Coordinator at the Cochrane Consumers and clear about the option features that matter most, discuss valu es Communication Group in the last update ( ). Stacey 2014b with their clinician, and become involved in decision making? fol- Therefore, the cumulative search of electronic databases is as lows. • Cochrane Central Register of Controlled Trials Other decision-making process variables (CENTRAL; 2015, Issue 6) in the Cochrane Library (searched to 24 April 2015). Decisional conflict • MEDLINE Ovid (1966 to 24 April 2015). • • Patient-clinician communication • Embase Ovid (1980 to 24 April 2015). • Participation in decision making • PsycINFO Ovid (1806 to 24 April 2015). Proportion undecided • CINAHL Ovid (1982 to September 2008), then in Ebsco • • Satisfaction with the choice, with the process of decision (to 24 April 2015). making, and with the preparation for decision making We present the search strategies in . Appendix 2 and Appendix 1 Secondary outcomes Searching other resources On 18 December 2015 we also searched trial registries (World Health Organization, ClinicalTrials.gov), the Internet usi ng Behaviour Google and Google Scholar, and the Decision Aid Library Inven- tory ( ). Finally, reference lists of all newly in- decisionaid.ohri.ca • Choice (the actual choice implemented; if not reported, the cluded trials were searched. participants’ preferred option was used as a surrogate measu re) Adherence to chosen option • Data collection and analysis Health outcomes For this current update, we focused only on new publications • Health status and quality of life (generic and condition- that had appeared since the previous publication ( Stacey 2014b ), specific) and we limited the inclusion to patient decision aids versus us ual Anxiety, depression, emotional distress, regret, confidence • t care. As such, we removed studies from the previous reviews tha compared detailed versus simple patient decision aids to prov ide a more focused review. Healthcare system Costs, cost-effectiveness • Selection of studies • Consultation length Pairs of eight review authors (CB, DS, RT, MB, MHR, KE, NC, • Litigation rates DR) screened all identified citations. We retrieved the full te xt of any papers identified as potentially relevant by at least one au- thor, listing all papers excluded from the review at this stag e, with reasons, in the ’Characteristics of excluded studies’ table. W e also Search methods for identification of studies provided citation details and any available information abo ut on- onal going studies, and we collated and reported details of additi Our search strategy for the review included: publications, so that each study (rather than each report) was th e 1. searching electronic medical and social science databases; unit of interest. We report the screening and selection process in and 2. searching other resources. Figure 1 . 9 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

13 Figure 1. Study flow diagram. 10 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

14 the effects were expected to be independent of the type of deci- Data extraction and management e sion studied. For example, we expected decision aids to improv We Two research assistants extracted data independently (KL, IS). knowledge and create accurate percetions of options, benefits, a nd compared findings and resolved inconsistencies through discus- harms; to reduce decisional conflict; and to enhance active partic- with sion with the principal investigator (DS) and, when necessary, ipation in decision making. Therefore, we pooled data from in - a co-author (CB). No review authors extracted data for their own cluded RCTs for these outcomes if trials used comparable mea- studies in this update nor in any other versions of this revie w. wl- sures. To facilitate pooling of data for some outcomes (e.g. kno ger One review author entered all extracted data into Review Mana edge, decisional conflict), we standardized the scores to range fr om 5 (RevMan 5), and a second one worked independently to check 0 to 100 points. When analysing the effects of decision aids on for accuracy against the data extraction sheets ( RevMan 2014 ). choices, we pooled outcomes on more homogeneous subgroups of decisions (choice of major surgery versus conservative option s; screening test or not, etc.). Assessment of risk of bias in included studies ing the Two research assistants independently appraised studies us Unit of analysis issues Higgins Cochrane ’Risk of bias’ tool (current update: KL, IS) ( We checked for unit-of-analysis errors. Where we found errors 2011 , Chapter 8). We judged each item as conferring high, low, ata and sufficient information was available, we re-analyzed the d or unclear risk of bias as set out in the criteria provided by Higgins using the appropriate unit of analysis by taking account of the , and we provided a quote from the study report and a justi- 2011 intracluster correlation (ICC). We obtained estimates of the IC C le. fication for our judgement for each item in the ’Risk of bias’ tab - by contacting authors of included studies, or we imputed them us ent in- For the item on ’other’ potential sources of bias, the assessm ing estimates from external sources. For two studies ( Kupke 2013 ; h cluded: whether the same clinician provided consultation to bot Lewis 2010 ), it was not possible to obtain sufficient information the intervention and usual care groups with measures taken po st- to re-analyze the data, and we reported these studies as being at sis; consultation, whether clustering was accounted for in the analy high risk for ’other’ bias based on these unit-of-analysis err ors. We and potential sources of bias reported by the authors in the st udy made no adjustments to the data based on these two studies tha t limitations. were included in meta-analysis for knowledge only. - We resolved inconsistencies by discussion with the principal in vestigator (DS) and, when necessary, with a co-author (CB). No review authors appraised risk of bias for their own studies i n this Dealing with missing data update nor in any other versions of this review. re Studies were deemed to be at the highest risk of bias if they we We contacted authors to obtain missing data. Where possible, we scored as at high risk on any of the items of the risk of bias tool conducted analysis on an intention-to-treat basis; otherwise, we Higgins 2011 ). ( o analyzed data as reported. We reported on the levels of loss t follow-up and assessed this as a source of potential bias. Measures of treatment effect Assessment of heterogeneity For dichotomous outcomes, we analyzed data based on the num- uped For this update and in previous versions of the review, we gro ber of events and the number of people assessed in the interve n- s to studies together across populations and settings. The aim wa tion and comparison groups. We will use these to calculate the ss enable an assessment of the effectiveness of decision aids acro risk ratio (RR) and 95% confidence interval (CI). For continuous measures, we analyzed data based on the mean, standard devia tion ven conditions, rather than to focus on disease-specific contexts. Gi nd (SD) and number of people assessed for both the intervention a e of that decision aids are a well-defined and clearly delineated typ comparison groups to calculate mean difference (MD) and 95% n intervention, we decided that this approach was defensible. O CI. aid the basis of grouping studies across populations and decision First, we described study characteristics individually. The a p ri- gree elements, we anticipated that there would be a substantial de in ori comparison was usual care versus decision aids. For studies of heterogeneity in our pooled effect estimates. However, we de- d which there were more than one intervention group, we extracte l- cided that we would consider the direction of effects and variabi data from the groups that provided the strongest contrast bet ween ity in these rather than variability in the size of effects, as the ma- the intervention and control groups. We pooled results across stud- jor basis for our interpretation of heterogeneity. This mea nt that ies in cases where investigators used similar outcome measure s and for those pooled effect estimates where the direction of effect was 11 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

15 excluded studies that were at high risk of bias for any of the cat e- consistent across studies, we did not downgrade for inconsiste ncy, gories in the ’Risk of bias’ assessment ( despite some variability in the size of effects across individ ual stud- ). Higgins 2011 ies. We did downgrade for inconsistency for one outcome: con- gruence between the chosen option and informed values. This wa s ’Summary of findings’ table g because there is no accepted gold standard measure for assessin ts We prepared a ’Summary of findings’ table to present the resul this outcome, and we considered that variability in measureme nt of meta-analysis, based on the methods described in Chapter 11 fects by the included studies added further uncertainty about the ef of the Cochrane Handbook for Systematic Reviews of Interventions of decision aids for this outcome. ). We presented the results of meta-analysis Schünemann 2011 ( e Where heterogeneity was present in pooled effect estimates, w for the major comparison of the review for each of the key out- roup explored possible reasons for variability by conducting subg comes. We provided a source and rationale for each assumed risk analysis in the 2009 update ( ). The post hoc O’Connor 2009b cited in the table and used the GRADE criteria to rank the qualit y analysis included the IPDAS effectiveness criteria to explore het- of the evidence for each outcome on each of the following do- ion erogeneity according to the following factors: the type of decis , and mains: risk of bias, inconsistency, imprecision, indirectness id (treatment versus screening), the type of media of the decision a publication bias. Two authors independently assessed the qu ality i- (video/computer versus audio booklet/pamphlet), and the poss of the evidence using the GRADEprofiler (GRADEpro) software the bility of a ceiling effect based on usual-care scores (resulting in ( ). GRADEpro GDT removal of studies with lower scores for knowledge and accurate risk perception and higher scores for decisional conflict using th e subscales measuring levels of informedness and clarity of val ues). We analyzed the effect of removing the biggest outlier(s) (defin ed by visual inspection of forest plots). Given that the post hoc a naly- R E S U L T S e- sis did not alter the findings from the 2009 update, we did not r conduct the post hoc analysis for the IPDAS effectiveness criter ia. Description of studies The current version of our review updates our 2014 version, Stacey 2014b , with 18 new studies ( Chabrera ; Brazell 2014 ; Bozic 2013 Assessment of reporting biases ; Fraenkel 2012 ; Knops 2014 ; Köpke 2014 Kuppermann ; 2015 We used funnel plots to assess publication bias. 2014 ; Legare 2012 ; Lepore 2012 ; Lam 2013 ; ; LeBlanc 2015 ; Shourie 2013 ; Sawka 2012 ; ; Stacey Mathers 2012 Mott 2014 2014a ). For this update, we excluded Williams 2013 ; Taylor 2006 ; Data synthesis 28 previously included studies due to the comparisons being li m- We used RevMan 5 software to estimate a weighted interventio n Deschamps ited to detailed versus simple patient decision aids ( RevMan 2014 ). For contin- effect with 95% confidence intervals ( ; ; Deyo 2000 ; 2004 Green 2004 Hunter ; ; Goel 2001 Dodin 2001 uous measures, we used mean differences (MD); for dichotomous 2005 ; Lalonde 2006 Legare ; ; Kuppermann 2009 Labrecque 2010 ; ll outcomes, we calculated pooled relative risks (RR). We analyzed a 2003 O’Connor 1998a ; Myers 2011 ; Myers 2005a ; Leung 2004 ; ; data with a random-effects model because of the diverse nature o f ; Rostom 2002 ; Raynes-Greenow 2010 Rothert O’Connor 1999a ; the studies being combined and then anticipated variability i n the 1997 Street 1995 ; Solberg 2010 ; Schapira 2007 Schapira 2000 ; ; ; populations and interventions of the included studies. We su m- ; Wakefield 2008a ; Volk 2008 ; Van Roosmalen 2004 ; Tiller 2006 he marized all of the results for the primary outcomes and rated t ). Wakefield 2008c ; Wakefield 2008b strength of evidence using GRADE ( ), presenting Andrews 2013 these in a ’Summary of findings’ table ( Higgins 2011 ). Results of the search Intotal,we identified46,054citationsfromthe electronicda tabase Subgroup analysis and investigation of heterogeneity ssed searchesand258citationsfromothersources.Of these,we asse For this update, we conducted a subgroup analysis to compare ). Figure 1 504 citations for eligibility using the full text (see the effects of the intervention when used in preparation for t he consultation with the effects of those used during the consulta tion Included studies to usual care. The remaining 151 citations provided data on 105 studies that met our inclusion criteria, 18 of which are new for this update. T he Sensitivity analysis 105 RCTs, involving 31,043 participants, presented results from fect We performed post hoc sensitivity analyses to examine the ef 10 countries: Australia (10 studies), Canada (15 studies), China (1 lysis of excluding studies of lower methodological quality. The ana study), Finland (2 studies), Germany (6 studies), Netherlands (2 12 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

16 for tooth decay (1 study); measles, mumps, and rubella vaccine studies), Spain (1 study), Sweden (1 study), the UK (16 studies), for infants (1 study); treatment of post-traumatic stress diso rder the USA (50 studies), and Australia plus Canada (1 study). We present study details below and in the Characteristics of included (1 study); and radioactive iodine treatment for thyroid cancer (1 studies study). table. The decision aids used different formats and were compared to a orma- variety of control interventions (e.g. usual care, general inf Unit of randomization noted tion, no intervention, guideline, placebo intervention). We Ninety studies randomized individual patients, and 15 rand om- the nature of usual care when reported (see Characteristics of ized clusters. For cluster trials, randomized 12 com- Allen 2010 table). For this review, we have grouped control included studies pany worksites; Fraenkel 2012 , 2 groups of primary care physi- interventions and refer to them as ’usual care’. cians; , Kupke 2013 , 12 inpatient psychiatric units; Hamann 2006 According to the definition of a patient decision aid, all of the Legare 2011 , 4 family medicine group prac- 49 dental students; ation studies evaluated patient decision aids that included inform tices; , 12 family medicine group practices; Lewis Legare 2012 about the options and outcomes and provided at least implicit 2010 Loh 2007 , 32 family medicine group practices; , 30 gen- clarification of values. Most patient decision aids included inf or- Mathers 2012 , 49 general medicine practices; eral practitioners; mation on the clinical problem (90.5%) as well as outcome proba- McAlister 2005 , 102 primary care practices; Mullan 2009 , 40 clin- bilities (89.5%). Fewer patient decision aids provided guidan ce in icians; Nagle 2008 Shourie 2013 , 60 general practitioners; , 50 the steps of decision making (65.7%), explicit methods to clarify Weymiller 2007 , 21 endocrinologists; general medicine practices; values (57.1%), and/or examples of others’ experiences (41.0%) and , 27 surgeons. Whelan 2004 (see table ). Characteristics of included studies For 10 studies ( Legare 2012 ; Loh Allen 2010 ; Legare 2011 ; ; ; ; Nagle 2008 ; Shourie 2013 ; Mathers 2012 Mullan 2009 2007 Excluded studies Whelan 2004 ), the cluster effect was taken into ; Weymiller 2007 account in the published outcome data, and the meta-analysis use d We excluded 302 studies upon close perusal of the relevant pape rs published results. Although did not account for Hamann 2006 (see ). The reasons for exclusion Characteristics of excluded studies the cluster effect in the published outcome data, the way this st udy were: the study was not a randomized controlled trial; the deci - was reported did not allow us to include it in the meta-analysis , t of sion was hypothetical, with participants not actually at a poin ly. so we did not re-analyze the data and report the study separate decision making; the intervention was not focused on making a For McAlister 2005 , meta-analysis was done applying the design choice; the intervention offered no decision support in the for m nt effect (based on the published intracluster correlation coefficie of a decision aid or did not provide enough information about (ICC)). For Fraenkel 2012 , the authors stated that adding a ran- the decision aid; no comparison outcome data were provided; the - dom effect for physician clusters did not contribute to better-fit l; study did not evaluate the decision aid; the study was a protoco The ting regression models, and we removed it from the analysis. the decision aid was about clinical trial entry, lifestyle choice , or did not account for clus- Kupke 2013 and Lewis 2010 analysis by ta- advanced care planning; the study involved testing the presen tering. tion of the decision aid, but with no difference in the content of the decision aid between study groups; or the study compared a detailed versus simple decision aid. Decision aids and comparisons n We also identified 61 ongoing RCTs through trial registratio The 105 included studies evaluated decision aids that focused o n ec- databases, personal contact, and published protocols in the el 50 different decisions ( Table 1 ). The most common decisions were tronic database searches (see references to Ongoing studies and about prostate cancer screening (14 studies), colon cancer screen- table Characteristics of ongoing studies ). ing (10 studies), medication for diabetes (4 studies), breast cance r i- genetic testing (4 studies), prenatal screening (4 studies), med cation for atrial fibrillation (4 studies), and surgery (mastecto my Risk of bias in included studies e can- for breast cancer, 4 studies; hysterectomy, 3 studies; prostat cer treatment, 4 studies). New decision topics added in this up- Details on the ratings and rationale for risk of bias are in th e d date included surgery for prolapsed pelvic organs (1 study) an Characteristics of included studies table and displayed in Figure 2 asymptomatic aortic abdominal aneurysm (1 study); restorati on and Figure 3 . 13 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

17 Figure 2. Risk of bias summary as percentages across all incl uded studies. 14 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

18 Figure 3. Risk of bias summary for each included study. 15 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

19 Allocation high risk of bias given that a large number of potential parti cipants When assessing risk of selection bias, we rated all 105 studies did not participate in the study. , Hamann 2006 , and Kupke 2013 as being at low or unclear risk of bias. Allocation concealment Lewis 2010 did not account for clustering in their analyses. methods prompted a rating of low or unclear risk of bias in 104 ). Kupke 2013 studies and high risk of bias in 1 study ( Effects of interventions Summary of findings for the main comparison See: , In addition to Summary of findings for the main comparison Blinding see the Data and analyses figures for pooled data and Additional We judged 102 studies to be at low or unclear risk of performance nts tables for outcome data that we did not pool. This section prese nel, and detection bias for the blinding of participants and person n the attributes of the choice made, the attributes of the decisio while 3 (2.9%) studies were at high risk of bias. High risk of process, and secondary outcomes. bias was due to lack of blinding of physicians to the status of patients randomized to the patient decision aid and alternat ive interventions ( Krist 2007 ; ). Man-Son-Hing 1999 ; Auvinen 2004 Primary outcomes We rated the blinding of outcome assessment as leading to low o r unclear risk of bias in all 105 studies. Attributes of the choice made: does the patient decision aid improve the match between the chosen option and the Incomplete outcome data features that matter most to the informed patient? For 103 studies, aspects related to incomplete outcome data con- re- The randomized controlled trials used three measures that cor ferred low or unclear risk of bias. In two (1.9%) studies ( Chambers spond to this outcome: knowledge scores, accuracy of risk percep- ), there was high risk of bias due to high attrition Mott 2014 ; 2012 ’s tions, and congruence between the chosen option and the patient rates. values. Selective reporting Knowledge We rated all 105 studies as being at either low risk of bias beca use e- Seventy-one of the 105 studies (67.6%) assessed the effects of d e- the protocol was registered publicly or at unclear risk of bias b ased cision aids on knowledge. The studies’ knowledge tests were b g cause we could not assess the extent or the impact of any reportin f on information contained in the decision aid. The proportion o bias. g accurate responses was transformed to a percentage scale rangin from 0% (no correct responses) to 100% (fully correct responses). more There is high-quality evidence that patient decision aids were Other potential sources of bias effective than usual care (52 studies) on knowledge scores (MD Of 105 studies, we rated 98 as being at low or unclear risk of 13.27, 95% CI 11.32 to 15.23; Analysis 1.1 ). In absolute terms other potential sources of bias. The other seven (6.7%) discuss ed s the group receiving usual care had, on average, 57 of 100 answer other potential risks of bias ( correct. Those in the decision aid group scored better, with 70 of ; Clancy 1988 Brazell 2014 ; Hamann 100 answers correct on average (from 68 to 72 correct). 2006 ; Kupke 2013 ; Knops 2014 ; ). Lewis 2010 ; LeBlanc 2015 t could Nineteen additional studies presented knowledge scores tha We rated Brazell 2014 LeBlanc 2015 as being at high risk and not be included in the pooled outcome (see Table 2 ). Most of these of bias given that the same physicians provided consultation t o other studies reported statistically-significantly higher kn owledge n both intervention and control groups, and measures were take describes a potential Clancy 1988 scores for those exposed to the decision aid compared to usual car e. after physician consultation. re for selection bias because non-randomized medical residents we The funnel plot for knowledge as an outcome in studies comparin g added to the decision aid group, and there was a low response ra te decision aid to usual care shows that these studies are at low ri sk among those offered decision aid. We rated for publication bias ( Figure 4 as being at Knops 2014 ). 16 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

20 Figure 4. Funnel plot of comparison: 1 Knowledge, outcome: 1 .1 Knowledge - all studies. Accurate risk perceptions (i.e. perceived probabilities o f re participants allocated to the decision aid had a significantly mo outcomes) out accurate perception of their estimated cardiovascular risk with statin therapy compared to the usual care group; this effect was Of 105 studies, 25 (23.8%) examined the effects of patient de- greater when the clinician used the decision aid during the con- of cision aids on the accuracy of patients’ perceived probabilities sultation rather than when the researcher used the decision ai d in outcomes (see ; Table 3 ). We classified the accuracy of Analysis 2.1 preparation for the consultation (P = 0.03). For the final interaction perceived outcome probabilities according to the percentage of i n- study by , three of eight knowledge test items mea- Mann E 2010 ence dividuals whose judgments corresponded to the scientific evid r total sured accurate risk perceptions, but results were presented fo about the chances of an outcome for similar people. For studies knowledge and not individual items.The funnel plot for accura te that elicited risk perceptions using multiple items, we avera ged the risk perception as an outcome in studies comparing decision aid proportion of accurate risk perceptions. to usual care shows low risk for publication bias. There is moderate-quality evidence that patient decision aids were r- more effective than usual care for transmitting accurate risk pe ceptions (risk ratio (RR) 2.10, 95% CI 1.66 to 2.66, 17 studies; Congruence between chosen option and values Analysis 2.1 ). This means that for every 1000 people receiving Of 105 studies, 16 (15.3%) measured congruence between the as far usual care, 269 were likely to accurately interpret risk, where chosen options and the patients’ values. Six measured values - more people (565 people per 1000; from 447 to 716) accurately choice congruence without considering knowledge ( interpreted risk after using a decision aid. Arterburn ling Eight studies reported results that were not amenable to poo Legare 2008a ; Berry 2013 ; 2011 ; ; Lerman 1997 ; Frosch 2008a (see Table 3 ). Fraenkel 2012 ; ). Of 10 studies that measured informed values- ; Kuppermann 2014 ; Vandemheen 2009 Hanson 2011 Smith 2010 ; and Mathieu 2010 reported a statistically significant choice congruence, eight used the Multi-Dimensional Measure improvement in accurate perceptions of outcomes for the decision ; Bjorklund 2012 ; Fagerlin 2011 Mathieu of Informed Choice ( aid group compared to usual care, and 2007 ; Steckelberg 2011 reported no ef- Miller 2005 Mathieu 2010 ; Nagle 2008 ; Smith 2010 ; ; fect on risk perception. In another study, reported Weymiller 2007 Trevena 2008 ), which assesses the extent to which the choice is 17 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

21 based on relevant knowledge, is consistent with a person’s va lues/ ards In relation to the International Patient Decision Aids Stand attitudes, and is behaviourally implemented ( ). These Michie 2002 (IPDAS) decision process criteria, no studies evaluated the ext ent cores studies operationalized the measure in terms of knowledge s that to which patient decision aids helped participants to recognize higher than the mid-point of the scale, attitude scale scores hig her e a decision needed to be made or understand that values affect th thanthe mid-point, andchoice beingcongruentwith attitude.Tw o bout decision. Some studies measured participants’ self-reports a Schwalm other studies measured informed values-based choice: feeling informed and clear about personal values. The measur es assessed the extent to which the choice was based on knowl- 2012 used to evaluate these criteria were two subscales of the previ ously edge score ≥ 60% and a score for three values-importance ratings O’Connor 1995 ). validated Decisional Conflict Scale (DCS) ( Stacey 2014a assessed the extent to that matched the choice; and 66% and mea- which the choice was based on knowledge score ≥ Decisional conflict l. sured values-choice congruence using a logistic regression mode For the 10 studies that measured informed values-choice congru- Of 105 studies, 63 (60.0%) evaluated decisional conflict using ence, two used preferred choice ( Trevena 2008 ; Mathieu 2010 ), the DCS ( ). The DCS is reliable, discriminates O’Connor 1995 and the other eight used actual choice. ge, between those who make or delay decisions, is sensitive to chan There is low quality evidence that patient decision aids were m ore and discriminates between different decision support interv en- effective than usual care for selecting an option that was congru ent Morgan 2000 ; O’Connor 1995 ; ). The O’Connor 1998b tions ( with their informed values (RR 2.06, 95% CI 1.46 to 2.91, 10 scale measures the constructs of overall decisional conflict and th e studies; ). Of the 10 studies, 8 individually showed Analysis 3.1 particular factors contributing to uncertainty (e.g. feeling un cer- statistically higher congruence scores for the patient decision aid tain, uninformed, unclear about values, and unsupported in d e- compared to usual care, and 2 showed no difference ( Bjorklund cision making). A final subscale measures perceived effective de- 2012 ). Repeating this analysis using the studies Mathieu 2010 ; cision making. The scores were standardized to range from 0 (no that measured actual choice and not preferred choice revealed a decisional conflict) to 100 points (extreme decisional conflict). pooled RR of 2.13 (95% CI 1.44 to 3.14; 8 studies). A sub-anal- Scores of 25 or lower are associated with follow-through with de - med ysis of studies using the Multi-Dimensional Measure of Infor cisions, whereas scores that exceed 38 are associated with delay Choice revealed a pooled RR of 2.08 (95% CI 1.40 to 3.08, 8 O’Connor 1998b ). When decision aids are in decision making ( ). studies; Analysis 3.3 compared to usual care, a negative score indicates a reduction in There was no difference between patient decision aid and usual decisional conflict, favouring the decision aid. care for the six studies that measured values-choice congruence .1 summarizes the decisional conflict results for the 42 Analysis 4.1 Berry Arterburn 2011 ; without considering knowledge scores ( studies that compared decision aids to usual care. We report on ; Lerman 1997 ; Vandemheen Frosch 2008a ; ; Legare 2008a 2013 21 studies that were not amenable to pooling in (original Table 5 2009 ; see Table 4 ). We did not pool these studies because of how DCS), (SURE test Table 7 (low literacy version), and Table 6 they reported results. Arterburn 2011 reported that, compared to version). d a the control group, those exposed to the decision aid experience 7.22 points − The mean difference (MD) for total DCS scores was more rapidearlyimprovementof value-choice concordance imme- out of 100, favouring the patient decision aid over usual care reported that women’s valu- Legare 2008a diately after exposure. groups (95% CI .1). Sixteen Analysis 4.1 − − 9.12 to 5.31; see ing of the non-chemical aspect of natural health products was studies that could not be pooled ( Table 5 ) reported mixed results positively associated with their choice of natural health prod ucts ver- on the original DCS. Of four studies that used the low literacy in managing menopausal symptoms (P 0.006). The other four = sion ( ; ; Fraenkel 2012 ), Williams 2013 ; Taylor 2006 Smith 2010 studies reported no differences between groups. However, Frosch all reported statistically significant improvement (i.e. redu ced) in observed that men exposed to the decision aid who chose 2008a total (or subscale) decisional conflict scores in the decision aid - not to have a prostate-specific antigen (PSA) test rated their con reported Stacey 2014a ). Table 6 group, compared to usual care ( cern about prostate cancer lower than men who requested a PSA no difference between groups using the SURE test version. test, while men assigned to the usual care group provided simi lar - The ’feeling uninformed’ subscale of the DCS measures self-re ratings of concern regardless of their PSA choice. d ported comfort with knowledge, not actual knowledge. We electe to consider this as a process measure and to reserve the gold sta n- dard of objective knowledge tests for assessing decision qual ity. lp Attributes of the decision process: does the decision aid he e There was high-quality evidence that patient decision aids wer patients to recognize that a decision needs to be made, more effective than usual care in reducing patients’ ’feeling u nin- know the options and their features, understand that values − 9.28, 95% CI formed’ about options, benefits, and harms (MD affect the decision, be clear about the features that matter − 12.20 to − 6.36; 27 studies; Analysis 4.1 .2). The funnel plot for most to them, discuss values with their clinician, and on ’feeling uninformed’ as an outcome in studies comparing decisi become involved in their preferred ways? Figure 5 ). aid to usual care shows low risk for publication bias ( 18 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

22 Figure 5. Funnel plot of comparison: 4.1 Decisional conflict : DA vs usual care - all studies, outcome: 4.1.2 Uninformed subscale There was high-quality evidence that patient decision aids wer e more effective thanusual care forreducingpatients’ ’feeling unclear − 8.81; 95% CI − 11.99 about values’ subscale of the DCS (MD to − 5.63; 23 studies; Analysis 4.1 .3). The funnel plot for using ’feeling unclear about values’ as an outcome in studies compari ng Figure decisionaidtousual care showslowriskforpublicationbias( 6 ). 19 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

23 : DA vs usual care - all studies, outcome: 4.1.3 Figure 6. Funnel plot of comparison: 4.1 Decisional conflict Unclear subscale Patient-clinician communication ). The fifth study showed no between-group difference in Table 8 discussion of cardiovascular disease with the clinician ( Sheridan Of 105 studies, 10 (9.5%) measured the effect of decision aids on ). Table 8 ; see 2006 patient-clinician communication. Of these 10 studies, 5 evaluat ed with a patient decision aid used primarily within the consultation Participation in decision making pa- the clinician, and 5 evaluated a patient decision aid used in pre ration for the consultation. Of 105 studies, 24 (22.9%) measured the effect of decision aids Five studies compared the effect of usual care versus a decision a id on patients’ perceived participation in decision making ( Analysis used within the clinical encounter (or, in 5.1 ). Davison 1997 used the Control Preferences Scale Table 9 ; Weymiller 2007 , half the ( ter), decision aid participants were exposed just prior to the encoun ). This scale uses five response statements to measure Degner 1992 evaluating the extent of shared decision making communication -con- the role in decision making: two represent an active or patient by analyzing the audio recordings using the OPTION scale ( onse trolled role; one a shared or collaborative role; and two resp Hess ost statements represent a passive or clinician-controlled role. M 2012 ; ; LeBlanc 2015 ; Montori 2011 Weymiller Mullan 2009 ; ). The OPTION scale measures the extent to which health- other studies used comparable response statements that could be 2007 care providers use behaviours that involve patients in decisi on f- classified within each of the three groupings of the Control Pre making ( erences Scale, except for Elwyn 2005 Hamann 2006 , which used the COM- ). All five studies reported statistically higher RADE instrument to measure patient perception of involvemen mean OPTION scores in the patient decision aid group compared t, to usual care (see ent and two others that used other measures of perceived involvem Table 8 ). ( l Four of five studies reported that compared to those in the usua Hanson 2011 Table 9 ; ). ; see Loh 2007 care group, significantly higher proportions of participants e x- Using the groupings of the Control Preferences Scale, 16 of 24 ta- posed to the patient decision aid in preparation for the consul is- studies reported on clinician-controlled decision making. Cons cian tion reported that they discussed the decision with their clini tent with the hypothesis that patient decision aids increase p atient ( Hanson 2011 ; Lepore 2012 ; Sheridan 2011 ; see Fraenkel 2012 ; participation in decision making, there was moderate-quality ev- 20 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

24 idence that patient decision aids were more effective than usua l ). For results that used a sim- Table 11 and Analysis 7.1 ; see 2007 care for reducing clinician-controlled decision making (RR 0.68; ilar measure ( Analysis 7.1 ), there was high satisfaction for all par- Analysis 5.1 .1). In this field, there is no 95% CI 0.55 to 0.83; ticipants, with a median score of 82.5% for the decision aid and consensus on the hypothesized effects of decision aids on measu res 80.0% for the usual care groups. g. of patient-controlled decision making or shared decision makin e Of 105 total studies, 11 (10.5%) measured satisfaction with th Of 24 studies, 15 reported on participants assuming an active k- decision, 11(10.5%)measuredsatisfactionwith the decision-ma r (patient-controlled) role in decision making and were pooled fo ing process (see Analysis 7.6 ; plus Hess 2012 and Vodermaier 2009 analysis. Compared to usual care, decision aid use increased pa - Table 12 in ), 4 measured satisfaction with information provided tient-controlled decision making (RR 1.28, 95% CI 1.05 to 1.55; LeBlanc 2015 ; Laupacis 2006 ), 3 Oakley 2006 ; Montori 2011 ; ( measured satisfaction with the clinician ( ; .2). The 15 studies that reported on a shared decision- Analysis 5.1 Laupacis 2006 Miller 2005 making role showed no difference between decision aid and usua l ), and 1 measured satisfaction with par- Vodermaier 2009 ; care (RR 0.95; 95% CI 0.83 to 1.10; ticipating in decision making ( Kennedy 2002 .3). Analysis 5.1 ). There were mixed , Leighl 2011, results, but no studies reported that those exposed to patie nt de- Of eightstudiesthatcouldnotbe pooled, Allen 2010 cision aids were significantly less satisfied compared to usual ca re. Rubel 2010, and reported no between- Van Peperstraten 2010 he de- For results that used a similar measure of satisfaction with t group differences in these roles ( ). Three studies reported Table 9 Analysis 7.4 ), there was high satisfaction cision-making process ( that a statistically significant proportion of patients expos ed to on for all participants, with median scores of 83.8% for the decisi ) - or at least Sheridan 2011 the decision aid either participated ( aid and 77.8% for the usual care groups. Although there were no felt involved - in decision making ( ; Hamann 2006 ). Loh 2007 he in- differences between participant groups in satisfaction with t did not analyze results accounting for However, Hamann 2006 formation in the the use of design clusters. reported that a higher , clinicians using the decision aid Montori 2011 Hanson 2011 proportion described feeling involved (83% vs. 77%), but the had higher satisfaction. difference between groups was not statistically significant. Three studies (2.9%) measured satisfaction with preparation for decision making using the Preparation for Decision Making Scal e Table 13 Bennett 2010 ) ( ). Compared to usual care, two stud- ( Proportion undecided on with ies reported significant improvements in people’s satisfacti their preparation for making decisions: in after us- Fraenkel 2007 Of 105 studies, 24 (22.9%) measured the proportion of partici- ing decision aids about management of knee osteoarthritis, a nd in pants remaining undecided: of these, 22 studies could be poole d. Vandemheen 2009 regarding referral to a lung transplant centre. A significantly lower proportion of people remained undecided n this The third study found no statistically significant difference o after exposure to a decision aid (RR 0.64; 95% CI 0.52 to 0.79; subscale’s four items ( Stacey 2014a ). Analysis 6.1 ). measured progress in decision making using a single Kasper 2008 item ranging from ’0 = completely undecided’ to ’100 = made my Secondary outcomes ese decision’. Given the difference in the measure Kasper used, th both results were not included in the meta-analysis. In this study, the patients exposed to a decision aid and the usual care group Behaviour en progressed in their decision making, with no difference betwe the groups ( reported that 10.8% in the ). Table 10 Sawka 2012 patient decision aid group versus 21.6% in the usual care group Choice reported not knowing if they preferred taking adjuvant radi oactive Choice was defined as the actual choice implemented. However, iodine. when studies did not report the actual choice, we used the patien ts’ preferred option as a surrogate measure. Actual choices or pref er- ences were reported as the percentage of individuals actually i m- Satisfaction plementing or stating a preference for the most intensive or m ost invasive option. Nineteen included studies (18.1%) measured satisfaction as it re- In summary, patient decision aids decreased the number of pa- lates to the choice and the preparation for and the process of deci - tients choosing elective surgical procedures (excluding prophy- sion making. When possible, we standardized the scores to a 0 to lactic mastectomy) and PSA testing in multiple studies. Singl e ion. 100 point scale, with higher scores reflecting greater satisfact le studies showed that decision aids increased the number of peop . Nineteen studies (18.1%) measured satisfaction with the choice for choosing hepatitis B vaccination, psycho-educational therapies sion Of these 19 studies, 4 reported that people exposed to the deci s- mental health conditions, and medication for cardiovascular di aid had higher satisfaction with their choice compared to usual ease prevention. In contrast, decision aids decreased the rate of fer- care, and the other 15 reported no statistically significant dif cardiac stress testing, the number of embryos being transpla nted, ences ( ; ; Montgomery Laupacis 2006 ; Chabrera 2015 Heller 2008 21 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

25 and the rate of antibiotic use for upper respiratory infectio pooled data ( Table 14 ). Frosch reported a reduction in screening ns. The effect on patients’ choice in other situations was more variabl rates and the other two reported no difference. e. There were mixed results for the choice of colon cancer screening, y, breast genetic testing, prenatal testing, anti-thrombosis therap e- screening, and diabetes medications. There was no difference b Choice for colon cancer screening en- tween groups for choices about natural health products, hypert - sion therapy, breast cancer chemotherapy, schizophraenia medi Of 10 studies (9.5%) on colon cancer screening, 3 reported sta- r cation, immunotherapy for multiple sclerosis, vaccines (for flu o tistically significant differences in choices, and 7 showed no dif - osteo- measles, mumps, rubella), diabetes screening, birth control, ference. Two studies reported that compared to usual care, the porosis treatment, chemotherapy for advanced cancer, chemopre- decision aid significantly increased the screening rates by 64% roce- ventive medications, use of blood transfusions, childbirth p ; Pignone 2000 ). The other study reported Ruffin 2007 and 70% ( iodine dures, treatment of prolapsed pelvic organs, or radioactive a statistically significant reduction of 21% for screening ( Smith treatment for thyroid cancer. 2010 ). There was an increase in screening rates in five studies, t by 6% to 39%, but the difference was not statistically significan ( Lewis 2010 ; Miller 2011 Wolf ; Steckelberg 2011 ; Schroy 2011 ; 2000 ). In two studies ( Dolan 2002 ; Trevena 2008 ), there was a Choice for major elective surgery 73% and 4% decrease in screening rates that was not statisticall y significant. The pooled RR was 1.12 (95% CI 0.95 to 1.31, 10 c- Eighteen studies (17.1%) focused on choices regarding major ele Analysis 8.2 .2). studies; tive surgery ( ). Analysis 8.1 Using intention-to-treat analysis, there was a non-significant re- duction in the number of patients choosing major elective surge ry in the group receiving the decision aid compared to usual care (RR Choice for cancer genetic screening 0.86; 95% CI 0.75 to 1.00, 18 studies; Analysis 8.1 Schwartz .2). reported a statistically significant uptake of prophylactic 2009a Four studies reported preferences or uptake rates for breast cancer mastectomy for women who are BRCA1/2 gene carriers (114%). genetic screening (3.8%). The decision aid did not significantly was a And after removing this study from the pooled results, there affect preferences for breast cancer genetic screening when com- oos- statistically significant reduction in the number of patients ch pared to usual care. The pooled RR was 0.99 (95% CI 0.71 to ing major elective surgery (RR 0.84 95% CI 0.73 to 0.97; 17 1.38, 3 studies; Analysis 8.2 .3). One study reported an increase in studies; Analysis 8.1 .3). screening rates by 14% ( Lerman 1997 ), a second study reported Four other studies showed statistically significant reduction s in Green 2001 ), and a third study reported a de- an increase of 18% ( surgery rates: − 29% for cardiac revascularization and bariatric crease of 29% ( Miller 2005 ). Schwartz 2001 reported that women surgery ( ), − ; Arterburn 2011 Morgan 2000 33% for orchiectomy cer exposed to the decision aid who were at higher risk of breast can ( 74% for mastectomy ( Whelan 2004 ). The Auvinen 2004 − ), and increased their intention to obtain genetic testing, while t hose at or other 15 studies showed no difference between the decision aid average risk decreased their intention ( ). Table 14 usual care groups. Choice for breast screening Choice for other elective surgery There were lower mammography screening rates among women care Two studies evaluated the effect of decision aids versus usual Mathieu 2010 ), but no between-group aged 38 to 45 years of age ( on other elective surgical decisions. Decision aids did not sign ifi- difference in women aged 70 or older who were exposed to a cantly influence surgical abortion rates in Wong 2006 or feeding decision aid versus usual care ( ). Mathieu 2007 ; Table 14 tube insertions in Hanson 2011 ( Table 14 ). Choice for prenatal screening Choice for prostate-specific antigen screening ning, In all four studies focusing on decisions around prenatal scree prenatal testing rates were not affected by a decision aid compa red The effects of decision aids on prostate-specific antigen (PSA) Kuppermann 2014 ; Bekker 2004 ; ; Bjorklund 2012 to usual care ( - screening decisions were variable in 13 studies (12.4%) that com Nagle 2008 ). Meta-analysis included two studies, showing no ef- pared decision aids to usual care. The pooled RR for 10 studies fect (RR 0.99, 95% CI 0.91 to 1.09, 2 studies; Bjorklund 2012 , Frosch 2008a .1); Analysis 8.2 ; was 0.88 (95% CI 0.80 to 0.98; , and Lepore 2012 Williams 2013 could not be included in the Kuppermann 2014 ; Analysis 8.2 .4). 22 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

26 Choice for stress test for chest pain Choice for chemotherapy for cancer There was no statistically significant difference in the rates o f che- Compared to usual care, adults presenting with chest pain in th e emergency department who received the decision aid had signifi- motherapy for adults with advanced colorectal cancer (77% ver- ). Table 14 ; Leighl 2011 Hess 2012 ; also found no Whelan 2003 sus 71%) ( cantly lower rates of stress testing (58% versus 77%) ( significant effect on preferences for adjuvant chemotherapy ver sus Table 14 ). no chemotherapy for early stage breast cancer. Choice for screening for diabetes Choice for diabetes treatment with new medications Compared to usual care, there was no difference in diabetes scre en- l care Four studies evaluated patient decision aids compared to usua ing rates in Marteau 2010 or preferences for screening in Mann E on decisions about starting new medications for diabetes ( Mann D 2010 Table 14 ). in adults exposed to a decision aid ( ; Mathers 2012 ; Mullan 2009 ; Weymiller 2007 ). Although 2010 there was no statistically significant difference between grou ps for crease in individual studies, pooled results indicated a significant in starting new medications (RR 1.65, 95% CI 1.06 to 2.56; Analysis Choice to take antibiotics for upper respiratory infection 8.3 ). Compared to usual care, using a decision aid in the consultation y infec- decreased prescriptions for antibiotics for upper respirator , although this difference was not statistically Legare 2012 tions in significant in Table 14 ). ( Legare 2011 Choice to take hypertension medication Montgomery 2003 found no significant effect for decision aids on over usual care on the initiation of medication for hypertensi ( Table 14 ). Choice for atrial fibrillation treatment Three studies evaluated the effect of a decision aid on the use ual care of anti-thrombotic therapy for atrial fibrillation versus us Table 14 ). One study demonstrated a non-significant reduction ( Choice for menopausal symptom treatment in warfarin use of 25% ( ). The second study Man-Son-Hing 1999 ), Murray 2001b In a study comparing a decision aid to usual care ( at was evaluated the proportions of patients choosing the option th there was a non-significant decrease of 8% in hormone therapy nificant appropriate relative to their level of risk, and found no sig ( ). Preferences for natural health products in women ex- Table 14 McAlister 2005 ). Thomson 2007 difference between the groups ( periencing menopausal symptoms were no different for women ce reported that patients in the usual care group (guided by practi exposed to the decision aid compared to women exposed to the recommendations) were much more likely to start warfarin (15/ usual education materials ( ). Legare 2008a 16; 93.8%) compared to the decision aid group (4/16; 25%; RR 0.27; 95% CI: 011 to 0.63). Choice for multiple sclerosis immunotherapy Choice to take breast cancer prevention medication reported no difference in the uptake of immunother- Kasper 2008 - apy in people with multiple sclerosis who were exposed to a deci There was no difference in medication use among women at risk sion aid compared to usual care based on practice guidelines ( Table of breast cancer who were exposed to the decision aid versus usua l ). 14 care ( ). Fagerlin 2011 Table 14 ; Choice to take osteoporosis treatment Choice for cardiovascular disease prevention There was no difference in prescriptions for bisphosphonates for r- There was an increase in patient preferences for any effective ca osteoporosis treatment ( LeBlanc 2015; Table 14 ). Montori 2011 diovascular disease risk-reducing strategy (including medicati on) found no significant effect of decision aids over usual care on the when using a decision aid versus usual care (63% versus 42%) uptake of medication for osteoporosis treatment. ( Sheridan 2011 ). Table 14 ; 23 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

27 Mental health ous There was no difference in the uptake of preoperative autolog blood donation when a decision aid was compared to usual care found no difference in prescription rates for an- Hamann 2006 ). Laupacis 2006 ( tipsychotic medications but reported a statistically significa nt in- crease in the uptake in psycho-education (P = 0.003) in people with schizophraenia exposed to the decision aid compared to usu al care ( ). reported that a higher proportion of Table 14 Mott 2014 Lung transplant referral ess participants in the decision aid group with post-traumatic str There was no difference in referral rates for consideration of lung disorder completed psychotherapy sessions (4 of 9) compared to transplant in people with advanced cystic fibrosis exposed to a usual care (1 of 11). decision aid versus usual care ( Vandemheen 2009 ). Obstetrical choices Pelvic organ prolapse treatment There was no difference in treatment rates for prolapsed pelv ic organs ( ). Brazell 2014 Childbirth procedures Three studies focused on childbirth issues, using a decision ai d compared to usual care. There was no difference in preference for vaginal birth in Shorten 2005 or actual vaginal mode of delivery Thyroid cancer radioactive iodine treatment in Montgomery 2007 following a previous cesarean section. An- dine There was no difference in the rates of adjuvant radioactive io - other study found no difference in actual choice to undergo exter treatment for thyroid cancer ( ). Sawka 2012 Nassar nal cephalic version for women with breech presentation ( 2007 ). Adherence (continuance/compliance) with chosen option Of 105 studies, 16 (15.2%) measured adherence using various Table 15 ). approaches ( Birth control approaches Based on the measurement framework by , we Trenaman 2016 There was no difference in the birth control methods chosen for grouped adherence according to adherence to the baseline choice those in the decision aid versus usual care groups ( ). Langston 2010 er- and adherence to the treatment. Six studies measured only adh Langston 2010 Lepore ; Legare 2012 ; ence to the baseline choice ( 2012 Man-Son-Hing 1999 ; ), 6 Trevena 2008 ; Mathers 2012 ; studies measured only adherence to treatment ( Loh 2007 ; Mann Embryo transplantation D 2010 Mott 2014 ; ; ), Sheridan 2011 ; Oakley 2006 Mullan 2009 ; Compared to usual care, those in the decision aid group were LeBlanc 2015 ; Montgomery 2003 ; and 4 studies measured both ( (43% significantlymore likelytochoose asingle embryotransplant Montori 2011 Weymiller 2007 ; ). versus 32%) ( ). Van Peperstraten 2010 For the 10 studies that measured adherence to choice, two studie s reported that patients exposed to decision aids had higher ad her- ence compared to usual care ( Mathers 2012 ), and Montori 2011 ; Vaccines 8 reported no difference between groups. For example, Mathers Compared to usual care, there was a non-significant increase in 2012 asked participants, 6 months after their decision, whether or intentions to get the flu vaccine in those exposed to the decision n for not they had changed their initial choice about starting insuli Chambers 2012 ), a statistically significant aid (46% versus 27%) ( type II diabetes (decision aid 68.1% versus 56.3% usual care; P increase in uptake of hepatitis B vaccination with decision aids = 0.041). Montori used pharmacy records to determine if partici- ( ), and no difference in uptake of measles, mumps, Clancy 1988 pants who chose bisphosphonates actually took their medicatio n rubella vaccine in infants ( Shourie 2013 ). on more than 80% of the days for which it was prescribed (100% decision aid versus 74% usual care; P = 0.009). For the 10 studies that measured adherence to treatment, 2 stu dies Other choices reported that patients exposed to decision aids had higher ad her- ence compared to usual care ( Mott 2014 ; Sheridan 2011 ), 1 study reported that patients exposed to decision aids had lower adh er- ence ( Mullan 2009 ), and 7 reported no difference. Mott reported Blood transfusions the percentage of participants at four months who engaged in ni ne 24 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

28 Other health outcomes or more psychotherapy sessions (4 of 4 decision aid group partic- ipants versus 1 of 5 usual care). Sheridan measured the percent- Seven studies (6.7%) reported on other health outcomes ( Table ere age of participants who, 3 months after initiating therapy, w ), glycated Knops 2014 ; ), including death ( Auvinen 2004 18 continuing (59% decision aid versus 34% usual care; P < 0.01). haemoglobin ( Mathers 2012 ), angina ( Morgan 2000 ), stroke Mullan used pharmacy records to determine the days covered by ( Thomson 2007 ), successful pregnancy ( Van Peperstraten 2010 ), medication use (97.5% decision aid versus 100% usual care). and pain ( ). There were no statistically significant Vuorma 2003 differences between groups. Health outcomes Preference-linked health outcomes None of the 105 studies measured preference-linked health out - comes - that is, whether the patients experienced the outcomes they preferred and avoided the outcomes they wanted to avoid. General health outcomes Eleven studies (10.5%) compared a decision aid to usual care in Anxiety terms of general health outcomes ( Table 16 ). Ten of these used either the previously validated Medical Outcomes Study 36-ite m Of 105 studies, 31 (29.5%) measured anxiety, with 24 using the Short-Form Health Survey (SF-36) or the 12-item Short-Form Spielberger previously validated State Trait Anxiety Inventory ( Health Survey (SF-12) ( ), while used Vuorma 2003 Stewart 1992 1970 ), 2 using the anxiety subscale of the Hospital Anxiety and the RAND-36 ( , there were no Table 16 ). As shown in Hays 1993 Depression Scale ( Knops 2014 ; Lam 2013 ), 2 using questions significant differences for mental health function or social fun c- about worry ( Fraenkel 2012 ; Smith 2010 ), 2 measuring intrusive Barry 1997 ), general tion in any of the seven studies. In one study ( Lewis 2010 ; McCaffery2010 ), and1usingasingle ques- thoughts( t- health andphysical functionoutcome scoreswere significantlybe ; see Table 19 Johnson 2006 tion on a seven-point Likert scale ( ). ter in the decision aid group compared to usual care for men con- hin Of 18 studies that used the State Trait Anxiety inventory wit sidering treatments for benign prostatic disease. Of the tw o stud- n 1 month postintervention, 2 (11.1%) reported that the decisio ies evaluating the effect of a decision aid for women considerin g aid group had significantly lower anxiety scores for people con- treatment for abnormal uterine bleeding, found Kennedy 2002 Montgomery sidering birthing options after a previous caesarean ( on, a statistically significant improvement in role physical functi 2007 ) and for women considering options for the treatment of and found a statistically significant improvement in Vuorma 2003 ). None of the studies demonstrated Protheroe 2007 menorrhagia ( emotional role functioning for women. one significant differences in effects on people’s state anxiety at In two studies measuring health utilities using the Euroqol EQ- month (2 studies), three months (6 studies), six months (4 stud- ), there was no difference be- Murray 2001a ; Murray 2001b 5D ( ies), or one year (2 studies). There was no significant difference tween the decision aid and usual care groups. There was also no ety. between groups for the other instruments that measured anxi between-group difference in the LeBlanc 2015 study, which used the Euroqol 5D health thermometer. Depression Of 105 studies, 6 (5.7%) measured the effect of decision aids on depression using various instruments ( ). None of the stud- Table 20 Condition-specific health outcomes ies reported a statistically significant difference between gr oups for decisions about cancer treatment ( ), Davison 1997 ; Whelan 2004 n- Seven studies (6.7%) used various measures to assess conditio depression ( Nagle 2008 ), prenatal genetic testing ( Loh 2007 ), or specific health outcomes ( Table 17 ). Outcomes included uri- for women considering the number of embryos to transplant ( Van nary symptoms ( Barry 1997 ; Murray 2001a ), angina ( Bernstein Peperstraten 2010 ). At 10 months’ postintervention, there were 1998 ), Leighl 2011 ), functional assessment of cancer therapy ( r lower levels of depression in women deciding about breast cance menopausal symptoms ( ), and menstrual symp- Murray 2001b e surgery who were exposed to the patient decision aid versus th toms ( ). Five studies found no sig- Protheroe 2007 ; Vuorma 2003 usual care, but no differences at 1 week, 1 month, or 4 months nificant effects on condition-specific health outcomes ( Bernstein Lam 2013 ). postintervention ( Vuorma 2003 ; ; ). ; ; Murray 2001b Murray 2001a Leighl 2011 1998 reported significantly higher menorrhagia-related Protheroe 2007 quality of life scores in women exposed to the decision aid com- Regret pared to usual care. showed an improvement in uri- Barry 1997 t nary symptoms in favour of the decision aid group, but it was no Of 105studies,7(6.7%)measuredthe effectof decisionaidsond e- statistically significant. cision regret, using the five-item Decisional Regret scale ( Brehaut 25 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

29 Consultation length see 2003; ). At 4 and 10 months postintervention, women Table 21 with breast cancer who were considering surgery and used a deci- Of 105 studies, 10 (9.5%) evaluated the effect of a decision aid sion aid reported lower regret scores compared to women receiv- Table 23 ). compared to usual care on consultation length (see Lam 2013 ). There was no statistically significant ing usual care ( The median consultation length was 24 minutes (range 3.8 to between-group difference in the other six studies. 68.3) for patient decision aid compared to 21 minutes (range 4.2 to 65.7) for usual care. The difference was 2.6 minutes longer (7.5% increase) than usual care consultations (range 0.4 minutes shorter to 23 minutes longer). The length of consultation was Confidence significantlylongerforthe patientdecisionaidgroupintwos tudies Of 105 studies, 8 (7.8%) measured the effect of decision aids on ( ; Bekker 2004 ), and eight studies reported no Thomson 2007 confidence levels (see ). Four of these studies used the De- Table 22 difference. reportedthatconsultationsaboutprenatal Bekker2004 cisional Self-efficacy Scale ( Fraenkel Allen 2010 ; Arterburn 2011 ; diagnostic testing were 5.9 minutes longer, and Thomson 2007 2007 Smith 2010 ). Four studies reported a statistically significant ; reported consultations about treatment for atrial fibrillat ion were improvement in confidence or self-efficacy with decision making 23 minutes longer when using a computerized decision aid with Chambers in the decision aid compared to the usual care groups ( standard gamble method within the consultation. 2012 Fraenkel 2007 McBride 2002 ), and the ; ; ; Gattellari 2003 other studies reported no difference between groups. Litigation rates None of the 105 studies examined the effect of decision aids on Healthcare system effects litigation. Adverse events Cost and resource use n, There were no adverse effects on health outcomes or satisfactio and no other adverse events reported. Of eight studies (7.6%) examining cost and resource use, one conducted a cost-effectiveness analysis ( Kennedy 2002 ), five eval- uated the effect of decision aids compared to usual care on to- Subgroup analysis - in preparation for versus during Van Montgomery 2007 ; Murray 2001a ; Murray 2001b ; tal costs ( the consultation Peperstraten 2010 Vuorma 2003 ; ), and two measured resource use Of 105 studies, 89 (84.8%) primarily evaluated the patient de ci- ( ) (see Table 23 ). ; Legare 2012 Thomson 2007 ta- sion aid when used by the patient in preparation for the consul The cost-effectiveness analysis ( ) was conducted Kennedy 2002 tion, and 16 (15.2%) primarily evaluated the patient decision aid from the healthcare system perspective, using USD values from when used within the consultation. The patient decision aids u sed 1999 to 2000 and calculating costs over two years. The decision during the consultation focused on prenatal screening ( Bekker aid with nurse coaching demonstrated the lowest mean cost (USD 2004 ); cardiac stress testing ( Hess 2012 ); dental surgery ( Johnson 1566) compared to decision aid alone (USD 2026) or usual care 2006 Kupke 2013 ); antibioticsforup- ); restorationof tooth decay( (USD 2751). per respiratory infection ( ); medication Legare 2012 ; Legare 2011 statis- Of the five studies that evaluated total costs, two reported no ; Loh 2007 ), diabetes ( Mann D 2010 Mullan use for depression ( tically significant difference in the patient decision aid compar ed 2009 LeBlanc 2015 Montori ), osteoporosis ( ; Weymiller 2007 ; Vuorma 2003 ; Montgomery 2007 ). Two studies re- to usual care ( 2011 Ozanne 2007 ), preventionof breastcancer( ), andatrial fibril- ported higher costs for the patient decision aid group when in- lation ( Whelan 2004 ); surgery for breast cancer ( Thomson 2007 ); cluding the cost of the interactive video disc equipment (USD 216 and chemotherapy for breast cancer ( ). Whelan 2003 at 1999 prices) and no statistically significant difference betw een Murray 2001a ; ). Murray 2001b groups when removing this cost ( The fifth study reported that the mean total savings in the deci - Knowledge sion aid group versus usual care was EUR 169.75 per couple ( Van Peperstraten 2010 ). When considered separately by subgroups, there was no differ- Legare For healthcare resource use in upper respiratory infection, ence between knowledge scores for those exposed to the decision 2012 reported no difference in the rates of repeat consultations in aid in preparation for the consultation compared to those used for the same reason, and : MD 13.77% versus 10.57%, Analysis 1.2 Thomson 2007 reported no difference the consultation itself ( 2 ths in the rates of general clinician consultations in the three mon test for subgroup difference P = 0.31, I : 3%). Weymiller 2007 following the intervention. Both studies used the patient d ecision reported a higher mean difference when the decision aid was ad- aid in the consultation. ministered during the consultation but not if it was administ ered 26 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

30 by research staff in preparation for the consultation. For the of pre-consultation decision aid delivery ( Fraenkel 2012 ; Hanson stud- 2011 ded ies evaluating decision aids used in the consultation not inclu Lepore 2012 ) reported that, compared to ; Sheridan 2011 ; - in the pooled outcome, two showed a statistically significant im s of those in the usual care group, significantly higher proportion ), and two ; LeBlanc 2015 Ozanne 2007 provement in knowledge ( participants exposed to the patient decision aid in preparati on for showed no difference ( ). Thomson 2007 ; Mann D 2010 the consultation reported that they discussed the decision wit h - their clinician, and the fifth study showed no between-group dif ). Sheridan 2006 ference ( Accurate risk perceptions When analyzing pre-consultation and in-consultation decision Participation in decision making aids further, accurate risk perceptions were not different bet ween lta- studies that used the decision aid in preparation for the consu There were too few studies on decision aids used during the con- tion and those where the intervention occurred during the con- sultation to interpret findings from the subgroup analysis ( Analysis sultation ( Analysis 2.2 : RR 2.25 versus RR 1.79, test for subgroup 5.2 Analysis 5.3 ; ). 2 : 0%). The only study evaluating a de- differences: P = 0.33, I cision aid within the consultation that was not included in the meta-analysis, Weymiller 2007 , reported a higher proportion with Length of the consultation accurate risk perception when the decision aid was administered Due to variation in the reporting of data for this outcome, we ps during the consultation, but found no difference between grou g on were unable to investigate the effects of intervention timin sul- when administered by research staff in preparation for the con the length of consultation. Of seven studies that evaluated d e- tation. Bekker 2004 LeBlanc ; cision aids used within the consultation ( 2015 ; ; Loh 2007 ; Ozanne 2007 ; Thomson 2007 ; Weymiller 2007 ), two reported that the length of the consultation Whelan 2003 Decisional conflict uninformed subscale was significantly longer for the patient decision aid group ( Bekker Too few studies measured the uninformed subscale in those ex- 2004 Thomson 2007 ). There was no difference for the other ; posed to decision aid within the consultation to be able to com- in studies. The three studies that evaluated decision aids used pare with those who used decision aids in preparation for the preparation for the consultation reported no between group d if- consultation. Weymiller 2007 reported that participants felt less Bozic 2013 ; ; Krist 2007 ference in the length of the consultation ( uninformed when the decision aid was administered during the ). Vodermaier 2009 consultation, but not if it was administered by research staff in preparation for the consultation. Other outcomes For values-choice congruence and proportion undecided, none of Decisional conflict unclear values subscale the studies of patient decision aids used during the consultat ion Too few studies measured the unclear values subscale in those e x- of measured these outcomes. For satisfaction, there were a range posed to decision aid within the consultation to be able to compa re ults different approaches to measuring this outcome with mixed res with those who used decision aids in preparation for the consul - and too few studies to make any descriptive comparisons. For reported that participants felt less unclear Weymiller 2007 tation. choice, there were too few studies to conduct a subgroup analysis about values when the decision aid was administered during th e of pooled comparisons. in consultation, but not if it was administered by research staff preparation for the consultation. Post hoc analysis Patient-clinician communication ere Due to variation in the reporting of data for this outcome, we w Effects of study quality unable to investigate the effect of intervention timing on th e vari- of low To examine the potential bias arising from including studies a- ation in the effect on communication. Five studies evaluated a p k methodological quality, we excluded 12 studies with a high ris h the tient decision aid primarily used within the consultation wit of bias for any of the seven risk of bias criteria from the analy sis a- clinician, and five evaluated a patient decision aid used in prep ( ration for the consultation (see Chambers 2012 ; Brazell 2014 ; Auvinen 2004 Table 8 ). All five studies that used ; Clancy 1988 ; higher the decision aid during consultations reported statistically ; ; Kupke 2013 Hamann 2006 LeBlanc Knops 2014 ; Krist 2007 ; mean OPTION scores in the patient decision aid group compared 2015 ; Man-Son-Hing 1999 ; Lewis 2010 ; Figure ; see Mott 2014 to usual care ( 3 ; ; LeBlanc 2015 ; Montori 2011 ; Mullan Hess 2012 ). Overall, the results remained the same ( Table 24 ; Analysis 1.3 2009 ; Weymiller 2007 ). Four of five studies assessing the effects Analysis 2.3 ; Analysis 3.5 ; Analysis 4.4 ). 27 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

31 (7.5%) in the decision aid group compared to usual care consulta- Heterogeneity tions.There continue to be too few studies to determine the ef fects When comparing patient decision aids to usual care, there was Trenaman 2014 ). Although of decision aids on costs/resource use ( statistically significant heterogeneity in five of six of the IP DAS aids, there may be additional costs involved in delivering decision effectiveness criteria: knowledge scores, accurate risk percept ions, an independent review of decision aid studies with economic ou t- congruence between values and choice; feeling uninformed, and comes concluded that “this was likely to be small relative to the feeling unclear regarding personal values. There was no stat istically benefit to patients in terms of improved decision quality when significant heterogeneity for participation in decision makin g. It effective decision aids are used” ( NCGC/NICE 2012 ). Given the should be noted that the heterogeneity of the effect was not ma n- mine the variability in measurement strategies, it difficult to deter te ifested in its direction but only in its size. For the 2009 upda effect of patient decision aids on adherence to the chosen option ( ), we explored the potential factors contributing O’Connor 2009b or treatment. Table 25 ). Overall, regardless of the subgroup to heterogeneity ( ids New for this update, we analyzed the pooled data for decision a all analyses conducted, scores for outcomes were similar to the over sion used in preparation for the consultation separately from deci effect, as indicated by overlapping confidence intervals. aids used in the consultation, and we found that there were sim - ilar improvements in knowledge, accurate risk perceptions, an d patient-clinician communication. D I S C U S S I O N Summary of main results Overall completeness and applicability of evidence In this updated review, we added 18 new studies for a total of 1 05 studies comparing patient decision aids to usual care. This up- mple date also removed 28 studies that compared detailed versus si Main effects of decision aids e. patient decision aids that were included in the previous updat Based on the GRADE assessment ( The largest and most consistent benefits of decision aids, rela tive Summary of findings for the d to usual care, are better knowledge of options and outcomes, an main comparison ), there is high-quality evidence that compared more accurate perceptions of outcome probabilities. These ob- to usual care, decision aids improve people’s knowledge regar ding servations are clinically important because the usual care grou ps’ options and reduce the decisional conflict stemming from feel- ere scores for knowledge and perception of outcome probabilities w ing uninformed and unclear about their personal values. Ther e is lower than the intervention groups’; both knowledge and per cep- moderate-quality evidence that decision aids stimulate peopl e to tion of outcome probabilities are important for ensuring inf ormed take a more active role in decision making and increase the accu- ’ may decisionmaking.These effectssuggestthatcurrent’usual care that racy of their risk perceptions. There is lower-quality evidence not be good enough when informing people about these complex, decision aids improve congruence between the chosen option and values-sensitive decisions. People need to comprehend the opt ions iffer- personal values. This outcome is measured using a variety of d and outcome probabilities in order to consider and communicate ent approaches, and the evidence could be strengthened by more to their clinicians the personal value they place on the benefits ver- d the standardized measurement. Moreover, decision aids decrease sus the harms. Likewise, pooling results from additional st udies in proportion of people remaining undecided. ed this update shows a significant increase in informed values-bas Although not a primary outcome of the review, the effect of de- choice when decision aids were compared to usual care, and the cision aids on patients’ choosing particular options continues to ies that results appear to be similar across subgroup analyses of stud ec- be variable. The numbers of patients choosing to have major el used the same composite measure. ve tive surgery continues to decrease in favour of more conservati Decision aids also help people feel more comfortable with thei r options, except when the baseline rates are low (e.g. surgery f or choices than usual care. This is revealed by the reduced scores for benign prostate hyperplasia, prophylactic mastectomy for wo men s. overall decisional conflict and for the decisional conflict subscale who are carriers of the BRCA gene). The numbers of men choos- ut People who use decision aids generally feel more informed abo ing prostate-specific antigen (PSA) testing were fewer after ex po- options and clearer regarding their personal values. sure to decision aids. cts Decision aids do no better than usual care in terms of their effe Compared to usual care strategies, decision aids improve indi vid- on people’s satisfaction with decision making or health outcom es r- uals’ perception of involvement in decision making. This obse ife. such as general quality of life or condition-specific quality of l s Stan- vation suggests that the International Patient Decision Aid es, However, no studies measured preference-linked health outcom dards criterion of helping patients participate ’in ways that they ce in nor were adverse events reported. There was also no differen prefer’ needs to be assessed after a patient has adequate inf orma- anxiety. For length of consultation, eight studies found no d iffer- tion about what involvement means using interventions such a s ence, while two studies found a median increase of 2.6 minutes patient decision aids. People may have a mistaken preference f or 28 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

32 Limited effects of decision aids passivity because they believe that the best choice relies on th e ?) expertise of the clinician (which option is medically reasonable The limited effects of decision aids on reported satisfaction w ith s rather than understand the importance of their own preference the decision-making process and with the actual choice made may for outcomes of options (which outcomes matter most to me?). n. indicate that decision aids have a limited effect on satisfactio f- Evidence continues to build that decision aids have a positive e . The null effects may also be due to measurement insensitivity fect on the patient-clinician consultation (in 9 of the 10 studies lready This is especially likely when satisfaction with usual care is a that assessed this effect). Of the studies that measured patie nt- quite high (e.g. ceiling effects) and when choices are inherently clinician communication, five involved using decision aids withi n difficult to make because of competing benefits and harms. Fur- the consultation and five in preparation for the consultation. At thermore, once the decision is made, people may find it psycho- her the same time, evidence on length of consultation indicates eit logically more comforting to say that they are satisfied rather than ations no difference (8 studies) or slightly longer (2 studies) consult ). Gruppen 1994 entertain doubts about what they have chosen ( in the decision aid group compared to usual care consultations. There is a need to establish the ’essential ingredients’ in d ecision However, few studies have reported on the impact of the context aids and to identify the people who are most likely to benefit f rom in which the patient decision aids are used. A previous subgrou p sier them. As the body of available research grows, it will become ea reatment analysis of 29 studies evaluating patient decision aids for t and more important to assess the usefulness of different comp o- decisions reported greater improvement in knowledge scores (P nents of decision support for different clinical contexts, decis ion = 0.03) when the patient decision aid was evaluated within the eci- problems, and groups of people. For example, an analysis of d clinical pathway of care, compared to when patients volunteered i- sion aids used in higher versus lower socioeconomic groups ind to participate in the study independent of their clinician ( Brown a- cated greater improvements for those of lower socioeconomic st 2015 ). tus( ). Recently, the IPDAS Collaborationcompleted Durand 2014 a set of evidence reviews underlying the IPDAS checklist ( IPDAS 2013 ), proposing criteria for defining the intervention as a pa- Joseph-Williams tient decision aid and minimal certifying criteria ( Variable effects of decision aids ). These are being used to inform the certification of patient 2013 decision aids in the USA, England, and Norway. There may be several reasons for the variable effect of decisio n aids ealth It is not surprising that decision aids had limited effects on h d on the outcome of choices. First, most studies were under-powere - outcomes. One reason for using a decision aid is that there is of to detect important differences in the outcome of choices. Second, ten no option with a clear health outcome advantage. For exam- ecific notenough isknownaboutbaseline ratesforoptimal use of sp ple, when men with localized prostate cancer consider active trea t- at options. Third, in the studies reporting the outcome ’choices’ ng ment options, their health outcomes can be different, dependi baseline and postdecision aid, some options may have been und er- on whether they choose surgery with higher risks of impotence o r usedandothersover-used,relative tothe choicesindividuals would radiation therapy with higher risks of longer term bowel irr itation. s, make if they were more fully informed. Under these circumstance Therefore, if health outcomes are used in future investigati ons of one could expect to observe directional effects on choices once ut- decision aids in situations in which there is clearly no health o people become better informed and more involved in decision i- come advantage, the key question to pose is: do patients exper making. ence the health outcomes they prefer and avoid the outcomes to Relatively under-used options at baseline were prostate sur gery which they are averse? y for for benign prostatic hyperplasia and prophylactic mastectom ich More recently, decision aids are being used in situations in wh breast cancer gene carriers. In this prostate-related example, there there may be a longer-term health advantage, for example, in p re- was a shortage of urologists and low referral rates for benig n pro- d/ ventive decisions about the management of type II diabetes an static hyperplasia, whereas the breast-related example refl ects the or hypertension, when the longer-term health outcome may be to growing number of women who test gene positive and become avoid stroke ( ; Mann D 2010 Montgomery 2003 ; ; Mathers 2012 aware of their options for preventing breast cancer. Hence, und er- Mullan 2009 ). Interestingly, the pooled results Weymiller 2007 ; use of an option may be corrected with exposure to a decision aid. ation showed a statistically significant increase in medication initi In the other surgical decision aid studies, there were higher n um- when participants were exposed to the decision aid compared to c bers of people choosing surgery in the control group (e.g. cardia usual care. revascularization, back surgery, hysterectomy, orchiectomy, m as- tectomy). The procedure may have been chosen due to people’s inflated perceptions of the probabilities of benefits, lack of a ppre- Unknown effects of decision aids al- ciation of the probabilities of harms, and lack of awareness of The effect of patient decision aids on adherence to the chosen ). Exposure to the decision aid reduced Hoffman 2015 ternatives ( option is an area of uncertainty. The adherence results are dif ficult e the number of people choosing elective surgery in favour of mor to interpret due to incomplete data, primarily self-reported data, conservative alternatives. 29 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

33 ve may have been bias arising from a failure to report all negati varying length of follow-ups, and small sample sizes. Moreov er, findings. studies reporting this outcome such as had Man-Son-Hing 1999 We rated the six primary outcomes in the ’Summary of findings’ very little variation in choice (over 90% of long-term aspirin us ers decided to stay on aspirin). When examining adherence, it would - table using GRADE and assessed outcomes as high quality (knowl be important to do so in the early phase, when presumably the edge, feeling uninformed, feeling unclear values), moderate qual- issue is actually decisional in nature (e.g. filling the prescrip tion, ity (accurate risk perception, clinician-controlled role in decisio n r than picking up the prescription, refilling the prescription) rathe making), and low quality (values-choice congruence). For values- involving the management of side effects and in a manner that choice congruence, the GRADE rating was downgraded for lack separates those choosing to change versus those remaining wit h of consistency, directness, and precision. More specifically, con- the status quo. gruence was measured using various approaches, as there is no an Despite the positive effects of decision aids on patient-clinici Munro 2016 ). Several of gold standard measurement approach ( communication, some authors are concerned about the potential the outcomes demonstrated statistically significant levels of het- negative influence that decision aids may have on the relationa l o- erogeneity. For the outcome of knowledge, for example, heter aspects of the decision-making process; this concern highlights t he geneity would be expected, given that the knowledge tests the m- ed as need for further evaluation when decision aids are implement selves were not standardized. However, we did not downgrade part of the routine process of care ( un- the ratings for knowledge, feeling uninformed, and feeling Charles 2010 ; LeBlanc 2010 ). clear values based on heterogeneity given the consistent dire ction In the context of decision aid use, cost-effectiveness and health d in tle utilities are other secondary outcome measures about which lit of findings across studies. Moreover, the heterogeneity foun se the various outcomes reflects differences across clinically diver is known and further evaluation is required ( Trenaman 2014 ). studies; therefore, the pooled effect size and confidence inter vals We also need to establish ways of measuring preference-linked t should be interpreted as a range across conditions, which may no health outcomes to better determine the effect on quality of li fe. be applicable to a specific condition. It is unlikely that we will observe the effect of decision aids o n litigation rates in studies of decision aids, given the time d elay to litigation and the rarity of this type of event. There do not a ppear to be any adverse events from using decision aids, but this coul d be more clearly examined in future studies. In fact, a mock trial that used a patient decision aid for prostate-specific antigen t esting found that the majority of jurors (94%) would indicate that the Potential biases in the review process standard of care had been met ( Barry 2008 ). A recent systematic on aids The strength of this systematic review is that patient decisi review concluded that there was insufficient evidence to determi ne improve several key primary outcomes across a wide variety of p op- if patient decision aids could reduce medical malpractice litigat ion ulations and decision contexts. The potential biases in the re view ( ). Durand 2014 process are due to limitations associated with having inadequ ate power to detect potentially important differences in effectiv eness ive between subgroups, to differentiate between the most effect elements within the patient decision aid, and to investigate any Quality of the evidence ons differences associated with the type of comparison interventi tically used in studies. Several of the outcomes demonstrated statis Risk of bias ratings reveal between-study variability. We ra ted few ically significant heterogeneity. This reflects differences across clin studies as being at low risk of bias for blinding of participan ts nce diverse studies; therefore, the pooled effect size and confide . and personnel and most studies as being at unclear risk of bias which intervals should be interpreted as a range across conditions, Likewise, the majority of studies were rated as being at uncle ar may not be applicable to a specific condition. In the Gentles 2013 risk of bias for selective reporting. When we conducted a post rogene- subgroup analysis exploring three potential sources of hete hoc analysis that involved removing studies at high risk of b ias ity ity (e.g. type of control intervention, decision aid IPDAS qual from the meta-analysis, there was no effect on the results. The pants’ score, participants’ baseline accurate risk perception), partici conclusions of this review are limited by inadequate power to r ex- baseline accurate risk perception was an important variable fo ss detect important between-subgroup differences in effectivene plaining heterogeneity. Authors reported that when partici pants’ and by the wide variability in the decision contexts, the eleme nts n baseline scores for accurate risk perception were lower, decisio within the patient decision aids, the type of comparison deliv ered x- aids led to great improvement. Furthermore, we limited the e (collectively referred to as usual care here), the targeted outcom es, tracted study data to only two comparison groups (e.g. most inte n- r and the evaluation procedures. The small number of studies fo sive intervention including a patient decision aid and usual ca re); most outcomes did not allow for analysis of publication bias du e therefore, we did not investigate the possibility of interm ediate ies were to failure to publish negative studies. Moreover, most stud effects with less intensive decision aid interventions. there at unclear risk of selective outcome reporting, indicating that 30 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

34 Implications for research Agreements and disagreements with other studies or reviews Studies are needed to deepen our understanding of interactio ns between patient decision aid use and the patterns of patient-cl in- Our results confirm many of the observations reported in the ician communication; format issues such as the web-based deliv- previous versions of our review and in a comparative effective ness - ery of patient decision aids; and downstream effects on cost, re review that focused on studies evaluating oncology-specific pat ient - source use, and adherence. Although this update shows new stud ). We published the first systematic Trikalinos 2014 decision aids ( ies conducted in Spain and China, most studies have taken place review of 17 randomized trials of decision aids in 1999 ( O’Connor in North America, the UK, Europe, and Australia. There were fa r 1999b O’Connor 2001 ), followed by updates in 2003 with a total ; ation fewer studies of patient decision aids used within the consult O’Connor 2003 ), in 2009 with a total of 55 studies of 35 studies ( urther than those delivered pre-consultation, and this is an area of f O’Connor 2009b ), in 2011 with a total of 86 studies ( Stacey ( research given the important issue of implementation. 2011 Stacey 2014b ). ), and 2014 with a total of 115 studies ( With the addition of more studies in the systematic review, i t may ults, be possible to tease out the reasons for heterogeneity of res including variability in: study quality; comparison interve ntion; A U T H O R S ’ C O N C L U S I O N S elements within patient decision aids; decision type; settin g where ook- it was used; and format of decision aid (e.g. video, Internet, b Implications for practice let). Research should also explore the degree of detail in pati ent decision aids that is required for positive effects according to the The positive effects of decision aids on improving people’s kn owl- IPDAS criteria. In particular, evaluation is needed to compare the edge of risks and benefits, feeling informed, and feeling clea r about effect of those decision aids that meet the minimal IPDAS criter ia their values across a wide variety of decision contexts provide s suf- S for certification versus those that meet the full roster of IPDA ficient evidence for using them in clinical practice. They proba- quality criteria ( Joseph-Williams 2013 ). tion bly also facilitate accurate risk perception and active participa in decision making. However, several conditions may be neces- sary for successful implementation, including: good quality d eci- sion aids that meet the needs of the population; clinicians who are willing to use decision aids in their practice; effective sys tems A C K N O W L E D G E M E N T S for delivering decision support; and clinicians and healthcare con- sumers who are skilled in shared decision making. Although th ere The Cochrane Consumers and Communication Group (editors, have been some strides in achieving these conditions ( Elwyn 2013 ; academic and consumer referees) provided peer review and ad- O’Connor 2007 ), the use of patient decision aids will not occur vice regarding the review and checked extracted data for newly without adequate attention to implementation barriers to i mple- included studies. We thank John Kis-Rigo at La Trobe Univer- oduc- mentation and careful design of effective strategies for intr . sity who revised the search strategy used since the 2014 update ing and maintaining their use in routine clinical practice ( Elwyn David Rovner and Nananda Col helped with screening studies 2013 Legare 2010 ; Legare 2014 ). ; Gravel 2006 ; Legare 2008b ; on and for inclusion, and Intissar Souli assisted with data extracti d a ’Risk of bias’ assessment. Anton Saarimaki set up and manage d New in this update was a subgroup analysis of the findings base ated web-based title and abstract screening application that facilit on timing of decision aid used either before or during a consul- independent screening of citations by the review authors, and he tation. Although knowledge scores and accurate risk perception s also verified references for patient decision aids in were significantly higher in the decision aid group compared to Table 1 . Alain the usual care, there was no difference in these outcomes when Mayhew provided guidance in the interpretation of the ’Summa ry comparing decision aids used in preparation for versus during the of findings’ table. Dean Fergusson provided consultation on t he consultation. statistical analysis. 31 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

35 R E F E R E N C E S References to studies included in this review Berry 2013 {published data only} ∗ Berry DL, Halpenny B, Hong F, Wolpin S, Lober WB, {published data only} Allen 2010 Russell KJ, et al. The personal patient profile-prostate Allen JD, Othus MK, Hart A Jr, Tom L, Li Y, Berry D, et decision support for men with localized prostate cancer: a al. A randomized trial of a computer-tailored decision aid Urologic Oncology multi-center randomized trial. 31 2013; to improve prostate cancer screening decisions: results from (7):1012–21. the take the wheel trial. Cancer Epidemiology, Biomarkers Berry DL, Wang Q, Halpenny B, Hong F. Decision and Prevention 19 2010; (9):2172–86. preparation, satisfaction and regret in a multi-center sample Allen JD, Othus MKD, Hart A Jr, Mohllajee AP, Bowen of men with newly diagnosed localized prostate cancer. D. Do men make informed decisions about prostate cancer Patient Education and Counseling 2012; 88 (2):262–7. screening? Baseline results from the “Take the Wheel” Trial. Bosco JLF, Halpenny B, Berry DL. Personal preferences 2011; Medical Decision Making 31 :108–120. and discordant prostate cancer treatment choice in an {published data only} Arterburn 2011 intervention trial of men newly diagnosed with localized Arterburn D, Westbrook E, Bogart T, Sepucha K, Bock 2012; prostate cancer. Health and Quality of Life Outcomes S, Weppner W. Randomized trial of a video-based patient 10 (123):1–8. (8): 19 2011; decision aid for bariatric surgery. Obesity Underhill ML, Hong F, Berry DL. When study site 1669–75. contributes to outcomes in a multi-center randomized Auvinen 2004 {published data only} trial: a secondary analysis of decisional conflict in men ∗ Auvinen A, Hakama M, Ala-Opas M, Vornanen T, with localized prostate cancer. Health and Quality of Life Leppilahti M, Salminen P, et al. A randomized trial of choice 12 Outcomes 2014; :159. of treatment in prostate cancer: the effect of intervention on {published data only} Bjorklund 2012 93 2004; BJU International the treatment chosen. (1):52–6. Bjorklund U, Marsk A, Levin C, Ohman SG. Audiovisual Auvinen A, Vornanen T, Tammela TL, Ala-Opas M, information affects informed choice and experience of Leppilahti M, Salminen P, et al. A randomized trial of the information in antenatal Down syndrome screening- choice of treatment in prostate cancer: design and baseline a randomized controlled trial. Patient Education and BJU International (7):708–15. 2001; characteristics. 88 2012; Counseling 86 (3):390–5. Huang RC, Auvinen A, Hakama M, Tammela TLJ, Ala- Öhman SG, Björklund U, Marsk A. Does an informational Opas M, Leppilahti M, et al. Effect of intervention on film increase women’s possibility to make an informed decision making of treatment for disease progression, Prenatal choice about second trimester ultrasound?. prostate-specific antigen biochemical failure and prostate 32 Diagnosis (9):833–9. 2012; 17 2014; cancer death. Health Expectations (6):776–83. Barry 1997 {published and unpublished data} Bozic 2013 {published data only} ∗ Barry MJ, Cherkin DC, Chang Y, Fowler FJ, Skates S. A Bozic KJ, Belkora J, Chan V, Youm J, Zhou T, Dupaix J, et randomized trial of a multimedia shared decision-making al. Shared decision making in patients with osteoarthritis of program for men facing a treatment decision for benign the hip and knee: results of a randomized controlled trial. Disease Management and Clinical prostatic hyperplasia. Journal of Bone and Joint Surgery: American Volume 2013; 95 Outcomes 1 (1):5–14. 1997; (18):1633–9. Rovner DR, Wills CE, Bonham V, Williams G, Lillie J, Bozic KJ, Chenok KE, Schindel J, Chan V, Huddleston JI, Kelly-Blake K, et al. Decision aids for benign prostatic Braddock C, Belkora J. Patient, surgeon, and healthcare Medical hyperplasia: applicability across race and education. purchaser views on the use of decision and communication Decision Making 2004; 24 (4):359–66. BMC aids in orthopaedic surgery: a mixed methods study. 14 2014; (366):1–10. Health Services Research Bekker 2004 {published data only} ∗ Youm J, Chan V, Belkora J, Bozic KJ. Impact of Bekker HL, Hewison J, Thornton JG. Applying decision socioeconomic factors on informed decision making and analysis to facilitate informed decision making about treatment choice in patients with hip and knee OA. The prenatal diagnosis for Down syndrome: a randomised (2):171–5. 30 2015; Journal of Arthroplasty controlled trial. Prenatal Diagnosis 2004; 24 (4):265–75. Bekker HL, Hewison J, Thornton JG. Understanding why Brazell 2014 {published data only} decision aids work: linking process with outcome. Patient Brazell HD, O’Sullivan DM, Forrest A, Greene JF. Effect of Education and Counseling 50 (3):323–9. 2003; a decision aid on decision making for the treatment of pelvic Bernstein 1998 {published and unpublished data} organ prolapse. Female Pelvic Medicine & Reconstructive Bernstein SJ, Skarupski KA, Grayson CE, Starling MR, 2014; 21 (4):231–5. Surgery Bates ER, Eagle KA. A randomized controlled trial of Chabrera 2015 {published data only} information-giving to patients referred for coronary Chabrera C, Zabalegui A, Bonet M, Caro M, Areal J, Health angiography: effects on outcomes of care. González JR, Font A. A decision aid to support informed 1 (1):50–61. Expectations 1998; 32 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

36 choices for patients recently diagnosed with prostate cancer. trial of a decision support tool to improve the quality of 38 (3):E42–E50. 2015; Cancer Nursing communication and decision-making in individuals with Journal of the American Geriatrics Society atrial fibrillation. Chambers 2012 {published data only} 60 2012; (8):1434–41. Chambers LW, Wilson K, Hawken S, Puxty J, Crowe L, Lam PP, et al. Impact of the Ottawa influenza decision aid {published data only} Frosch 2008a on healthcare personnel’s influenza immunization decision: Frosch DL, Bhatnagar V, Tally S, Hamori CJ, Kaplan RM. Journal of Hospital Infection 2012; (3): a randomized trial. 82 Internet patient decision support: a randomized controlled 194–202. trial comparing alternative approaches for men considering Clancy 1988 {published data only} Archives of Internal Medicine prostate cancer screening. Clancy CM, Cebul RD, Williams SV. Guiding individual (4):363–9. 168 2008; decisions: a randomized, controlled trial of decision Gattellari 2003 {published data only} 1988; analysis. American Journal of Medicine 84 (2):283–8. Gattellari M, Ward JE. Does evidence-based information {published data only} Davison 1997 about screening for prostate cancer enhance consumer Davison BJ, Degner LF. Empowerment of men newly decision-making? A randomised controlled trial. Journal of (3): Cancer Nursing diagnosed with prostate cancer. 1997; 20 Medical Screening 2003; (1):27–39. 10 187–96. Gattellari 2005 {published data only} De Achaval 2012 {published data only} Gattellari M, Ward JE. A community-based randomised De Achaval S, Fraenkel L, Volk R, Cox V, Suarez-Almazor controlled trial of three different educational resources for M. Impact of educational and patient decision aids on Patient Education and men about prostate cancer screening. decisional conflict associated with total knee arthroplasty. 57 Counseling (2):168–82. 2005; 64 (2):229–37. 2012; Arthritis Care & Research {published data only} Dolan 2002 {published data only} Green 2001 Dolan JG, Frisina S. Randomized controlled trial of a Green MJ, Biesecker BB, McInerney AM, Mauger D, Fost Medical patient decision aid for colorectal cancer screening. N. An interactive computer program can effectively educate (2):125–39. 22 2002; Decision Making patients about genetic testing for breast cancer susceptibilit y. (1):16–23. American Journal of Medical Genetics 103 2001; Evans 2010 {published data only} Evans R, Joseph-Williams N, Edwards A, Newcombe R, {published data only} Hamann 2006 Wright P, Kinnersley P, et al. Supporting informed decision Hamann J, Cohen R, Leucht S, Busch R, Kissling W. Shared making for prostate specific antigen (PSA) testing on the decision making and long-term outcome in schizophrenia web: an online randomized controlled trial. Journal of 68 Journal of Clinical Psychiatry treatment. (7):992–7. 2007; ∗ Medical Internet Research 2010; 12 (3):e27. Hamann J, Langer B, Winkler V, Busch R, Cohen R, Fagerlin 2011 {published data only} Leucht S, et al. Shared decision making for in-patients with Banegas MP, McClure JB, Barlow WE, Ubel PA, Smith 114 (4): Acta Psychiatrica Scandinavica schizophrenia. 2006; DM, Zikmund-Fisher BJ, et al. Results from a randomized 265–73. trial of a web-based, tailored decision aid for women at high Hanson 2011 {published data only} risk for breast cancer. Patient Education and Counseling Ersek M, Sefcik JS, Feng-Chang L, Lee TJ, Gilliam R, 91 :364-71. 2013; Hanson LC. Provider staffing effect on a decision aid Fagerlin A. Randomization for Guide to Decide phase II. :36–53. Clinical Nursing Research intervention. 23 2014; Word document provided by the authors. Hanson L, Carey T, Caprio A, Joon Lee T, Ersek M, Garrett ∗ Fagerlin A, Dillard AJ, Smith DM, Zikmund-Fisher BJ, J, et al. Improving decision making for feeding options in Pitsch R, McClure JB, et al. Women’s interest in taking advanced dementia: a randomized, controlled trial. Journal tamoxifen and raloxifene for breast cancer prevention: (11):2009–16. of the American Geriatrics Society 59 2011; response to a tailored decision aid. Breast Cancer Research Snyder EA, Caprio AJ, Wessell K, Lin FC, Hanson LC. 127 2011; and Treatment (3):681–8. Impact of a decision aid on surrogate decision-makers’ Korfage IJ, Fuhrel-Forbis A, Ubel PA, Zikmund-Fisher BJ, perceptions of feeding options for patients with dementia. Greene SM, McClure JB, et al. Informed choice about 2013; 14 (2):114–8. American Medical Directors Association breast cancer prevention: randomized controlled trial of an {published data only} Heller 2008 Breast Cancer Research online decision aid intervention. Heller L, Parker PA, Youssef A, Miller MJ. Interactive 2013; 15 (R74):1–9. Plastic & digital education aid in breast reconstruction. Fraenkel 2007 {published data only} Reconstructive Surgery 2008; 122 (3):717–24. Fraenkel L, Rabidou N, Wittink D, Fried T. Improving informed decision-making for patients with knee pain. Hess 2012 {published data only} 2007; Journal of Rheumatology 34 (9):1894–8. Hess EP, Knoedler MA, Shah ND, Kline JA, Breslin M, {published data only} Fraenkel 2012 Branda ME, et al. The chest pain choice decision aid: a Fraenkel L, Street RL Jr, Towle V, O’Leary JR, Iannone L, randomized trial. Circulation: Cardiovascular Quality and Van Ness PH, Fried TR. A pilot randomized controlled Outcomes 2012; 5 (3):251–9. 33 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

37 {published data only} Laupacis 2006 {published data only} Jibaja-Weiss 2011 Laupacis A, O’Connor AM, Drake ER, Rubens FD, Jibaja-Weiss M, Volk R, Granchi T, Neff N, Robinson E, Robblee JA, Grant FC, et al. A decision aid for autologous Spann S, et al. Entertainment education for breast cancer Patient pre-donation in cardiac surgery - a randomized trial. surgery decisions: a randomized trial among patients with 2006; Education and Counseling 61 (3):458–66. Patient Education and Counseling 2011; low health literacy. 84 (1):41–8. LeBlanc 2015 {published data only} {published data only} Johnson 2006 LeBlanc A, Wang AT, Wyatt K, Branda ME, Shah ND, Johnson BR, Schwartz A, Goldberg J, Koerber A. A Van Houten H, et al. Encounter decision aid vs. clinical chairside aid for shared decision making in dentistry: a decision support or usual care to support patient-centered Journal of Dental Education randomized controlled trial. e treatment decisions in osteoporosis: the osteoporosis choic (2):133–41. 70 2006; (5):1–13. 2015; randomized trial II. PLOS ONE 10 {published data only} Kasper 2008 Legare 2008a {published data only} Kasper J, Kopke S, Muhlhauser I, Nubling M, Heesen C. Legare F, Dodin S, Stacey D, Leblanc A, Tapp S. Patient Informed shared decision making about immunotherapy for decision aid on natural health products for menopausal patients with multiple sclerosis (ISDIMS): A randomized symptoms: randomized controlled trial. Menopause 2008; European Journal of Neurology 15 (12): controlled trial. 14 2008; International (3):105–10. 1345–52. Legare 2011 {published data only} Kennedy 2002 {published data only} Legare F, Labrecque M, LeBlanc A, Njoya M, Laurier C, Kennedy AD, Sculpher MJ, Coulter A, Dwyer N, Rees M, Cote L, et al. Training family physicians in shared decision Abrams KR, et al. Effects of decision aids for menorrhagia making for the use of antibiotics for acute respiratory on treatment choices, health outcomes, and costs: a infections: a pilot clustered randomized controlled trial. JAMA 2002; (21):2701–8. randomized controlled trial. 288 2011; :96–110. 14 Health Expectations {published data only} Knops 2014 Legare 2012 {published and unpublished data} Knops AM, Goossens A, Ubbink DT, Balm R, Koelemay Legare F, Labrecque M, Cauchon M, Castel J, Turcotte S, MJ, Vahl AC, et al. DECAID Trial Group. A decision Grimshaw J. Training family physicians in shared decision- aid regarding treatment options for patients with an making to reduce the overuse of antibiotics in acute asymptomatic abdominal aortic aneurysm: a randomised respiratory infections: a cluster randomized trial. Canadian European Journal of Vascular and Endovascular clinical trial. (13):E726–34. 2012; 184 Medical Association Journal 2014; Surgery (3):276–283. 48 Leighl 2011 {published data only} Krist 2007 {published data only} Leighl NB, Shepherd HL, Butow PN, Clarke SJ, McJannett Krist AH, Woolf SH, Johnson RE, Kerns JW. Patient M, Beale PJ, et al. Supporting treatment decision making education on prostate cancer screening and involvement in in advanced cancer: a randomized trial of a decision aid decision making. Annals of Family Medicine 2007; 5 (2): for patients with advanced colorectal cancer considering 112–9. (15): 29 2011; chemotherapy. Journal of Clinical Oncology Kupke 2013 {published data only (unpublished sought but not used)} 2077–84. Kupke J, Wicht MJ, Stützer H, Derman SH, Lichtenstein NV, Noack MJ. Does the use of a visualised decision Lepore 2012 {published data only} board by undergraduate students during shared decision- Lepore SJ, Wolf RL, Basch CE, Godfrey M, McGinty making enhance patients’ knowledge and satisfaction? A E, Shmukler C, et al. Informed decision making about randomised controlled trial. European Journal of Dental prostate cancer testing in predominantly immigrant black (1):19–25. Education 17 2013; Annals of Behavioral men: a randomized controlled trial. Medicine 44 2012; (3):320–30. {published data only} Kuppermann 2014 Kuppermann M, Pena S, Bishop JT, Nakagawa S, Gregorich {published data only} Lerman 1997 SE, Sit A, et al. Effect of enhanced information, values Lerman C, Biesecker B, Benkendorf JL, Kerner J, Gomez- clarification, and removal of financial barriers on use of Caminero A, Hughes C, et al. Controlled trial of pretest prenatal genetic testing: a randomized clinical trial. Journal education approaches to enhance informed decision-making 2014; 312 of the American Medical Association (12):1210–7. Journal of the National Cancer for BRCA1 gene testing. Institute 1997; (2):148–57. 89 {published data only} Lam 2013 Lam WW, Chan M, Or A, Kwong A, Suen D, Fielding R. {published data only} Lewis 2010 Reducing treatment decision conflict difficulties in breast ∗ Lewis C, Pignone M, Schild L, Scott T, Winquist A, Journal of cancer surgery: a randomized controlled trial. Rimer B, et al. Effectiveness of a patient and practice-level 2013; 31 Clinical Oncology (23):2879–85. colorectal cancer screening intervention in health plan {published data only} Langston 2010 members: design and baseline findings of the CHOICE Langston A, Rosario L, Westhoff C. Structured contraceptive trial. Cancer 2010; 116 (7):1664–73. Patient Education counseling: a randomised controlled trial. Pignone M, Winquist A, Schild L, Lewis C, Scott T, 81 and Counseling (3):362–7. 2010; Hawley J, et al. Effectiveness of a patient and practice- 34 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

38 level colorectal cancer screening intervention in health plan {published data only} Mathieu 2007 Cancer 2011; 117 (15):3252–62. members. Mathieu E, Barratt A, Davey HM, McGeechan K, Howard Pignone MP, Brenner AT, Hawley S, Sheridan SL, Lewis K, Houssami N. Informed choice in mammography CL, Jonas DE, et al. Conjoint analysis versus rating and screening: a randomized trial of a decision aid for 70-year- ranking for values elicitation and clarification in colorectal (19): 167 2007; Archives of Internal Medicine old women. 2011; cancer screening. Journal of General Internal Medicine 2039–46. (1):45–50. 27 {published data only} Mathieu 2010 Mathieu E, Barratt AL, McGeechan K, Davey HM, Howard {published data only} Loh 2007 K, Houssami N. Helping women make choices about Loh A, Simon D, Harter M. Effects of shared decision mammography screening: an online randomized trial of a making in primary care of depressive patients - better Patient Education and decision aid for 40-year-old women. (1): compliance and treatment effects. Klinikarzt 2007; 36 Counseling 2010; 81 (1):63–72. 38–41. ∗ Loh A, Simon D, Wills CE, Kriston L, Niebling W, Harter McAlister 2005 {published data only} M. The effects of a shared decision-making intervention in McAlister FA, Man-Son-Hing M, Straus SE, Ghali WA, primary care of depression: a cluster-randomized controlled Anderson D, Majumdar SR, et al. Impact of a patient (3):324–32. 67 2007; Patient Education and Counseling trial. decision aid on care among patients with nonvalvular atrial CMAJ fibrillation: a cluster randomized trial. (5): 173 2005; {published data only} Mann D 2010 496–501. Mann DM, Ponieman D, Montori VM, Arciniega J, McBride 2002 {published data only} McGinn T. The statin choice decision aid in primary care: a Bastian LA, McBride CM, Fish L, Lyna P, Farrell D, randomized trial. Patient Education and Counseling 2010; Lipkus IM, et al. Evaluating participants’ use of a hormone (1):138–40. 80 replacement therapy decision-making intervention. Patient Mann E 2010 {published data only} Education and Counseling 2002; 48 (3):283–91. Mann E, Kellar I, Sutton S, Kinmonth AL, Hankins M, ∗ McBride CM, Bastian LA, Halabi S, Fish L, Lipkus IM, Griffin S, et al. Impact of informed-choice invitations on Bosworth HB, et al. A tailored intervention to aid decision diabetes screening knowledge, attitude and intentions: an American making about hormone replacement therapy. BMC Public Health :768. 10 2010; analogue study. 92 2002; Journal of Public Health (7):1112–4. Man-Son-Hing 1999 {published and unpublished data} {published data only} McCaffery 2010 Man-Son-Hing M, Laupacis A, O’Connor AM, Biggs J, McCaffery KJ, Irwig L, Turner R, Chan SF, Macaskill P, Drake E, Yetisir E, et al. A patient decision aid regarding Lewicka M, et al. Psychosocial outcomes of three triage antithrombotic therapy for stroke prevention in atrial methods for the management of borderline abnormal 1999; fibrillation: a randomized controlled trial. JAMA 282 2010; BMJ cervical smears: an open randomised trial. : 340 (8):737–43. b4491. Miller 2005 {published data only} {published data only} Marteau 2010 Miller SM, Fleisher L, Roussi P, Buzaglo JS, Schnoll R, Kellar I, Mann E, Kinmonth AL, Prevost AT, Sutton Slater E, et al. Facilitating informed decision making about S, Marteau TM. Can informed choice invitations lead breast cancer risk and genetic counseling among women to inequities in intentions to make lifestyle changes calling the NCI’s Cancer Information Service. Journal of among participants in a primary care diabetes screening 2005; 10 (Suppl 1):119–36. Health Communication programme? Evidence from a randomized trial. Public 2011; (9):645–52. 125 Health Miller 2011 {published data only} ∗ Marteau TM, Mann E, Prevost AT, Vasconcelos JC, Duren-Winfield V, Onsomu EO, Case DL, Pignone Kellar I, Sanderson S, et al. Impact of an informed choice M, Miller D. Health literacy and computer-assisted invitation on uptake of screening for diabetes in primary instruction: usability and patient preference. Journal of care (DICISION): randomised trial. :c2138. BMJ 2010; 340 :491–8. 20 2015; Health Communication Miller D, Spangler J, Case D, Goff D, Singh S, Pignone Mathers 2012 {published data only} M. Effectiveness of a web-based colorectal cancer screening Brown I, Bradley A, Ng CJ, Colwell B, Mathers N. patient decision aid: a randomized controlled trial in a Investigating active ingredients in a complex intervention: American Journal of Preventive mixed-literacy population. a nested study within the Patient and Decision Aids Medicine (6):608–15. 40 2011; (PANDAs) randomised controlled trial for people with type {published and unpublished data} Montgomery 2003 2014; 2 diabetes. 7 BMC Research Notes :347. Emmett CL, Montgomery AA, Peters TJ, Fahey T. Three- Mathers N, Ng CJ, Campbell MJ, Colwell B, Brown year follow-up of a factorial randomised controlled trial of I, Bradley A. Clinical effectiveness of a patient decision two decision aids for newly diagnosed hypertensive patients. aid to improve decision quality and glycaemic control in (516):551–3. British Journal of General Practice 2005; 55 people with diabetes making treatment choices: a cluster ∗ Montgomery AA, Fahey T, Peters TJ. A factorial randomised controlled trial (PANDAs) in general practice. randomised controlled trial of decision analysis and 2 (6):1–12. 2012; BMJ Open 35 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

39 an information video plus leaflet for newly diagnosed decision aid on hormone replacement therapy in primary British Journal of General Practice hypertensive patients. 323 2001; BMJ care. (7311):490–3. 2003; (491):446–53. 53 {published data only} Nagle 2008 ∗ Montgomery 2007 {published data only} Nagle C, Gunn J, Bell R, Lewis S, Meiser B, Metcalfe Frost J, Shaw. Women’s views on the use of decision aids S, et al. Use of a decision aid for prenatal testing of fetal for decision making about the method of delivery following abnormalities to improve women’s informed decision a previous caesarean section: qualitative interview study. BJOG: An making: a cluster randomised controlled trial. (7): 116 2009; British Journal of Obstetrics and Gynecology International Journal of Obstetrics and Gynaecology 2008; 115 896–905. (3):339–47. Hollinghurst S, Emmett C, Peters TJ, Watson H, Fahey T, Nagle C, Lewis S, Meiser B, Metcalfe S, Carlin JB, Murphy DJ, et al. Economic evaluation of the DIAMOND Bell R, et al. Evaluation of a decision aid for prenatal randomized trial: cost and outcomes of 2 decision aids for testing of fetal abnormalities: a cluster randomised trial mode of delivery among women with previous caesarian 6 :96. BMC Public Health [ISRCTN22532458]. 2006; BMJ 30 2010; :453–63. section. Nassar 2007 {published data only} ∗ Montgomery AA, Emmett CL, Fahey T, Jones C, Ricketts Nassar N, Roberts CL, Raynes-Greenow CH, Barratt I, Patel RR, et al. Two decision aids for mode of delivery A, Peat B, Decision Aid for Breech Presentation Trial among women with previous caesarean section: randomised Collaborators. Evaluation of a decision aid for women with (7607):1305. 2007; BMJ controlled trial. 334 breech presentation at term: a randomised controlled trial [ISRCTN14570598]. BJOG: An International Journal of Montori 2011 {published data only} ∗ 114 Obstetrics & Gynaecology 2007; (3):325–33. Montori VM, Shah ND, Pencille LJ, Branda ME, Van Houten HK, Swiglo BA. Use of a decision aid to improve {published data only} Oakley 2006 e treatment decisions in osteoporosis: the osteoporosis choic Oakley S, Walley T. A pilot study assessing the effectiveness randomized trial. American Journal of Medicine 124 2011; of a decision aid on patient adherence with oral (6):549–56. 2006; Pharmaceutical Journal bisphosphonate medication. Pencille LJ, Campbell ME, Van Houten HK, Shah ND, (7399):536–8. 276 Mullan RJ, Swiglo BA, et al. Protocol for the Osteoporosis {published data only} Ozanne 2007 Choice trial. A pilot randomized trial of a decision aid in Ozanne EM, Annis C, Adduci K, Showstack J, Esserman L. 2009; :113. 10 primary care practice. Trials Pilot trial of a computerized decision aid for breast cancer Morgan 2000 {published and unpublished data} Breast Journal prevention. (2):147–54. 13 2007; Morgan MW. A Randomized Trial of the Ischemic Heart Partin 2004 {published and unpublished data} Disease Shared Decision Making Program: An Evaluation Partin MR, Nelson D, Flood AB, Friedemann-Sanchez G, of a Decision Aid [Masters Thesis] . Toronto: University of Wilt TJ. Who uses decision aids? Subgroup analyses from Toronto, 1997. a randomized controlled effectiveness trial of two prostate ∗ Morgan MW, Deber RB, Llewellyn-Thomas HA, cancer screening decision support interventions. Health Gladstone P, Cusimano RJ, O’Rourke K, et al. Randomized, 2006; 9 (3):285–95. Expectations controlled trial of an interactive videodisc decision aid for ∗ Partin MR, Nelson D, Radosevich D, Nugent S, Flood patients with ischemic heart disease. Journal of General AB, Dillon N, et al. Randomized trial examining the effect Internal Medicine 2000; (10):685–93. 15 of two prostate cancer screening educational interventions Mott 2014 {published data only} Journal on patient knowledge, preferences, and behaviors. Mott JM, Stanley MA, Street RL Jr, Grady RH, Teng EJ. (8):835–42. of General Internal Medicine 2004; 19 Increasing engagement in evidence-based PTSD treatment Pignone 2000 {published data only} Military through shared decision-making: a pilot study. Pignone M, Harris R, Kinsinger L. Videotape-based (2):143–9. Medicine 2014; 179 decision aid for colon cancer screening. A randomized, {published data only} Mullan 2009 2000; (10): 133 controlled trial. Annals of Internal Medicine Mullan RJ, Montori VM, Shah ND, Christianson TJ, 761–9. Bryant SC, Guyatt GH, et al. The diabetes mellitus Protheroe 2007 {published data only} medication choice decision aid: a randomized trial. Archives Patel S, Ngunjiri A, Hee SW, Yang Y, Brown S, Friede T, et (17):1560–8. 169 2009; of Internal Medicine al. Primum non nocere: shared informed decision making {published and unpublished data} Murray 2001a in low back pain - a pilot cluster randomised trial. BMC Murray E, Davis H, Tai SS, Coulter A, Gray A, Haines A. 15 2014; Musculoskeletal Disorders :282. Randomised controlled trial of an interactive multimedia Protheroe J, Bower P, Chew-Graham C. The use of decision aid on benign prostatic hypertrophy in primary mixed methodology in evaluating complex interventions: 323 2001; BMJ care. (7311):493–6. identifying patient factors that moderate the effects of a (6):594–600. 24 decision aid. Family Practice 2008; {published and unpublished data} Murray 2001b ∗ Protheroe J, Bower P, Chew-Graham C, Peters TJ, Fahey Murray E, Davis H, Tai SS, Coulter A, Gray A, Haines A. T. Effectiveness of a computerized decision aid in primary Randomized controlled trial of an interactive multimedia 36 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

40 care on decision making and quality of life in menorrhagia: patients’ discussions with their doctor and their plans for results of the MENTIP randomized controlled trial. BMC Health Services prevention: a pilot randomized trial. Medical Decision Making 2007; 27 (5):575–84. Research :121. 6 2006; Rubel 2010 {published data only} {published data only} Sheridan 2011 Rubel SK, Miller JW, Stephens RL, Xu Y, Scholl LE, Sheridan SL, Draeger LB, Pignone MP, Keyserling TC, Holden EW, et al. Testing the effects of a decision aid for Simpson RJ Jr, Rimer B, et al. A randomized trial of an prostate cancer screening. Journal of Health Communication intervention to improve use and adherence to effective (3):307–21. 15 2010; BMC Health coronary heart disease prevention strategies. :331. Services Research 2011; 11 {published data only} Ruffin 2007 Sheridan SL, Draeger LB, Pignone MP, Rimer B, Bangdiwala Ruffin MT, Fetters MD, Jimbo M. Preference-based SI, Cai J, Gizlice Z, Keyserling TC, Simpson RJ. The effect electronic decision aid to promote colorectal cancer of a decision aid intervention on decision making about screening: results of a randomized controlled trial. coronary heart disease risk reduction: secondary analyses of Preventive Medicine 45 (4):267–73. 2007; BMC Medical Informatics and Decision a randomized trial. Sawka 2012 {published and unpublished data} 14 (14):1–11. 2014; Making Sawka AM, Straus S, Rotstein L, Brierley JD, Tsang RW, Asa S, et al. Randomized controlled trial of a computerized Shorten 2005 {published and unpublished data} decision aid on adjuvant radioactive iodine treatment for Shorten A, Shorten B, Keogh J, West S, Morris J. Making Journal of patients with early-stage papillary thyroid cancer. choices for childbirth: a randomized controlled trial of a 30 (23):2906–11. 2012; Clinical Oncology decision-aid for informed birth after cesarean. Birth 2005; Schroy 2011 {published data only} (4):252–61. 32 ∗ Schroy PC 3rd, Emmons K, Peters E, Glick JT, Robinson Shourie 2013 {published data only} PA, Lydotes MA, et al. The impact of a novel computer- Shourie S, Jackson C, Cheater FM, Bekker HL, Edlin based decision aid on shared decision making for colorectal R, Tubeuf S, et al. A cluster randomised controlled trial cancer screening: a randomized trial. Medical Decision of a web based decision aid to support parents’ decisions (1):93–107. 2011; Making 31 about their child’s Measles Mumps and Rubella (MMR) Schroy PC 3rd, Emmons KM, Peters E, Glick JT, Robinson 31 (50):6003–10. Vaccine vaccination. 2013; PA, Lydotes MA, et al. Aid-assisted decision making and colorectal cancer screening: a randomized controlled Smith 2010 {published data only} (6): 43 2012; American Journal of Preventive Medicine trial. Smith SK, Barratt A, Trevana L, Simpson JM, Jansen J, 573–83. McCaffery KJ. A theoretical framework for measuring Patient Education knowledge in screening decision aid trials. {published data only} Schwalm 2012 2012; :330–6. 89 and Counseling Schwalm JD, Stacey D, Pericak D, Natarajan MK. Radial Smith SK, Kearney P, Trevena L, Barratt A, Nutbeam D, artery versus femoral artery access options in coronary McCaffery KJ. Informed choice in bowel cancer screening:a angiogram procedures: randomized controlled trial of a qualitative study to explore how adults with lower education Circulation: Cardiovascular Quality patient-decision aid. :511–22. 17 2012; Health Expectations use decision aids. 5 (3):260–6. 2012; and Outcomes Smith SK, Simpson JM, Trevena LJ, McCaffery KJ. Factors {published data only} Schwartz 2001 associated with informed decisions and participation in Schwartz MD, Benkendorf J, Lerman C, Isaacs C, Ryan- bowel cancer screening among adults with lower education Robertson A, Johnson L. Impact of educational print and literacy. Medical Decision Making 2014; 34 (6):756–72. materials on knowledge, attitudes, and interest in BRCA1/ Smith SK, Trevena L, Simpson JM, Barratt A, Nutbeam Cancer BRCA2: testing among Ashkenazi Jewish women. D, McCaffery KJ. A decision aid to support informed 2001; (4):932–40. 92 choices about bowel cancer screening among adults with Schwartz 2009a {published data only} BMJ 2010; low education: randomised controlled trial. Hooker GW, Leventhal KG, DeMarco T, Peshkin BN, 341 :c5370. Finch C, Wahl E, et al. Longitudinal changes in patient Stacey 2014a {published and unpublished data} distress following interactive decision aid use among Stacey D, Hawker G, Dervin G, Tugwell P, Boland L, BRCA1/2 carriers: a randomized trial. Medical Decision Pomey MP, et al. Decision aid for patients considering (3):412–21. Making 2011; 31 ∗ total knee arthroplasty with preference report for surgeons: Schwartz MD, Valdimarsdottir HB, DeMarco TA, BMC Musculoskeletal A pilot randomized controlled trial. Peshkin BN, Lawrence W, Rispoli J, et al. Randomized trial Disorders 15 :54. 2014; of a decision aid for BRCA1/BRCA2 mutation carriers: impact on measures of decision making and satisfaction. Steckelberg 2011 {published data only} Health Psychology 28 2009; (1):11–9. Steckelberg A, Hulfenhaus C, Haastert B, Muhlhauser I. Sheridan 2006 {published data only} Effect of evidence based risk information on “informed Sheridan SL, Shadle J, Simpson RJ Jr, Pignone MP. The choice” in colorectal cancer screening: randomised impact of a decision aid about heart disease prevention on controlled trial. BMJ 2011; 342 :d3193. 37 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

41 Taylor 2006 {published data only} treatment decision - a randomized trial among women with Taylor KL, Davis JL 3rd, Turner RO, Johnson L, Schwartz 6 2003; Health Expectations heavy menstruation. (4):290–7. MD, Kerner JF, et al. Educating African American men Vuorma S, Teperi J, Aalto AM, Hurskainen R, Kujansuu about the prostate cancer screening dilemma: a randomized E, Rissanen P. A randomized trial among women with intervention. Cancer Epidemiology, Biomarkers & Prevention heavy menstruation - impact of a decision aid on treatment (11):2179–88. 2006; 15 Health Expectations (4):327–37. outcomes and costs. 2004; 7 Thomson 2007 {published data only} {published data only} Watson 2006 Kaner E, Heaven B, Rapley T, Murtagh M, Graham R, Watson E, Hewitson P, Brett J, Bukach C, Evans R, Edwards Thomson R, et al. Medical communication and technology: A, et al. Informed decision making and prostate specific a video-based process study of the use of decision aids in antigen (PSA) testing for prostate cancer: a randomised primary care consultations. BMC Medical Informatics and controlled trial exploring the impact of a brief patient (2):1–11. 7 Decision Making 2007; decision aid on men’s knowledge, attitudes and intention ∗ Thomson RG, Eccles MP, Steen IN, Greenaway J, Patient Education and Counseling 63 to be tested. 2006; (3): Stobbart L, Murtagh MJ, et al. A patient decision aid 367–79. to support shared decision-making on anti-thrombotic Weymiller 2007 {published data only} treatment of patients with atrial fibrillation: randomised Jones LA, Weymiller AJ, Shah N, Bryant SC, Christianson Quality & Safety in Health Care controlled trial. 16 2007; TJH, Guyatt GH, et al. Should clinicians deliver decision (3):216–23. aids? further exploration of the statin choice randomized Trevena 2008 {published data only} 29 2009; trial results. Medical Decision Making (4):468–74. Trevena LJ, Irwig L, Barratt A. Randomized trial of a self- Nannenga MR, Montori VM, Weymiller AJ, Smith SA, administered decision aid for colorectal cancer screening. Christianson TJ, Bryant SC, et al. A treatment decision aid Journal of Medical Screening 2008; 15 (2):76–82. may increase patient trust in the diabetes specialist. The 2009; Health Expectations Statin Choice randomized trial. {published data only} Vandemheen 2009 (1):38–44. 12 Vandemheen KL, O’Connor A, Bell SC, Freitag A, Bye ∗ Weymiller AJ, Montori VM, Jones LA, Gafni A, Guyatt P, Jeanneret A, et al. Randomized trial of a decision GH, Bryant SC, et al. Helping patients with type 2 aid for patients with cystic fibrosis considering lung diabetes mellitus make treatment decisions: statin choice transplantation. American Journal of Respiratory & Critical 2007; randomized trial. 167 Archives of Internal Medicine (8):761–8. 180 2009; Care Medicine (10):1076–82. Van Peperstraten 2010 {published data only} Whelan 2003 {published and unpublished data} Kreuwel I, van Peperstraten A, Hulscher M, Kremer J, Whelan T, Sawka C, Levine M, Gafni A, Reyno L, Willan Grol R, Nelen W, Hermens R. Evaluation of an effective A, et al. Helping patients make informed choices: a multifaceted implementation strategy for elective single- randomized trial of a decision aid for adjuvant chemotherapy embryo transfer after in vitro fertilization. Human in lymph node-negative breast cancer. Journal of the 2013; Reproduction 28 (2):336–42. 95 (8):581–7. 2003; National Cancer Institute Van Peperstraten A, Nelen W, Grol R, Zielhuis G, Adang E, Stalmeier P, et al. The effect of a multifaceted empowerment Whelan 2004 {published and unpublished data} strategy on decision making about the number of embryos Whelan T, Levine M, Willan A, Gafni A, Sanders K, transferred in in vitro fertilisation: randomised controlled Mirsky D, et al. Effect of a decision aid on knowledge and :c2501. 341 BMJ 2010; trial. treatment decision making for breast cancer surgery: a randomized trial. JAMA 2004; (4):435–41. 292 {published data only} Vodermaier 2009 Vodermaier A, Caspari C, Koehm J, Kahlert S, Ditsch N, Williams 2013 {published and unpublished data} Untch M. Contextual factors in shared decision making: Williams RM, Davis KM, Luta G, Edmond SN, Dorfman a randomised controlled trial in women with a strong CS, Schwartz MD, et al. Fostering informed decisions: 2009; suspicion of breast cancer. British Journal of Cancer A randomized controlled trial assessing the impact of 100 (4):590–7. a decision aid among men registered to undergo mass screening for prostate cancer. Patient Education and {published and unpublished data} Volk 1999 ∗ Counseling :329–36. 91 2013; Volk RJ, Cass AR, Spann SJ. A randomized controlled trial of shared decision making for prostate cancer screening. Wolf 1996 {published data only} 8 1999; Archives of Family Medicine (4):333–40. ∗ Wolf AM, Nasser JF, Wolf AM, Schorling JB. The impact Volk RJ, Spann SJ, Cass AR, Hawley ST. Patient education of informed consent on patient interest in prostate-specific for informed decision making about prostate cancer Archives of Internal Medicine 156 1996; antigen screening. screening: a randomized controlled trial with 1-year follow- (12):1333–6. up. 2003; 1 (1):22–8. Annals of Family Medicine Wolf AM, Schorling JB. Preferences of elderly men for Vuorma 2003 {published data only} prostate-specific antigen screening and the impact of ∗ Vuorma S, Rissanen P, Aalto AM, Hurskainen R, Kujansuu informed consent. Journals of Gerontology Series A-Biological E, Teperi J. Impact of patient information booklet on Sciences & Medical Sciences 1998; 53 (3):M195–200. 38 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

42 {published data only} Armstrong 2005 Wolf 2000 {published and unpublished data} Armstrong K, Weber B, Ubel PA, Peters N, Holmes Wolf AM, Schorling JB. Does informed consent alter elderly J, Schwartz JS. Individualized survival curves improve patients’ preferences for colorectal cancer screening? Results satisfaction with cancer risk management decisions in Journal of General Internal Medicine of a randomized trial. women with BRCA1/2 mutations. Journal of Clinical 2000; 15 (1):24–30. (36):9319–28. 23 2005; Oncology Wong 2006 {published data only} Wong SS, Thornton JG, Gbolade B, Bekker HL. A {published data only} Arterburn 2013 Arterburn D, Flum DR, Westbrook EO, Fuller S, Shea M, randomised controlled trial of a decision-aid leaflet to facilitate women’s choice between pregnancy termination Bock SN, Landers J, Kowalski K, Turnbull E, Cummings DE, CROSSROADS Study Team. A population- methods. BJOG: An International Journal of Obstetrics & 2006; Gynaecology (6):688–94. 113 based, shared decision-making approach to recruit for a ype randomized trial of bariatric surgery versus lifestyle for t References to studies excluded from this review 2 diabetes. 9 Surgery for Obesity and Related Diseases 2013; (6):837–44. {published data only} Abadie 2009 {published data only} Au 2011 Abadie R, Weymiller AJ, Tilburt J, Shah ND, Charles C, Au AH, Lam WW, Chan MC, Or AY, Kwong A, Suen D, et Gafni A, et al. Clinician’s use of the Statin Choice decision al. Development and pilot-testing of a decision aid for use aid in patients with diabetes: a videographic study nested in Health among Chinese women facing breast cancer surgery. a randomized trial. Journal of Evaluation in Clinical Practice Expectations 2011; 14 (4):405–16. 2009; 15 (3):492–7. {published data only} Bakken 2014 Adab 2003 {published data only} Bakken S, Jia H, Chen ES, Choi J, John RM, Lee NJ, et Adab P, Marshall T, Rouse A, Randhawa B, Sangha H, al. The effect of a mobile health decision support system Bhangoo N. Randomised controlled trial of the effect on diagnosis and management of obesity, tobacco use, and of evidence based information on women’s willingness depression in adults and children. The Journal for Nurse Journal of to participate in cervical cancer screening. (10):774–80. Practitioners 2014; 10 57 (8):589–93. 2003; Epidemiology & Community Health {published data only} Becker 2009 Alegría 2014 {published data only} Becker H, Stuifbergen AK, Dormire SL. The effects of Alegría M, Carson N, Flores M, Li X, Shi P, Lessios AS, et hormone therapy decision support for women with mobility al. Activation, self-management, engagement, and retention 30 Health Care for Women International impairments. 2009; in behavioral health care: a randomized clinical trial of (9):845–54. JAMA Psychiatry the DECIDE intervention. 71 (5): 2014; {published data only} Belkora 2012 557–65. Belkora J, Stupar L, O’Donnell S, Loucks A, Moore D, Al Saffar 2008 {published data only} Jupiter C, et al. Decision support by telephone: randomized Al Saffar N, Abdulkareem A, Abdulhakeem A, Salah AQ, Patient controlled trial in a rural community setting. Heba M. Depressed patients’ preferences for education 89 2012; Education and Counseling (1):134–42. about medications by pharmacists in Kuwait. Patient {published data only} Bellmunt 2010 Education and Counseling (1):94–101. 72 2008; Bellmunt J, Eisen T, Szczylik C, Mulders P, Porta C. A new {published data only} Altiner 2007 patient-focused approach to the treatment of metastatic Altiner A, Brockmann S, Sielk M, Wilm S, Wegscheider renal cell carcinoma: establishing customized treatment K, Abholz HH. Reducing antibiotic prescriptions for (8):1190–9. BJU International 2010; options. 107 acute cough by motivating GPs to change their attitudes Bennett 2011 {published data only} to communication and empowering patients: a cluster- Bennett PA. Making the choice: cesarean delivery by randomized intervention study. Journal of Antimicrobial maternal request versus planned vaginal birth [PhD thesis]. (3):638–44. Chemotherapy 2007; 60 University of Colorado at Denver. ProQuest. Ann Arbor: {published data only} Anderson 2011 University of Colorado at Denver, 2011. Anderson C, Carter J, Nattress K, Beale P, Philp S, Harrison Bieber 2006 {published data only} J, et al. “The booklet helped me not to panic”: a pilot of a Bieber C, Muller KG, Blumenstiel K, Eich W. Participative decision aid for asymptomatic women with ovarian cancer decision-making as a measure to improve the doctor- International Journal of and with rising CA-125 levels. patient interaction with fibromyalgia patients [Partizipative Gynecological Cancer (4):737–43. 21 2011; Entscheidungsfindung als Maßnahme zur Verbesserung der {published data only} Arimori 2006 Arzt–Patient–Interaktion mit Fibromyalgie–Patientinnen]. Arimori N. Randomized controlled trial of decision aids 15 Zeitschrift fur Medizinische Psychologie 2006; (2):53–60. for women considering prenatal testing: the effect of the Bieber C, Muller KG, Blumenstiel K, Hochlehnert A, Ottawa Personal Decision Guide on decisional conflict. Wilke S, Hartmann M, et al. A shared decision-making Japan Journal of Nursing Science (2):119–30. 3 2006; communication training program for physicians treating 39 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

43 fibromyalgia patients: effects of a randomized controlled breast reconstruction: a pilot randomized controlled trial. Journal of Psychosomatic Research 2008; 64 (1):13–20. trial. 23 2015; Support Cancer Care :1365–1375. {published data only} Branda 2013 {published data only} Chadwick 1991 Branda ME, LeBlanc A, Shah ND, Tiedje K, Ruud K, Van Chadwick DJ, Gillatt DA, Gingell JC. Medical or surgical Houten H, et al. Shared decision making for patients with (6776): orchidectomy: the patients’ choice. 1991; 302 BMJ BMC type 2 diabetes: a randomized trial in primary care. 572. 13 Health Services Research (301):1–10. 2013; {published data only} Chan 2011 {published data only} Brenner 2014 Chan EC, McFall SL, Byrd TL, Mullen PD, Volk RJ, Brenner A, Howard K, Lewis C, Sheridan S, Crutchfield Ureda J, et al. A community-based intervention to promote T, Hawley S, et al. Comparing 3 values clarification informed decision making for prostate cancer screening methods for colorectal cancer screening decision-making: a among Hispanic American men changed knowledge and randomized trial in the US and Australia. Journal of General Patient Education and role preferences: a cluster RCT. 29 (3):507–13. Internal Medicine 2014; 84 Counseling 2011; (2):e44–51. {published data only} Breslin 2008 Chewning 1999 {published data only} Breslin M, Mullan RJ, Montori VM. The design of a Chewning B, Mosena P, Wilson D, Erdman H, Potthoff S, decision aid about diabetes medications for use during the Murphy A, et al. Evaluation of a computerized contraceptive consultation with patients with type 2 diabetes. Patient decision aid for adolescent patients. Patient Education and (3):465–72. 73 2008; Education and Counseling Counseling 1999; 38 (3):227–39. {published data only} Brown 2004 {published data only} Chiew 2008 Brown RF, Butow PN, Sharrock MA, Henman M, Boyle F, Chiew KS, Shepherd H, Vardy J, Tattersall MHN, Butow Goldstein D, et al. Education and role modelling for clinical PN, Leighl NB. Development and evaluation of a decision Health Expectations decisions with female cancer patients. aid for patients considering first-line chemotherapy for 2004; (4):303–16. 7 11 metastatic breast cancer. Health Expectations 2008; (1): Brundage 2001 {published data only} 35–45. Brundage MD, Feldman-Stewart D, Cosby R, Gregg R, {published data only} Clouston 2014 Dixon P, Youssef Y, et al. Phase I study of a decision aid for Clouston K, Katz A, Martens PJ, Sisler J, Turner D, patients with locally advanced non-small-cell lung cancer. Lobchuk M, et al. CIHR/CCMB Team in Primary Care Journal of Clinical Oncology 2001; (5):1326–35. 19 Oncology (PCO-NET). Does access to a colorectal cancer Burton 2007 {published data only} screening website and/or a nurse-managed telephone help Burton MJ. Booklet-based education in vestibular line provided to patients by their family physician increase rehabilitation or symptom control improved subjective fecal occult blood test uptake? Results from a pragmatic health in Meniere disease. Evidence-Based Medicine 2007; cluster randomized controlled trial. 2014; : 14 BMC Cancer 12 (4):111. 263. Buzhardt 2011 {published data only} Col 2007 {published data only} Buzhardt J, Greenwood CR, Walker D, Anderson R, Col NF, Ngo L, Fortin JM, Goldberg RJ, O’Connor AM. Howard W, Carta JJ. Effects of web-based support on early Can computerized decision support help patients make head start home visitors’ use of evidence-based intervention complex treatment decisions? A randomized controlled decision making and growth in children’s expressive Medical trial of an individualized menopause decision aid. NHSA Dialog (3):121–46. 2011; communication. 14 2007; Decision Making 27 (5):585–98. {published data only} Campbell 2014 {published data only} Colella 2004 Campbell SR, Holter MC, Manthey TJ, Rapp CA. The Colella KM, DeLuca G. Shared decision making in patients effect of CommonGround Software and Decision Support with newly diagnosed prostate cancer: a model for treatment Center. 2014; American Journal of Psychiatric Rehabilitation 24 (3):187- Urologic Nursing education and support. 2004; (2):166–80. 17 91, 195-6. Carling 2008 {published data only} {published data only} Costanza 2011 Carling C, Kristoffersen DT, Herrin J, Treweek S, Oxman Costanza ME, Luckmann RS, Rosal M, White MJ, LaPelle AD, Schunemann H, et al. How should the impact of N, Partin M, et al. Helping men make an informed decision different presentations of treatment effects on patient choice about prostate cancer screening: a pilot study of telephone be evaluated? A pilot randomized trial. 2008; 3 PLOS ONE 2011; (2): counseling. Patient Education and Counseling 82 (11):e3693. 193–200. {published data only} Causarano 2015 Causarano N, Platt J, Baxter NN, Bagher S, Jones JM, Coulter 2003 {published data only} Metcalfe KA, Hofer SOP, O’Neill AC, Cheng T, Starenkyj E, Coulter A. Patient information and shared decision-making Zhong T. Pre-consultation educational group intervention in cancer care. British Journal of Cancer 2003; 89 (Suppl 1): to improve shared decision-making for postmastectomy S15–6. 40 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

44 {published and unpublished data} Deyo 2000 Cox 2012 {published data only} ∗ Deyo RA, Cherkin DC, Weinstein J, Howe J, Ciol M, Cox CE, Lewis CL, Hanson LC, Hough CL, Kahn JM, Mulley AG. Involving patients in clinical decisions: impact White DB, et al. Development and pilot testing of a of an interactive video program on use of back surgery. decision aid for surrogates of patients with prolonged Medical Care (9):959–69. 38 2000; 40 (8): 2012; Critical Care Medicine mechanical ventilation. Phelan EA, Deyo RA, Cherkin DC, Weinstein JN, Ciol 2327–34. MA, Kreuter W, et al. Helping patients decide about back {published data only} Crang-Svalenius 1996 surgery: a randomized trial of an interactive video program. Crang-Svalenius E, Dykes AK, Jorgensen C. Women’s Spine (2):206–12. 26 2001; informed choice of prenatal diagnosis: early ultrasound examination-routine ultrasound examination-age- Diefenbach 2012 {published data only} independent amniocentesis. Fetal Diagnosis & Therapy Diefenbach MA, Mohamed NE, Butz BP, Bar-Chama N, (1):20–5. 1996; 11 Stock R, Cesaretti J, et al. Acceptability and preliminary feasibility of an internet/CD-ROM-based education and Davison 1999 {published data only} decision program for early-stage prostate cancer patients: Davison BJ, Kirk P, Degner LF, Hassard TH. Information randomized pilot study. Journal of Medical Internet Research and patient participation in screening for prostate cancer. (1):e6. 14 2012; Patient Education and Counseling 37 1999; (3):255–63. {published data only} Davison 2007 Dobke 2008 {published data only} Davison BJ, Goldenberg SL, Wiens KP, Gleave ME. Dobke MK, Bhavsar D. Pilot trial of telemedicine as a Comparing a generic and individualized information decision aid for patients with chronic wounds. Telemedicine decision support intervention for men newly diagnosed (3):245–9. 14 2008; Journal and e-health (5): with localized prostate cancer. Cancer Nursing 2007; 30 Dodin 2001 {published and unpublished data} E7–15. Dodin S, Legare F, Daudelin G, Tetroe J, O’Connor A. {published data only} De Boer 2012 Making a decision about hormone replacement therapy. A De Boer JC, van Blijderveen G, van Dijk G, randomized controlled trial [Prise de decision en matière Duivenvoorden HJ, Williams M. Implementing structured, d’hormonothérapie de remplacement]. Canadian Family multiprofessional medical ethical decision-making in a Physician 2001; 47 :1586–93. Journal of Medical Ethics 2012; neonatal intensive care unit. {published data only} Donovan 2012 38 (10):596–601. Donovan JL. Presenting treatment options to men with Deen 2012 {published data only} clinically localized prostate cancer: the acceptability of Deen D, Lu WH, Weintraub MR, Maranda MJ, Elshafey S, active surveillance/monitoring. Journal of the National Gold MR. The impact of different modalities for activating Cancer Institute. Monographs 2012; :191–6. 45 patients in a community health center setting. Patient 2012; 89 Education and Counseling (1):178–83. Driscoll 2008 {published data only} Driscoll DL, Rupert DJ, Golin CE, McCormack LA, De Haan 2013 {published data only} Sheridan SL, Welch BM. Promoting prostate-specific De Haan MC, de Wijkerslooth TR, Stoop E, Bossuyt P, antigen informed decision-making. Evaluating two Fockens P, Thomeer M, et al. Informed decision-making community-level interventions. American Journal of in colorectal cancer screening using colonoscopy or CT- Preventive Medicine 35 2008; (2):87–94. Patient Education and Counseling 2013; 91 colonography. (3):318–25. Dunn 1998 {published and unpublished data} Dunn RA, Shenouda PE, Martin DR, Schultz AJ. Videotape {published data only} Deinzer 2009 increases parent knowledge about poliovirus vaccines and Deinzer A, Veelken R, Kohnen R, Schmieder RE. Is a choices of polio vaccination schedules. Pediatrics 1998; 102 shared decision-making approach effective in improving (2):e26. hypertension management?. Journal of Clinical Hypertension (5):266–70. 11 2009; {published data only} Eaton 2011 Eaton L, Cherry C, Cain D, Pope H. A novel approach to Denig 2014 {published data only} prevention for at-risk HIV negative men who have sex with Denig P, Schuling J, Haaijer-Ruskamp F, Voorham J. Effects men: creating a teachable moment to promote informed of a patient oriented decision aid for prioritising treatment American Journal of Public Health sexual decision making. goals in diabetes: pragmatic randomised controlled trial. 2011; 101 (3):539–45. 2014; :g5651. 349 BMJ {published and unpublished data} Deschamps 2004 {published data only} Eden 2009 Deschamps MA, Taylor JG, Neubauer SL, Whiting S, Green Eden KB, Dolan JG, Perrin NA, Kocaoglu D, Anderson N, Case J, et al. Patients were more consistent in K. Impact of pharmacist consultation versus a decision randomized trial at prioritizing childbirth preferences using aid on decision making regarding hormone replacement graphic-numeric than verbal formats. Journal of Clinical 12 2004; International Journal of Pharmacy Practice therapy. 2009; Epidemiology (4):415–24. 62 (1):21–8. 41 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

45 Eden 2014 {published data only} {published data only} Fiks 2013a Eden KB, Perrin NA, Vesco KK, Guise JM. A randomized Fiks AG, Grundmeier RW, Mayne S, Song L, Feemster comparative trial of two decision tools for pregnant women K, Karavite D, et al. Effectiveness of decision support with prior cesareans. Journal of Obstetric, Gynecologic, & for families, clinicians, or both on HPV vaccine receipt. Neonatal Nursing 43 :568–79. 2014; 131 2013; Pediatrics (6):1114–24. {published data only} Eden 2015 Flood 1996 {published data only} Eden KB, Perrin NA, Hanson GC, Messing JT, Bloom TL, Flood AB, Wennberg JE, Nease RF Jr, Fowler FJ Jr, Ding Campbell JC, et al. Use of online safety decision aid by J, Hynes LM. The importance of patient preference in American Journal of Preventive Medicine abused women. the decision to screen for prostate cancer. Prostate Patient (4):372–83. 48 2015; Outcomes Research Team. Journal of General Internal 1996; Medicine (6):342–9. 11 {published data only} Edwards 2012 Edwards JA, Snyder FJ, Allen PM, Makinson KA, Hamby Francis 2009 {published data only} DM. Decision making for risk management: a comparison Francis NA, Butler CC, Hood K, Simpson S, Wood F, of graphical methods for presenting quantitative uncertainty. Nuttall J. Effect of using an interactive booklet about (12):2055–70. 32 2012; Risk Analysis childhood respiratory tract infections in primary care {published data only} El-Jawahri 2010 consultations on reconsulting and antibiotic prescribing: a El-Jawahri A, Podgurski LM, Eichler AF, Plotkin SR, :b2885. BMJ 339 cluster randomised controlled trial. 2009; Temel JS, Mitchell SL. Use of video to facilitate end-of- {published data only} Fraval 2015 life discussions with patients with cancer: a randomized Fraval A, Chandrananth J, Chong YM, Tran P, Coventry Journal of Clinical Oncology controlled trial. 28 2010; (2): LS. Internet based patient education improves informed 305–10. consent for elective orthopaedic surgery: a randomized {published data only} Ellison 2008 16 2015; :14. BMC Musculoskeletal Disorders controlled trial. Ellison GL, Weinrich SP. A randomized trial comparing {published data only} Frosch 2001 web-based decision aids on prostate cancer knowledge for Frosch DL, Kaplan RM, Felitti V. Evaluation of two African-American men. Journal of the National Medical methods to facilitate shared decision making for men (10):1139–45. 100 2008; Association considering the prostate-specific antigen test. Journal of {published data only} Elwyn 2004 General Internal Medicine 16 (6):391–8. 2001; Elwyn G, Edwards A, Hood K, Robling M, Atwell C, Russell I, et al. Achieving involvement: process outcomes {published data only} Frosch 2003 from a cluster randomized trial of shared decision making Frosch DL, Kaplan RM, Felitti VJ. A randomized controlled skill development and use of risk communication aids in trial comparing internet and video to facilitate patient Family Practice general practice. (4):337–46. 2004; 21 education for men considering the prostate specific antigen test. Journal of General Internal Medicine 2003; (10): 18 {published data only} Emery 2007 781–7. Emery J, Morris H, Goodchild R, Fanshawe T, Prevost AT, Bobrow M, et al. The GRAIDS Trial: a cluster randomised {published data only} Frosch 2008b controlled trial of computer decision support for the Frosch DL, Legare F, Mangione CM. Using decision aids in British management of familial cancer risk in primary care. community-based primary care: a theory-driven evaluation Journal of Cancer 2007; 97 (4):486–93. Patient Education and with ethnically diverse patients. (3):490–6. Counseling 73 2008; {published data only} Emmett 2007 Emmett CL, Murphy DJ, Patel RR, Fahey T, Jones C, {published data only} Frosch 2011 Ricketts IW, et al. Decision-making about mode of delivery Frosch DL, Uy V, Ochoa S, Mangione CM. Evaluation of after previous caesarean section: development and piloting a behavior support intervention for patients with poorly of two computer-based decision aids. Health Expectations 2011; Archives of Internal Medicine controlled diabetes. 171 (2):161–72. 10 2007; (22):2011–7. {published data only} Feldman-Stewart 2006 {published data only} Frost 2009 Feldman-Stewart D, Brennenstuhl S, Brundage MD, Frost J, Shaw A, Montgomery A, Murphy DJ. Women’s Roques T. An explicit values clarification task: development views on the use of decision aids for decision making about and validation. Patient Education and Counseling 2006; 63 the method of delivery following a previous caesarean (3):350–6. section: qualitative interview study. BJOG: An International {published data only} Feldman-Stewart 2012 2009; Journal of Obstetrics & Gynaecology 116 (7):896–905. Feldman-Stewart D, Tong C, Siemens R, Alibhai S, Pickles T, Robinson J, Brundage MD. The impact of explicit values Fujiwara 2015 {published data only} clarification exercises in a patient decision aid emerges aft er Fujiwara H, Shimoda A, Ishikawa Y, Taneichi A, Ohashi the decision is actually made: evidence from a randomized M, Takahashi Y, et al. Effect of providing risk information 32 2012; Medical Decision Making controlled trial. (4): on undergoing cervical cancer screening: a randomized 616–26. controlled trial. Archives of Public Health 2015; 73 :7. 42 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

46 {published data only} Greenfield 1985 Garvelink 2013 {published data only} Greenfield S, Kaplan S, Ware JE Jr. Expanding patient Garvelink MM, ter Kuile MM, Fischer MJ, Louwé LA, involvement in care. Effects on patient outcomes. Annals of Hilders CG, Kroep JR, et al. Development of a decision aid 102 (4):520–8. 1985; Internal Medicine about fertility preservation for women with breast cancer in the Netherlands. Journal of Psychosomatic Obstetrics and {published data only} Griffith 2008a (4):170–8. 2013; 34 Gynecology Griffith JM, Lewis CL, Brenner AR, Pignone MP. The effect of offering different numbers of colorectal cancer screening {published data only} Genz 2012 test options in a decision aid: a pilot randomized trial. Genz J, Haastert B, Müller H, Verheyen F, Cole D, :4. 8 2008; BMC Medical Informatics and Decision Making Rathmann W, et al. Blood glucose testing and primary Griffith 2008b {published data only} prevention of Type 2 diabetes - evaluation of the effect Griffith JM, Fichter M, Fowler FJ, Lewis C, Pignone MP. of evidence-based patient information: a randomized Should a colon cancer screening decision aid include the controlled trial. Diabetic Medicine 2012; 29 (8):1011–20. option of no testing? A comparative trial of two decision Giordano 2014 {published data only} 2008; aids. BMC Medical Informatics and Decision Making Giordano A, Lugaresi A, Confalonieri P, Granella F, Radice 8 :10. D, Trojano M, et al. Implementation of the “Sapere {published data only} Gruppen 1994 Migliora” information aid for newly diagnosed people with Gruppen LD, Margolin J, Wisdom K, Grum CM. Outcome multiple sclerosis in routine clinical practice: a late-phas e bias and cognitive dissonance in evaluating treatment (9): controlled trial. Multiple Sclerosis Journal 2014; 20 decisions. 69 (10 Suppl):S57–9. 1994; Academic Medicine 1234–43. {published data only} Gummersbach 2015 {published and unpublished data} Goel 2001 Gummersbach E, in der Schmitten J, Mortsiefer A, Abholz Goel V, Sawka CA, Thiel EC, Gort EH, O’Connor AM. HH, Wegscheider K, Pentzek M. Willingness to participate Randomized trial of a patient decision aid for choice of in mammography screening - a randomized controlled Medical Decision surgical treatment for breast cancer. questionnaire study of responses to two patient information 21 (1):1–6. Making 2001; Deutsches Ärzteblatt leaflets with different factual content. (5):61–8. 112 2015; International {published data only} Graham 2000 {published data only} Hacking 2013 Graham W, Smith P, Kamal A, Fitzmaurice A, Smith N, Hacking B, Wallace L, Scott S, Kosmala-Anderson J, Hamilton N. Randomised controlled trial comparing Belkora J, McNeill A. Testing the feasibility, acceptability effectiveness of touch screen system with leaflet for and effectiveness of a ’decision navigation’ intervention providing women with information on prenatal tests. BMJ for early stage prostate cancer patients in Scotland - a 2000; 320 (7228):155–60. randomised controlled trial. 2013; Psycho-Oncology (5): 22 Gray 2009 {published data only} 1017–1024. Gray SW, O’Grady C, Karp L, Smith D, Schwartz JS, Hall 2007 {published data only} Hornik RC, et al. Risk information exposure and direct- Hall S, Chitty L, Dormandy E, Hollywood A, Wildschut to-consumer genetic testing for BRCA mutations among HIJ, Fortuny A, et al. Undergoing prenatal screening for women with a personal or family history of breast or ovarian Down’s syndrome: presentation of choice and information 2009; Cancer Epidemiology, Biomarkers & Prevention cancer. European Journal of Human Genetics in Europe and Asia. (4):1303–11. 18 15 (5):563–9. 2007; {published data only} Green 2001b Hall 2011 {published data only} Green MJ, McInerney AM, Biesecker BB, Fost N. Education Hall MJ, Manne SL, Winkel G, Chung DS, Weinberg DS, nt about genetic testing for breast cancer susceptibility: patie Meropol NJ. Effects of a decision support intervention on preferences for a computer program or genetic counselor. y decisional conflict associated with microsatellite instabilit (1):24–31. 103 2001; American Journal of Medical Genetics Cancer Epidemiology, Biomarkers and Prevention testing. 20 (2):249–54. 2011; {published data only} Green 2004 Hamann 2014 {published data only} Green MJ, Peterson SK, Baker MW, Friedman LC, Harper Hamann J, Maris N, Iosifidou P, Mendel R, Cohen R, GR, Rubinstein WS, et al. Use of an educational computer Wolf P, Kissling W. Effects of a question prompt sheet on program before genetic counseling for breast cancer active patient behaviour: a randomized controlled trial susceptibility: effects on duration and content of counseling International Journal of Social with depressed outpatients. (4):221–9. 2005; sessions. Genetics in Medicine 7 (3):227–35. Psychiatry 60 2014; ∗ Green MJ, Peterson SK, Baker MW, Harper GR, Friedman {published data only} Harmsen 2014 LC, Rubinstein WS, et al. Effect of a computer-based Harmsen CG, Kristiansen IS, Larsen PV, Nexøe J, Støvring decision aid on knowledge, perceptions, and intentions H, Gyrd-Hansen D, et al. Communicating risk using about genetic testing for breast cancer susceptibility: a absolute risk reduction or prolongation of life formats: 2004; JAMA randomized controlled trial. (4):442–52. 292 43 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

47 cluster-randomised trial in general practice. British Journal {published data only} Hoffman 2009 64 (621):e199–207. of General Practice 2014; Hoffman RM, Walter LC. Colorectal cancer screening in the elderly: the need for informed decision making. Journal Harwood 2011 {published data only} 2009; (12):1336–7. 24 of General Internal Medicine Harwood R, Douglas C, Clark D. Decision aids for breast and nodal surgery in patients with early breast cancer: {published data only} Holbrook 2007 Asia-Pacific Journal of development and a pilot study. Holbrook A, Labiris R, Goldsmith CH, Ota K, Harb S, Clinical Oncology :114–22. 7 2011; Sebaldt RJ. Influence of decision aids on patient preferences CMAJ for anticoagulant therapy: a randomized trial. 2007; Healton 1999 {published data only} 176 (11):1583–7. Healton C, Taylor S, Messeri P, Weinberg G, Bamji M. Hollen 2013 {published data only} Effects of ZDV-based patient education on intentions Hollen PJ, Tyc VL, Donnangelo SF, Shannon SV, toward ZDV use, HIV testing and reproduction among a O’Laughlen MC, Hinton I, et al. A substance use decision 1999; 11 (6):675–86. US cohort of women. AIDS Care d aid for medically at-risk adolescents: results of a randomize {published data only} Henderson 2013 Cancer controlled trial for cancer-surviving adolescents. Henderson C, Brohan E, Clement S, Williams P, Lassman Nursing 2013; 36 (5):355–67. F, Schauman O, et al. Decision aid on disclosure of mental health status to an employer: feasibility and outcomes of {published data only} Holloway 2003 a randomised controlled trial. British Journal of Psychiatry Holloway RM, Wilkinson C, Peters TJ, Russell I, 2013; 203 (5):350–7. Cohen D, Hale J, et al. Cluster-randomised trial of risk communication to enhance informed uptake of cervical {published data only} Herrera 1983 2003; British Journal of General Practice (493): screening. 53 Herrera AJ, Cochran B, Herrera A, Wallace B. Parental 620–5. information and circumcision in highly motivated couples 71 Pediatrics (2):233–4. with higher education. 1983; Holmes-Rovner 2011 {published data only} Holmes-Rovner M, Kelly-Blake K, Dwamena F, Dontje K, {published data only} Hess 2015 Henry R, Olomu A, et al. Shared decision making guidance Hess LM, Litwiller A, Byron J, Stutsman J, Kasper K, Patient Education and reminders in practice (SDM-GRIP). Learman LA. Preference elicitation tool for abnormal Counseling 85 2011; (2):219–24. uterine bleeding treatment: a randomized controlled trial. 2015; (2): 8 The Patient: Patient Centered Outcomes Research Holt 2009 {published data only} 217–27. Holt CL, Wynn TA, Litaker MS, Southward P, Jeames S, Schulz E. A comparison of a spiritually based and non- Hewison 2001 {published data only} spiritually based educational intervention for informed Hewison J, Cuckle H, Baillie C, Sehmi I, Lindow S, decision making for prostate cancer screening among Jackson F, et al. Use of videotapes for viewing at home to Urologic Nursing church-attending African-American men. inform choice in Down syndrome screening: a randomised 29 2009; (4):249–58. controlled trial. Prenatal Diagnosis 2001; 21 (2):146–9. Hope 2010 {published data only} {published data only} Heyn 2013 Hope N, Rombauts L. Can an educational DVD improve Heyn L, Finset A, Eide H, Ruland CM. Effects of an the acceptability of elective single embryo transfer? A interactive tailored patient assessment on patient-clinician 2010; Fertility and Sterility 94 randomized controlled study. Psycho-Oncology 2013; communication in cancer care. 22 (2):489–95. (1):89–96. Huijbregts 2013 {published data only} {published data only} Hickish 1995 Huijbregts KML, de Jong FJ, van Marwijk HWJ, Beekman Hickish TF, Smith IE, Middleton G, Nicolson M. Patient ATF, Adèr HJ, Hakkaart-van Roijen L, et al. A target-driven preference for extended palliative chemotherapy for non- collaborative care model for major depressive disorder is 345 small cell lung cancer. Lancet 1995; (8953):857–8. effective in primary care in the Netherlands. A randomized {published data only} Hochlehnert 2006 clinical trial from the depression initiative. Journal of Hochlehnert A, Richter A, Bludau HB, Bieber C, Affective Disorders 2013; 146 :328–37. Blumenstiel K, Mueller K, et al. A computer-based Hunt 2005 {published data only} information-tool for chronic pain patients: computerized Hunt LM, de Voogd KB, Castaneda H. The routine and information to support the process of shared decision- the traumatic in prenatal genetic diagnosis: does clinical 61 Patient Education and Counseling (1): 2006; making. information inform patient decision-making?. Patient 92–8. 2005; Education and Counseling (3):302–12. 56 {published data only} Hofbauer 2008 Hofbauer GFL, Buhler RPN, French LE, Brockes M, {published data only} Hunter 1999 Scheuer E. Patient-centered care in dermatology: an online Hunter M, O’Dea I. An evaluation of a health education system that provides accessible and appropriate information intervention for mid-aged women: five year follow-up of to guide patients’ decision making. Archives of Dermatology effects upon knowledge, impact of menopause and health. 2008; 144 (9):1225–7. Patient Education and Counseling 1999; 38 (3):249–55. 44 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

48 on mental health literacy, help-seeking and symptoms. Hunter 2005 {published data only} (6):1071–9. 33 2003; Psychological Medicine Hunter AG, Cappelli M, Humphreys L, Allanson JE, Chiu TT, Peeters C, et al. A randomized trial comparing Kakkilaya 2011 {published data only} alternative approaches to prenatal diagnosis counseling in Kakkilaya V, Groome L, Platt D, Kurepa D, Pramanik A, 67 2005; Clinical Genetics advanced maternal age patients. Caldito G, et al. Use of a visual aid to improve counseling at (4):303–13. (6):e1511–9. the threshold of viability. Pediatrics 128 2011; {published data only} Huyghe 2009 Kaplan 2014a {published data only} Huyghe E, Martinetti P, Sui D, Schover LR. Banking on Kaplan CP, Livaudais-Toman J, Tice JA, Kerlikowske Fatherhood: pilot studies of a computerized educational K, Gregorich SE, Pérez-Stable EJ, et al. A randomized, Psycho- tool on sperm banking before cancer treatment. controlled trial to increase discussion of breast cancer in Oncology (9):1011–4. 18 2009; primary care. Cancer Epidemiology, Biomarkers & Prevention 2014; (7):1245–53. 23 {published data only} Ilic 2008 Ilic D, Egberts K, McKenzie JE, Risgridger G, Green {published data only} Kaplan 2014b S. Informing men about prostate cancer screening: a Kaplan AL, Crespi CM, Saucedo JD, Connor SE, Litwin randomized controlled trial of patient education materials. MS, Saigal CS. Decisional conflict in economically 23 (4):466–71. 2008; Journal of General Internal Medicine disadvantaged men with newly diagnosed prostate cancer. 2014; Cancer 120 (17):2721–7. {published data only} Isebaert 2007 Isebaert S, Van Audenhove C, Haustermans K, DeRidder {published data only} Kassan 2012 K, Junius S, Joniau S, et al. A decision aid for patients with Kassan EC, Williams RM, Kelly SP, Barry SA, Penek S, p localized prostate cancer: first results [Een beslissingshul Fishman MB, Cole CA, et al. Men’s use of an internet- voor patienten met gelokaliseerde prostaatkanker: eerste based decision aid for prostate cancer screening. Journal of (1):15–21. 63 2007; Tijdschrift voor Geneeskunde resultaten]. 2012; Health Communication 17 (6):677–97. {published data only} Kellar 2008 Jackson 2011 {published data only} Kellar I, Sutton S, Griffin S, Prevost AT, Kinmonth AL, Jackson C, Cheater FM, Harrison W, Peacock R, Bekker H, Marteau TM. Evaluation of an informed choice invitation West R, et al. Randomised cluster trial to support informed for type 2 diabetes screening. Patient Education and parental decision-making for the MMR vaccine. BMC Counseling (2):232–8. 2008; 72 Public Health 2011; 11 :475. Jerant 2007 {published data only} Kiatpongsan 2014 {published data only} Jerant A, Kravitz RL, Rooney M, Amerson S, Kreuter Kiatpongsan S, Carlson K, Feibelmann S, Sepucha K. M, Franks P. Effects of a tailored interactive multimedia Decision aid reduces misperceptions about hormone computer program on determinants of colorectal cancer Menopause: The therapy: a randomized controlled trial. screening: a randomized controlled pilot study in physician 21 Journal of The North American Menopause Society 2014; Patient Education and Counseling (1): offices. 2007; 66 (1):33–38. 67–74. Kobelka 2009 {published data only} Jibaja-Weiss 2006 {published data only} Kobelka C, Mattman A, Langlois S. An evaluation of the Jibaja-Weiss ML, Volk RJ, Granchi TS, Neff NE, Spann decision-making process regarding amniocentesis following SJ, Aoki N, et al. Entertainment education for informed Prenatal a screen-positive maternal serum screen result. breast cancer treatment decisions in low-literate women: Diagnosis (5):514–9. 29 2009; development and initial evaluation of a patient decision aid. Koelewijn-van Loon 2009 {published data only} 21 2006; Journal of Cancer Education (3):133–9. Koelewijn-van Loon MS, van der Weijden T, van Steenkiste {published data only} Joosten 2009 B, Ronda G, Winkens B, Severens JL, et al. Involving Joosten EA, de Jong CA, de Weert-van Oene GH, Sensky T, patients in cardiovascular risk management with nurse-led van der Staak CP. Shared decision-making reduces drug use clinics: a cluster randomized controlled trial. CMAJ 2009; and psychiatric severity in substance-dependent patients. (12):E267–74. 181 Psychotherapy and Psychosomatics 2009; 78 :245–53. Köpke 2009 {published data only} Köpke S, Kasper J, Mühlhauser I, Nübling M, Heesen C. Joosten 2011 {published data only} Patient education program to enhance decision autonomy Joosten EA, De Jong CA, de Weert-van Oene GH, Sensky in multiple sclerosis relapse management: a randomized- T, van der Staak CP. Shared decision-making: increases (1):96–104. 15 2009; Multiple Sclerosis controlled trial. autonomy in substance-dependent patients. Substance Use and Misuse 2011; 48 :1037–48. {published data only} Köpke 2014 {published data only} Jorm 2003 Köpke S, Kern S, Ziemssen T, Berghoff M, Kleiter I, Jorm AF, Griffiths KM, Christensen H, Korten AE, Marziniak M, et al. Evidence-based patient information Parslow RA, Rodgers B. Providing information about the programme in early multiple sclerosis: a randomised effectiveness of treatment options to depressed people in controlled trial. Journal of Neurology, Neurosurgery, and the community: a randomized controlled trial of effects Psychiatry 2014; 85 (4):411–18. 45 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

49 Krawczyk 2012 {published data only} {published data only} Lancaster 2009 Krawczyk A. Cancer Prevention and the Human Lancaster T. Physician training in the use of a decision aid Papillomavirus Vaccine: Psychosocial and Behavioural Fac tors increased patient participation in decision making for CVD . Involved in Vaccination Decision-making [PhD thesis] prevention. 2009; (1):24. 14 Evidence-Based Medicine Montreal: McGill Library, 2012. Landrey 2013 {published data only} Landrey AR, Matlock DD, Andrews L, Bronsert M, {published data only} Kripalani 2007 Denberg T. Shared decision making in prostate-specific Kripalani S, Sharma J, Justice E, Justice J, Spiker C, antigen testing: the effect of a mailed patient flyer prior to Laufman LE, et al. Low-literacy interventions to promote an annual exam. Journal of Primary Care & Community discussion of prostate cancer: a randomized controlled trial. (1):67–74. Health 2013; 4 (2):83–90. American Journal of Preventive Medicine 2007; 33 Lazcano Ponce 2000 {published data only} {published data only} Krones 2008 Lazcano Ponce EC, Sloan NL, Winikoff B, Langer A, Krones T, Keller H, Becker A, Sonnichsen A, Baum E, Coggins C, Heimburger A, et al. The power of information Donner-Banzhoff N. The theory of planned behaviour in and contraceptive choice in a family planning setting in a randomized trial of a decision aid on cardiovascular risk 76 2000; Sexually Transmitted Infections Mexico. (4):277–81. (2): 78 2009; prevention. Patient Education and Counseling 169–76. {published data only} Legare 2003 Krones T, Keller H, Sönnichsen A, Sadowski EM, Baum E, Legare F, O’Connor AM, Graham ID, Wells GA, Jacobsen Wegscheider K, et al. Absolute cardiovascular disease risk MJ, Elmslie T, et al. The effect of decision aids on the and shared decision making in primary care: a randomized agreement between women’s and physicians’ decisional Annals of Family Medicine controlled trial. (3): 2008; 6 conflict about hormone replacement therapy. Patient 218–27. 2003; (2):211–21. Education and Counseling 50 {published data only} Kuppermann 2009 Leung 2004 {published data only} Kuppermann M, Norton ME, Gates E, Gregorich SE, Leung KY, Lee CP, Chan HY, Tang MH, Lam YH, Lee Learman LA, Nakagawa S, et al. Computerized prenatal A. Randomised trial comparing an interactive multimedia genetic testing decision-assisting tool: a randomized decision aid with a leaflet and a video to give information (1): 113 2009; Obstetrics & Gynecology controlled trial. Prenatal about prenatal screening for Down syndrome. 53–63. (8):613–8. Diagnosis 2004; 24 {published data only} Kurian 2009 Levin 2011 {published data only} Kurian B, Trivedi M, Grannemann B, Claassen C, Daly E, Levin W, Campbell D, McGovern K, Gau J, Kosty D, Sunderajan P. A computerized decision support system for Seeley J, Lewinsohn P. A computer-assisted depression depression in primary care. Primary Care Companion to the intervention in primary care. Psychological Medicine 2011; Journal of Clinical Psychiatry 2009; 11 (4):140–6. 41 (7):1373–83. {published data only} Labrecque 2010 Lewis 2003 {published data only} Labrecque M, Paunescu C, Plesu I, Stacey D, Legare F. Lewis CL, Pignone MP, Sheridan SL, Downs SM, Kinsinger Evaluation of the effect of a patient decision aid about LS. A randomized trial of three videos that differ in the vasectomy on the decision-making process: a randomized framing of information about mammography in women 40 (6):556–62. trial. Contraception 2010; 82 2003; Journal of General Internal Medicine to 49 years old. 18 (11):875–83. {published data only} LaCroix 1999 Lewis 2012 {published data only} LaCroix AZ, Newton KM, Buist DSM, Curry SJ, Scholes D, Anderson LA, et al. Population-based strategy for Lewis CL, Brenner AT, Griffith JM, Moore CG, Pignone MP. Two controlled trials to determine the effectiveness of improving informed decision making about hormone replacement therapy in managed care settings. Women’s a mailed intervention to increase colon cancer screening. North Carolina Medical Journal 2012; 73 (2):93–8. 1999; Health Issues 9 (6):306–18. Lopez-Jornet 2012 {published data only} Lairson 2011 {published data only} López-Jornet P, Camacho-Alonso F, Sanchez-Siles M. Lairson DR, Chan W, Chang YC, del Junco DJ, Vernon SW. Patient information preferences and behaviour in relation to Cost-effectiveness of targeted versus tailored interventions British Journal of Oral & Maxillofacial Surgery oral biopsies. to promote mammography screening among women 50 (8):e115–8. 2012; military veterans in the United States. Evaluation and 2011; 34 (2):97–104. Program Planning {published data only} Lukens 2013 Lukens JM, Solomon P, Sorenson SB. Shared decision- Lalonde 2006 {published data only} making for clients with mental illness: a randomized Lalonde L, O’Connor AM, Duguay P, Brassard J, Drake E, factorial survey. Research on Social Work Practice 2013; 23 Grover SA. Evaluation of a decision aid and a personal risk (6):694–705. profile in community pharmacy for patients considering options to improve cardiovascular health: the OPTIONS Lurie 2011 {published data only} 2006; International Journal of Pharmacy Practice pilot study. Lurie J, Spratt K, Blood E, Tosteson T, Tosteson A, 14 (1):51–62. Weinstein J. Effects of viewing an evidence based video 46 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

50 decision aid on patients’ treatment preferences for spine {published data only} Mazur 1994 Spine 2011; 36 (18):1501–4. surgery. Mazur DJ, Hickam DH. The effect of physician’s explanations on patients’ treatment preferences: five-year {published data only} Maisels 1983 Medical Decision Making 1994; 14 survival data. (3):255–8. Maisels MJ, Hayes B, Conrad S, Chez RA. Circumcision: {published data only} McCaffery 2007 the effect of information on parental decision making. McCaffery K, Irwig L, Bossuyt P. Patient decision aids to Pediatrics (3):453–5. 71 1983; support clinical decision making: evaluating the decision {published data only} Mancini 2006 Medical Decision Making or the outcomes of the decision. Mancini J, Santin G, Chabal F, Julian-Reynier C. Cross- 2007; 27 (5):619–25. cultural validation of the Decisional Conflict Scale in a {published data only} McGinley 2002 15 sample of French patients. Quality of Life Research 2006; McGinley AM. Effect of Web-based Computer-tailoring on (6):1063–8. Women’s Intention to Continue or Begin to Use Hormone Replacement Therapy to Lower their Risk for Osteoporosis [P hD Manne 2009 {published data only} . Philadelphia: University of Pennsylvania, 2002. thesis] Manne SL, Coups EJ, Markowitz A, Meropol NJ, Haller McGowan 2008 {published data only} D, Jacobsen PB, et al. A randomized trial of generic versus tailored interventions to increase colorectal cancer screening McGowan J, Hogg W, Campbell C, Rowan M. Just-in-time information improved decision-making in primary care: Annals of Behavioral among intermediate risk siblings. 2009; Medicine 37 (2):207–17. (11): 3 a randomized controlled trial. PLOS ONE 2008; e3785. Manns 2005 {published data only} {published data only} McInerney-Leo 2004 Manns B J, Taub K, Vanderstraeten C, Jones H, Mills C, McInerney-Leo A, Biesecker BB, Hadley DW, Kase RG, Visser M, et al. The impact of education on chronic kidney Giambarresi TR, Johnson E, et al. BRCA1/2 testing in disease patients” plans to initiate dialysis with self-care hereditary breast and ovarian cancer families: effectiveness 68 dialysis: a randomized trial. Kidney International 2005; of problem-solving training as a counseling intervention. (4):1777–83. 130 2004; (3): American Journal of Medical Genetics. Part A Markham 2003 {published data only} 221–7. Markham R, Smith A. Limits to patient choice: example {published data only} Mclaren 2012 from anaesthesia. BMJ 2003; 326 (7394):863–4. Mclaren PJ, Hyde MK, White KM. Exploring the role of gender and risk perceptions in people’s decisions to register Martin 2012 {published data only} as a bone marrow donor. Health Education Research 2011; Martin R, Brower M, Geralds A, Gallagher P, Tellinghuisen 27 (3):513–22. D. An experimental evaluation of patient decision aid {published data only} Meropol 2013 design to communicate the effects of medications on the Meropol NJ, Egleston BL, Buzaglo JS, Balshem A, Benson rate of progression of structural joint damage in rheumatoid AB 3rd, Cegala DJ, et al. A web-based communication aid Patient Education and Counseling 2012; 86 arthritis. (3): for patients with cancer: the CONNECT study. Cancer 329–34. 2013; 119 (7):1437–45. Maslin 1998 {published data only} {published data only} Michie 1997 ∗ Maslin AM, Baum M, Walker JS, A’Hern R, Prouse A. Michie S, Smith D, McClennan A, Marteau TM. Patient Shared decision-making using an interactive video disk decision making: An evaluation of two different methods NT Research system for women with early breast cancer. of presenting information about a screening test. British 3 (6):444–55. 1998; Journal of Health Psychology 1997; 2 (4):317–26. Maslin AM, Baum M, Walker JS, A’Hern R, Prouse A. Miller 2014a {published data only} Using an interactive video disk in breast cancer patient Miller MJ, Allison JJ, Cobaugh DJ, Ray MN, Saag KG. A 94 (44):4–10. support. Nursing Times 1998; group-randomized trial of shared decision making for non- {published data only} Matlock 2014 steroidal anti-inflammatory drug risk awareness: primary Matlock DD, Keech TA, McKenzie MB, Bronsert MR, Journal of Evaluation in Clinical results and lessons learned. Nowels CT, Kutner JS. Feasibility and acceptability of a 20 2014; Practice :638–48. decision aid designed for people facing advanced or terminal {published data only} Miller 2014b illness: a pilot randomized trial. Health Expectations 2014; Miller SM, Roussi P, Scarpato J, Wen KY, Zhu F, Roy G. (1):49–59. 17 Randomized trial of print messaging: the role of the partner and monitoring style in promoting provider discussions {published data only} Matloff 2006 about prostate cancer screening among African American Matloff ET, Moyer A, Shannon KM, Niendorf KB, Col Psycho-Oncology :404–11. 23 2014; men. NF. Healthy women with a family history of breast cancer: impact of a tailored genetic counseling intervention on risk {published data only} Mishel 2009 perception, knowledge, and menopausal therapy decision Mishel MH, Germino BB, Lin L, Pruthi RS, Wallen EM, (7):843–56. 2006; Journal of Women’s Health making. 15 Crandell J, et al. Managing uncertainty about treatment 47 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

51 decision making in early stage prostate cancer: a randomized {published data only} O’Cathain 2002 77 2009; Patient Education and Counseling clinical trial. (3): O’Cathain A, Walters SJ, Nicholl JP, Thomas KJ, Kirkham 349–59. M. Use of evidence based leaflets to promote informed choice in maternity care: randomised controlled trial in Mohammad 2012 {published data only} everyday practice. 324 (7338):643–6. 2002; BMJ Mohammad-Alizadeh-Charandabi S, Shahnazi M, Jahanbakhsh. Communicating contraceptive effectiveness: O’Connor 1996 {published data only} Journal of Caring Sciences a randomized controlled trial. O’Connor AM, Pennie RA, Dales RE. Framing effects 1 2012; (1):1–9. on expectations, decisions, and side effects experienced: the case of influenza immunization. Journal of Clinical {published data only} Molenaar 2001 (11):1721–6. 49 1996; Epidemiology Molenaar S, Sprangers MA, Rutgers EJ, Luiten EJ, Mulder J, Bossuyt PM, et al. Decision support for patients with O’Connor 1998a {published and unpublished data} early-stage breast cancer: effects of an interactive breast O’Connor AM, Tugwell P, Wells GA, Elmslie T, Jolly E, cancer CDROM on treatment decision, satisfaction, and Hollingworth G, et al. Randomized trial of a portable, (6): 19 2001; Journal of Clinical Oncology quality of life. self-administered decision aid for postmenopausal women 1676–87. Medical considering long-term preventive hormone therapy. Mulley 2006 {published data only} 1998; Decision Making 18 :295–303. Mulley AG Jr. Developing skills for evidence-based surgery: O’Connor 1999a {published data only} Surgical ensuring that patients make informed decisions. O’Connor AM, Wells GA, Tugwell P, Laupacis A, Elmslie 2006; (1):181–92. 86 Clinics of North America T, Drake E. The effects of an ’explicit’ values clarification {published data only} Myers 2005a exercise in a women’s decision aid regarding postmenopausal Myers RE, Daskalakis C, Cocroft J, Kunkel EJ, Delmoor 2 1999; Health Expectations :21–32. hormone therapy. E, Liberatore M, et al. Preparing African-American men in O’Connor 2009a {published data only} community primary care practices to decide whether or not O’Connor PJ, Sperl-Hillen J, Johnson PE, Rush WA, Crain to have prostate cancer screening. Journal of the National AL. Customized feedback to patients and providers failed 97 2005; (8):1143–54. Medical Association to improve safety or quality of diabetes care: a randomized Myers 2005b {published data only} 2009; 32 trial. Diabetes Care (7):1158–63. Myers RE. Decision counseling in cancer prevention and {published data only} O’Connor 2011 control. 2005; 24 (4 Suppl):S71–7. Health Psychology Connor PJ, Sperl-Hillen JM, Rush WA, Johnson PE, Myers 2007 {published data only} Amundson GH, Asche SE, et al. Impact of electronic Myers RE, Sifri R, Hyslop T, Rosenthal M, Vernon SW, health record clinical decision support on diabetes care: a Cocroft J, et al. A randomized controlled trial of the impact 2011; randomized trial. (1): Annals of Family Medicine 9 of targeted and tailored interventions on colorectal cancer 12–21. 110 2007; Cancer (9):2083–91. screening. {published data only} Owens 2014A Myers 2011 {published data only} Owens OL. A Community-driven Approach to the Myers RE, Daskalakis C, Kunkel EJ, Cocroft JR, Riggio Development of a Digital Decision Aid to Facilitate Informe d JM, Capkin M, et al. Mediated decision support in prostate Decision Making for Prostate Cancer Screening among cancer screening: a randomized controlled trial of decision African-American men in Communities of Faith [PhD thesis] . 83 2011; Patient Education and Counseling counseling. (2): Columbia: University of South Carolina, 2014. 240–6. {published data only} Patanwala 2011 Myers 2013 {published data only} Patanwala IM, Brocklebank V, Inglis J, Trewby PN. A Myers RE, Bittner-Fagan H, Daskalakis C, Sifri R, Vernon randomized questionnaire-based study on the impact of SW, Cocroft J, et al. A randomized controlled trial of a providing numerical information on colorectal cancer tailored navigation and a standard intervention in colorectal screening. Journal of the Royal Society of Medicine Short Cancer Epidemiology, Biomarkers & cancer screening. 2011; (6):48. 2 Reports Prevention 2013; (1):109–17. 22 {published data only} Patel 2014 {published data only} Neubeck 2008 Patel S, Ngunjiri A, Wan Hee S, Yang Y, Brown S, Friede Neubeck L, Redfern J, Briffa T, Bauman A, Hare D, T, et al. Primum non nocere: shared informed decision Freedman SB. The CHOICE (Choice of Health Options In making in low back pain - a pilot cluster randomised trial. dy prevention of Cardiovascular Events) replication trial: stu :282. 15 BMC Musculoskeletal Disorders 2014; :25. 2008; BMC Cardiovascular Disorders protocol. 8 {published data only} Newton 2001 Pearson 2005 {published data only} Newton KM, LaCroix AZ, Buist DS, Delaney KM, Pearson S, Maddern GJ, Hewett P. Interacting effects of Anderson LA. Women’s responses to a mailed hormone preoperative information and patient choice in adaptation Menopause 2001; 8 (5): replacement therapy workbook. Diseases of the Colon & Rectum 2005; 48 to colonoscopy. 361–7. (11):2047–54. 48 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

52 {published data only} Peele 2005 {published data only} Raynes-Greenow 2009 Raynes-Greenow CH, Roberts CL, Nassar N, Trevena Peele PB, Siminoff LA, Xu Y, Ravdin PM. Decreased use of L. Do audio-guided decision aids improve outcomes? A adjuvant breast cancer therapy in a randomized controlled randomized controlled trial of an audio-guided decision aid trial of a decision aid with individualized risk information. compared with a booklet decision aid for Australian women 2005; Medical Decision Making 25 (3):301–7. considering labour analgesia. 12 2009; Health Expectations {published data only} Petty 2014 (4):407–16. Petty J. Exploring the effectiveness of an interactive, {published data only} Raynes-Greenow 2010 technology-enabled learning tool to enhance knowledge Raynes-Greenow CH, Nassar N, Torvaldsen S, Trevena L, for neonatal nurses. Neonatal, Paediatric and Child Health Roberts CL. Assisting informed decision making for labour Nursing 2014; 17 (1):2–10. analgesia: a randomised controlled trial of a decision aid for Philip 2010 {published data only} BMC Pregnancy and labour analgesia versus a pamphlet. Philip E, DuHamel K, Jandorf L. Evaluating the impact of Childbirth :15. 2010; 10 an educational intervention to increase CRC screening rates Rimer 2001 {published data only} in the African American community: a preliminary study. Rimer BK, Halabi S, Sugg Skinner C, Kaplan EB, Crawford 2010; Cancer Causes Control (10):1685–91. 21 Y, Samsa GP, et al. The short-term impact of tailored Phillips 1995 {published data only} mammography decision-making interventions. Patient Phillips C, Hill BJ, Cannac C. The influence of video 43 (3):269–85. Education and Counseling 2001; imaging on patients’ perceptions and expectations. Angle {published data only} Rimer 2002 1995; (4):263–70. Orthodontist 65 Rimer BK, Halabi S, Sugg Skinner C, Lipkus IM, Strigo Pignone 2013 {published data only} TS, Kaplan EB, et al. Effects of a mammography decision- Pignone MP, Howard K, Brenner AT, Crutchfield TM, American making intervention at 12 and 24 months. Hawley ST, Lewis CL, et al. Comparing 3 techniques for Journal of Preventive Medicine (4):247–57. 2002; 22 eliciting patient values for decision making about prostate- {published data only} Robinson 2013 specific antigen screening: a randomized controlled trial. Robinson JK, Gaber R, Hultgren B, Eilers S, Blatt H, 173 2013; JAMA Internal Medicine (5):362–8. Stapleton J, et al. Skin self-examination education for early Pinto 2008 {published data only} detection of melanoma: a randomized controlled trial of Pinto H, Rumball D, Maskrey V, Holland R. A pilot study Journal of internet, workbook and in-person interventions. for a randomized controlled and patient preference trial of 2013; Medical Internet Research 16 (1):1–11. buprenorphine versus methadone maintenance treatment in Ronda 2014 {published and unpublished data} the management of opiate dependent patients. Journal of Ronda G, Grispen JEJ, Ickenroth M, Dinant GJ, de Vries 2008; 13 (2):73–82. Substance Use NK, van der Weijden T. The effects of a web-based decision Powers 2011 {published data only} aid on the intention to diagnostic self-testing for cholesterol Powers B, Danus S, Grubber J, Olsen M, Oddone E, and diabetes: a randomized controlled trial. BMC Public Bosworth H. The effectiveness of personalized coronary Health :921. 14 2014; American heart disease and stroke risk communication. {published data only} Rostom 2002 Heart Journal 161 2011; (4):673–80. Rostom A, O’Connor A, Tugwell P, Wells G. A randomized Proctor 2006 {published data only} trial of a computerized versus an audio-booklet decision Proctor A, Jenkins TR, Loeb T, Elliot M, Ryan A. Patient aid for women considering post-menopausal hormone satisfaction with 3 methods of postpartum contraceptive Patient Education and Counseling replacement therapy. counseling: a randomized, prospective trial. Journal of 2002; 46 (1):67–74. (5):377–82. 2006; Reproductive Medicine 51 {published data only} Roter 2012 Prunty 2008 {published data only} Roter DL, Wexler R, Naragon P, Forrest B, Dees J, Prunty MC, Sharpe L, Butow P, Fulcher G. The motherhood Almodovar A, Wood J. The impact of patient and physician choice: a decision aid for women with multiple sclerosis. computer mediated communication skill training on Patient Education and Counseling 2008; 71 (1):108–15. reported communication and patient satisfaction. Patient {published data only} Ranta 2015 Education and Counseling 2012; 88 (3):406–13. Ranta A, Dovey S, Weatherall M, O’Dea D, Gommans Rothert 1997 {published and unpublished data} J, Tilyard M. Cluster randomized controlled trial of TIA Holmes-Rovner M, Kroll J, Rovner DR, Schmitt N, American electronic decision support in primary care. Rothert M, Padonu G, et al. Patient decision support (15):1545–51. 2015; Academy of Neurology 84 intervention: increased consistency with decision analytic 1999; 37 Medical Care models. (3):270–84. Rapley 2006 {published data only} ∗ Rothert ML, Holmes-Rovner M, Rovner D, Kroll J, Rapley T, May C, Heaven B, Murtagh M, Graham R, Breer L, Talarczyk G, et al. An educational intervention Kaner EF, et al. Doctor-patient interaction in a randomised as decision support for menopausal women. Research in controlled trial of decision-support tools. Social Science & Nursing & Health 1997; 20 (5):377–87. (9):2267–78. 2006; 62 Medicine 49 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

53 Schapira 2000 {published data only} Rovner 2004 {published data only} Schapira MM, VanRuiswyk J. The effect of an illustrated Rovner DR, Wills CE, Bonham V, Williams G, Lillie J, pamphlet decision-aid on the use of prostate cancer Kelly-Blake K, et al. Decision aids for benign prostatic (5): 49 2000; Journal of Family Practice screening tests. hyperplasia: applicability across race and education. Medical 418–24. 2004; (4):359–66. 24 Decision Making Schapira 2007 {published data only} Rubinstein 2011 {published data only} Schapira MM, Gilligan MA, McAuliffe T, Garmon G, Rubinstein W, Acheson L, O’Neill S, Ruffin M, Wang C, Carnes M, Nattinger AB. Decision-making at menopause: Beaumont J, et al. Clinical utility of family history for a randomized controlled trial of a computer-based hormone cancer screening and referral in primary care: a report from Patient Education and Counseling therapy decision-aid. Genetics in Medicine the Family Healthware Impact Trial. 67 2007; (1-2):100–7. 2011; (11):956–65. 13 {published data only} Schwartz 2009b Ruddy 2009 {published data only} Schwartz LM, Woloshin S, Welch HG. Using a drug Ruddy KJ, Partridge AH. Breast cancer in young women: facts box to communicate drug benefits and harms: two Oncology clinical decision-making in the face of uncertainty. (8): Annals of Internal Medicine 2009; randomized trials. 150 23 2009; (6):474–7. 516–27. Ruehlman 2012 {published data only} {published data only} Sears 2007 Ruehlman LS, Karoly P, Enders C. A randomized controlled Sears SR, Woodward JT, Twillman RK. What do I have to evaluation of an online chronic pain self management lose? effects of a psycho-educational intervention on cancer 2012; (2):319–30. 153 program. Pain Journal of Behavioral patient preference for resuscitation. (6):533–44. 30 2007; Medicine Ruland 2013 {published data only} {published data only} Sequist 2011 Ruland CM, Andersen T, Jeneson A, Moore S, Grimsbø Sequist T, Zaslavsky A, Colditz G, Ayanian J. Electronic GH, Børøsund E, Ellison MC. Effects of an internet support patient message to promote colorectal cancer screening. system to assist cancer patients in reducing symptom 2011; (7):636–41. Archives of Internal Medicine 171 distress: a randomized controlled trial. Cancer Nursing 36 2013; (1):6–17. {published data only} Shah 2012 Shah S, Singh K, Ali MK, Mohan V, Kadir MM, {published data only} Ryser 2004 Unnikrishnan AG, et al. CARRS Trial Writing Group. Ryser FG. Breastfeeding attitudes, intention, and initiation Improving diabetes care: multi-component cardiovascular in low-income women: the effect of the best start program. disease risk reduction strategies for people with diabetes 2004; 20 (3):300–5. Journal of Human Lactation in South Asia - the CARRS multi-center translation trial. {published data only} Sassen 2014 (2):285–94. 2012; Diabetes Research and Clinical Practice 98 Sassen B, Kok G, Schepers J, Vanhees L. Supporting health {published data only} Sheppard 2012 care professionals to improve the processes of shared decisi on Sheppard VB, Wallington SF, Williams KP, Lucas W. A making and self-management in a web-based intervention: decision-support intervention for black women eligible Journal of Medical Internet randomized controlled trial. for adjuvant systematic therapy: Sisters informing sisters (10):e211. Research 2014; 16 about breast cancer treatment - An intervention to reduce {published data only} Saver 2007 treatment disparities. In: Elk R, Landrine H editor(s). Saver BG, Gustafson D, Taylor TR, Hawkins RP, Woods Cancer Disparities: Causes and Evidence-Based Solutions . NF, Dinauer S, et al. A tale of two studies: the importance of American Cancer Society, 2012. setting, subjects and context in two randomized, controlled {published data only} Sheridan 2004 trials of a web-based decision support for perimenopausal Sheridan SL, Felix K, Pignone MP, Lewis CL. Information and postmenopausal health decisions. Patient Education and needs of men regarding prostate cancer screening and 66 Counseling 2007; (2):211–22. Patient Education and the effect of a brief decision aid. (3):345–51. 54 2004; Counseling {published data only} Sawka 2011 Sawka AM, Straus S, Gafni A, Brierley JD, Tsang RW, Sheridan 2010 {published data only} Rotstein L, et al. How can we meet the information Sheridan SL, Griffith JM, Behrend L, Gizlice Z, Jianwen C, needs of patients with early stage papillary thyroid cancer Pignone MP. Effect of adding a values clarification exercise Clinical considering radioactive iodine remnant ablation?. to a decision aid on heart disease prevention: a randomized 2011; Endocrinology :419–23. 74 30 2010; Medical Decision Making (4):E28–39. trial. {published data only} Scaffidi 2014 Sheridan 2012 {published data only} Scaffidi RM, Posmontier B, Bloch JR, Wittmann-Price R. Sheridan SL, Golin C, Bunton A, Lykes JB, Schwartz B, The relationship between personal knowledge and decision McCormack L, Driscoll D, Bangdiwala SI, Harris RP. self-efficacy in choosing trial of labor after cesarean. Journal Shared decision making for prostate cancer screening: (3):246–53. 2014; of Midwifery & Women’s Health 59 the results of a combined analysis of two practice-based 50 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

54 randomized controlled trials. BMC Medical Informatics & {published data only} Sorenson 2004 Decision Making 2012; 12 :130. Sorenson JR, Lakon C, Spinney T, Jennings-Grant T. Assessment of a decision aid to assist genetic testing research {published data only} Sherman 2014 Genetic Testing participants in the informed consent process. Sherman KA, Harcourt DM, Lam TC, Shaw LK, Boyages (3):336–46. 8 2004; J. BRECONDA: Development and acceptability of an {published data only} Sparano 2006 interactive decisional support tool for women considering Sparano JA. TAILORx: trial assigning individualized :835–8. 23 breast reconstruction. 2014; Psycho-Oncology options for treatment (Rx). 2006; Clinical Breast Cancer 7 Shirai 2012 {published data only} (4):347–50. Shirai Y, Fujimori M, Ogawa A, Yamada Y, Nishiwaki {published data only} Stalmeier 2009 Y, Ohtsu A, Uchitomi Y. Patients’ perception of the Stalmeier PF, Roosmalen MS. Concise evaluation of usefulness of a question prompt sheet for advanced cancer (1): decision aids. Patient Education and Counseling 2009; 74 patients when deciding the initial treatment: a randomized, 104–9. Psycho-Oncology (7):706–713. controlled trial. 2012; 21 {published data only} Starosta 2015 Starosta AJ, Luta G, Tomko CA, Schwartz MD, Taylor KL. {published data only} Silver 2012 Baseline attitudes about prostate cancer screening moderate Silver B, Zaman IF, Ashraf K, Majed Y, Norwood EM, the impact of decision aids on screening rates. Annals of Schuh LA, et al. A randomized trial of decision-making :762–768. 49 2015; Behavioral Medicine 78 Neurology in asymptomatic carotid stenosis. 2012; (5): 315–21. Stein 2013 {published data only} Stein RA, Sharpe L, Bell ML, Boyle FM, Dunn SM, Clarke Siminoff 2006 {published data only} SJ. Randomized controlled trial of a structured intervention Siminoff LA, Gordon NH, Silverman P, Budd T, Ravdin to facilitate end-of-life decision making in patients with PM. A decision aid to assist in adjuvant therapy choices for 31 (27): Journal of Clinical Oncology advanced cancer. 2013; breast cancer. Psycho-Oncology 2006; 15 (11):1001–13. 3403–10. Vickers AJ, Elkin EB, Peele PB, Dickler M, Siminoff LA. {published data only} Steiner 2003 Long-term health outcomes of a decision aid: data from Steiner MJ, Dalebout S, Condon S, Dominik R, Trussell a randomized trial of adjuvant! in women with localized J. Understanding risk: a randomized controlled trial of (4):461–7. breast cancer. Medical Decision Making 2009; 29 Obstetrics & communicating contraceptive effectiveness. {published data only} Simon 2012a Gynecology 102 (4):709–17. 2003; Simon D, Kriston L, von Wolff A, Buchholz A, Vietor C, Stephens 2008 {published data only} Hecke T, et al. Effectiveness of a web-based, individually Stephens RL, Xu Y, Volk RJ, Scholl LE, Kamin SL, Holden tailored decision aid for depression or acute low back EW. Influence of a patient decision aid on decisional conflict Patient Education and pain: a randomized controlled trial. related to PSA testing: a structural equation model. Health Counseling (3):360–8. 87 2012; 27 2008; Psychology (6):711–21. {published data only} Simon 2012b {published data only} Stiggelbout 2008 Simon W, Lambert MJ, Harris MW, Busath G, Vazquez A. Stiggelbout AM, Molewijk AC, Otten W, van Bockel Providing patient progress information and clinical support JH, Bruijninckx CM, van der Salm I, et al. The impact tools to therapists: Effects on patients at risk of treament of individualized evidence-based decision support on Psychotherapy Research failure. 2012; 22 (6):638–47. aneurysm patients’ decision making, ideals of autonomy, and quality of life. 28 2008; (5): Medical Decision Making Smith 2011a {published data only} 751–62. Smith T, Dow L, Virago E, Khatcheressian J, Matsuyama R, {published data only} Stirling 2012 Lyckholm L. A pilot trial of decision aids to give truthful Stirling C, Leggett S, Lloyd B, Scott J, Blizzard L, Quinn prognostic and treatment information to chemotherapy S, Robinson A. Decision aids for respite service choices by patients with advanced cancer. Journal of Supportive carers of people with dementia: development and pilot Oncology 2011; 9 (2):79–86. BMC Medical Informatics and Decision Making 2012; RCT. {published data only} Smith 2011b :21. 12 Smith SW, Nazione S, LaPlante C, Clark-Hitt R, Park HS, Street 1995 {published data only} Sung R, Leichtman A. Living kidney donor decision making Street RLJ, Voigt B, Geyer CJ, Manning T, Swanson GP. and communication. Journal of Health Communication: Increasing patient involvement in choosing treatment for 16 International Perspectives (8):870–88. 2011; Cancer 1995; (11):2275–85. 76 early breast cancer. {published data only} Solberg 2010 Street 1998 {published data only} Solberg LI, Asche SE, Sepucha K, Thygeson NM, Madden Street RL Jr, Van Order A, Bramson R, Manning T. JE, Morrissey L, et al. Informed choice assistance for women Preconsultation education promoting breast cancer making uterine fibroid treatment decisions: a practical screening: does the choice of media make a difference?. 2010; Medical Decision Making 30 clinical trial. (4):444–52. Journal of Cancer Education 1998; 13 (3):152–61. 51 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

55 Sundaresan 2011 {published data only} Journal of Health Care for the Poor and and screening. Sundaresan P, Turner S, Kneebone A, Pearse M, Butow P. 2013; Underserved (1):311–331. 24 Evaluating the utility of a patient decision aid for potential {published data only} Valdez 2001 participants of a prostate cancer trial (RAVES-TROG Valdez A, Banerjee K, Fernandez M, Ackerson L. Impact of (3):521–4. 08.03). Radiotherapy and Oncology 2011; 101 a multimedia breast cancer education intervention on use of Journal of Cancer mammography by low-income Latinas. {published data only} Tabak 1995 16 Education 2001; (4):221–4. Tabak N. Decision making in consenting to experimental cancer therapy. Cancer Nursing 1995; 18 (2):89–96. Van der Krieke 2013 {published data only} Van der Krieke L, Emerencia AC, Boonstra N, Wunderink {published data only} Taylor 2013 L, de Jonge P, Sytema S. A web-based tool to support shared Taylor KL, Williams RM, Davis K, Luta G, Penek S, Barry decision making for people with a psychotic disorder: S, et al. Decision making in prostate cancer screening using randomized controlled trial and process evaluation. Journal decision aids vs usual care: a randomized clinical trial. of Medical Internet Research 2013; 15 (10):e216. 173 JAMA Internal Medicine 2013; (18):1704–12. Ten 2008 {published data only} {published and unpublished data} Van Roosmalen 2004 Ten Wolde GB, Dijkstra A, van Empelen P, van den Van Roosmalen MS, Stalmeier PF, Verhoef LC, Hoekstra- Hout W, Neven AK, Zitman F. Long-term effectiveness Weebers JE, Oosterwijk JC, Hoogerbrugge N, et al. of computer-generated tailored patient education on Randomised trial of a decision aid and its timing for women Addiction benzodiazepines: a randomized controlled trial. British Journal of being tested for a BRCA1/2 mutation. 103 (4):662–70. 2008; 90 Cancer (2):333–42. 2004; ∗ Van Roosmalen MS, Stalmeier PF, Verhoef LC, Hoekstra- Thomas 2013 {published data only} Weebers JE, Oosterwijk JC, Hoogerbrugge N, et al. Thomas KL, Zimmer LO, Dai D, Al-Khatib SM, Allen Randomized trial of a shared decision-making intervention LaPointe NM, Peterson ED. Educational videos to reduce consisting of trade-offs and individualized treatment racial disparities in ICD therapy via innovative designs Journal of information for BRCA1/2 mutation carriers. American Heart (VIVID): a randomized clinical trial. (16):3293–301. 22 2004; Clinical Oncology (1):157–63. 166 2013; Journal Van Steenkiste 2008 {published data only} Thomson 2006 {published data only} Van Steenkiste B, van der Weijden TM, Stoffers JHEH, Thomson P, Dowding D, Swanson V, Bland R, Mair Grol RPTM. Patients’ responsiveness to a decision support C, Morrison A, et al. A computerised guidance tree tool for primary prevention of cardiovascular diseases in (decision aid) for hypertension, based on decision analysis: Patient Education and Counseling 2008; 72 (1): primary care. development and preliminary evaluation. European Journal 63–70. of Cardiovascular Nursing 2006; 5 (2):146–9. Van Til 2009 {published data only} {published data only} Thornton 1995 Van Til JA, Stiggelbout AM, IJzerman MJ. The effect of Thornton JG, Hewison J, Lilford RJ, Vail A. A randomised information on preferences stated in a choice-based conjoint trial of three methods of giving information about prenatal analysis. (2): 74 2009; Patient Education and Counseling 311 (7013):1127–30. testing. BMJ 1995; 264–71. Tiller 2006 {published data only} {published data only} Van Tol-Geerdink 2013 Tiller K, Meiser B, Gaff C, Kirk J, Dudding T, Phillips KA, Van Tol-Geerdink JJ, Leer JW, Weijerman PC, van Oort et al. A randomized controlled trial of a decision aid for IM, Vergunst H, van Lin EN, et al. Choice between women at increased risk of ovarian cancer. Medical Decision prostatectomy and radiotherapy when men are eligible for Making 26 (4):360–72. 2006; both: a randomized controlled trial of usual care vs decision Tinsel 2013 {published data only} BJU International aid. 2013; 111 (4):564–73. Tinsel I, Buchholz A, Vach W, Siegel A, Dürk T, Buchholz Veroff 2012 {published data only} A, et al. Shared decision-making in antihypertensive Veroff D, Sullivan L, Shoptaw EJ, Venator B, Ochoa-Arvelo therapy: a cluster randomised controlled trial. BMC Family T, Baxter J, et al. Improving self-care for heart failure for Practice :135. 2013; 14 seniors: Impact of video and written education and decision {published data only} Tomko 2015 aids. (1):37–45. 15 2012; Population Health Management Tomko C, Davis K, Ludin S, Kelly S, Stern A, Luta G, et {published data only} Volandes 2009 al. Decisional outcomes following use of an interactive Volandes AE, Paasche-Orlow MK, Barry MJ, Gillick MR, web-based decision aid for prostate cancer screening. Minaker KL, Chang Y, et al. Video decision support tool for earch Translational Behavioral Medicine: Practice, Policy, Res advance care planning in dementia: randomised controlled (2):189–97. 5 2015; BMJ :b2159. 338 2009; trial. {published data only} Ukoli 2013 Ukoli FA, Patel K, Hargreaves M, Beard K, Moton PJ, Bragg Volandes 2011 {published data only} R, et al. A tailored prostate cancer education intervention Volandes A, Ferguson L, Davis A, Hull N, Green M, Chang for low-income African Americans: impact on knowledge Y, et al. Assessing end-of-life preferences for advanced 52 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

56 {published data only} Warner 2015 dementia in rural patients using an educational video: a Warner DO, LeBlanc A, Kadimpati S, Vickers KS, Shi Y, randomised controlled trial. Journal of Palliative Medicine Montori V. Decision aid for cigarette smokers scheduled for 14 (2):169–77. 2011; elective surgery. (1):18–28. 123 2015; Anesthesiology {published data only} Volandes 2013 {published data only} Watts 2014 Volandes AE, Paasche-Orlow MK, Mitchell SL, El-Jawahri Watts KJ, Meiser B, Wakefield CE, Barratt AL, Howard K, A, Davis AD, Barry MJ, et al. Randomized controlled Cheah BC, et al. Online prostate cancer screening decision trial of a video decision support tool for cardiopulmonary Health Psychology aid for at-risk men: a randomized trial. Journal of resuscitation decision making in advanced care. 2014; (9):986–97. 33 (3):380–6. 31 Clinical Oncology 2013; {published data only} Welschen 2012 {published data only} Volk 2008 Welschen LM, Bot SD, Kostense PJ, Dekker JM, Volk RJ, Jibaja-Weiss ML, Hawley ST, Kneuper S, Spann Timmermans DR, van der Weijden T, et al. Effects of SJ, Miles BJ, et al. Entertainment education for prostate cardiovascular disease risk communication for patients cancer screening: a randomized trial among primary care with type 2 diabetes on risk perception in a randomized Patient Education and patients with low health literacy. Diabetes Care controlled trial: the @RISK study. : 2012; 35 (3):482–9. 2008; Counseling 73 2485–92. {published data only} Von Wagner 2011 {published data only} Wennberg 2010 Von Wagner C. A decision aid to support informed choice Wennberg DE, Marr A, Lang L, O’Malley S, Bennett G. A about bowel cancer screening in people with low educational randomized trial of a telephone care-management strategy. level improves knowledge but reduces screening uptake. 363 New England Journal of Medicine (13):1245–55. 2010; (2):36–7. 14 Evidence-Based Nursing 2011; Westermann 2013 {published data only} Wagner 1995 {published data only} Westermann GM, Verheij F, Winkens B, Verhulst FC, Van Wagner EH, Barrett P, Barry MJ, Barlow W, Fowler FJ Jr. Oort FV. Structured shared decision-making using dialogue The effect of a shared decision making program on rates Patient and visualization: a randomized controlled trial. of surgery for benign prostatic hyperplasia. Pilot results. 2013; Education and Counseling (1):74–81. 90 Medical Care 1995; 33 (8):765–70. {published data only} Weymann 2015 Wakefield 2008a {published data only} Weymann N, Dirmaier J, von Wolff A, Kriston L, Härter Wakefield CE, Meiser B, Homewood J, Ward R, O’Donnell M. Effectiveness of a web-based tailored interactive health S, Kirk J, et al. Randomized trial of a decision aid for communication application for patients with type 2 diabetes individuals considering genetic testing for hereditary or chronic low back pain: randomized controlled trial. nonpolyposis colorectal cancer risk. Cancer 2008; 113 (5): Journal of Medical Internet Research (3):e53 1-21. 17 2015; 956–65. {published data only} Wilhelm 2009 Wakefield 2008b {published data only} Wilhelm D, Gillen S, Wirnhier H, Kranzfelder M, Wakefield CE, Meiser B, Homewood J, Peate M, Taylor A, Schneider A, Scmidt A, et al. Extended preoperative patient Lobb E, et al. A randomized controlled trial of a decision education using a multimedia DVD: impact on patients aid for women considering genetic testing for breast and receiving a laparoscopic cholecystectomy: a randomised ovarian cancer risk. Breast Cancer Research and Treatment 394 controlled trial. Langenbeck’s Archives of Surgery 2009; (2):289–301. 107 2008; (2):227–33. Wakefield 2008c {published data only} {published data only} Wilkes 2013 Wakefield CE, Meiser B, Homewood J, Taylor A, Gleeson Wilkes MS, Day FC, Srinivasan M, Griffin E, Tancredi M, Williams R. A randomized trial of a breast/ovarian DJ, Rainwater JA, et al. Pairing physician education with cancer genetic testing decision aid used as a communication patient activation to improve shared decisions in prostate 2008; Psycho-Oncology aid during genetic counseling. (8): 17 cancer screening: a cluster randomized controlled trial. 844–54. (4):324–34. 2013; Annals of Family Medicine 11 {published data only} Wallston 1991 {published data only} Wilkie 2013 Wallston KA, Smith RA, King JE, Smith MS, Rye P, Burish Wilkie DJ, Gallo AM, Yao Y, Molokie RE, Stahl C, TG. Desire for control and choice of antiemetic treatment Hershberger PE, et al. Reproductive health choices for Western Journal of Nursing for cancer chemotherapy. young adults with sickle cell disease or trait: randomized (1):12–23. 13 Research 1991; controlled trial immediate posttest effects. Nursing Research (5):352–61. 62 2013; Wang 2004 {published data only} Wang C, Gonzalez R, Milliron KJ, Strecher VJ, Merajver Wilkins 2006 {published data only} SD. Genetic counseling for BRCA1/2: a randomized Wilkins EG, Lowery JC, Copeland LA, Goldfarb SL, Wren n controlled trial of two strategies to facilitate the educatio PA, Janz NK. Impact of an educational video on patient American Journal of Medical and counseling process. decision making in early breast cancer treatment. Medical Genetics. Part A (1):66–73. 134 2005; Decision Making 2006; 26 (6):589–98. 53 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

57 Qualitative Health Research (1): 22 2012; intervention. {published data only} Willemsen 2006 103–18. Willemsen MC, Wiebing M, van Emst A, Zeeman G. Helping smokers to decide on the use of efficacious smoking References to ongoing studies cessation methods: a randomized controlled trial of a decision aid. 2006; Addiction 101 (3):441–9. {published data only} ACTRN12615000523505 {published data only} Williamson 2014 ACTRN12615000523505. The Motherhood Choices Williamson LEA, Lawson KL, Downe PJ, Pierson RA. Decision Aid for Women with Rheumatoid Arthritis Informed reproductive decision-making: The impact of Increases Knowledge and Reduces Decisional Conflict: providing fertility information on fertility knowledge and A Randomized Controlled Study. http://apps.who.int/ Journals of Obstetrics and intentions to delay childbearing. trialsearch/Trial2.aspx?TrialID=ACTRN12615000523505 (5):400–5. 2014; Gynaecology Canada 36 (first received May 25, 2015). {published data only} Williams-Piehota 2008 {published data only} ACTRN12615000843550 Williams-Piehota PA, McCormack LA, Treiman K, Bann ACTRN12615000843550. Evaluation of decision aids CM. Health information styles among participants in for parents about the benefits and harms of antibiotic a prostate cancer screening informed decision-making use for coughs and colds in children [Pilot randomised 2008; (3): 23 Health Education Research intervention. controlled trial of decision aids for parents about the benefit 440–53. and harm of antibiotics for common acute respiratory Woltmann 2011 {published data only} infections in children to aid informed decision making]. Woltmann EM, Wilkniss SM, Teachout A, McHugo GJ, http://apps.who.int/trialsearch/Trial2.aspx?TrialID= Drake RE. Trial of an electronic decision support system ACTRN12615000843550 (first received August 13, 2015). to facilitate shared decision making in community mental {published data only} Al-Itejawi 2015 (1):54–60. 62 health. Psychiatric Services 2011; Al-Itejawi HH, van Uden-Kraan CF, Vis AN, Wroe 2005 {published data only} Nieuwenhuijzen JA, Hofstee MJ, van Moorselaar RJ, Wroe AL, Turner N, Owens RG. Evaluation of a decision- Verdonck-de Leeuw IM. Development of a patient decision making aid for parents regarding childhood immunizations. aid for the treatment of localised prostate cancer: a (6):539–47. 24 2005; Health Psychology J Clin Nurs participatory design approach. (7-8): 25 2016; 1131–1144. Yee 2014 {published data only} Yee LM, Wolf M, Mullen R, Bergeron AR, Cooper Bailey {published data only} Anderson 2014 S, Levine R, Grobman WA. A randomized trial of a prenatal Anderson RT, Montori VM, Shah ND, Ting HH, Pencille genetic testing interactive computerized information aid. LJ, Demers M, et al. Effectiveness of the Chest Pain Choice 2014; Prenatal Diagnosis 34 (6):552–7. decision aid in emergency department patients with low- risk chest pain: study protocol for a multicenter randomized Yun 2011 {published data only} 15 2014; trial. (166):1–11. Trials Yun YH, Lee MK, Park S, Lee JL, Park J, Choi YS, et al. Use of a decision aid to help caregivers discuss terminal disease Aslani 2014 {published data only} status with a family member with cancer: a randomized Aslani A, Tara F, Ghalichi L, Eslami S. The impact of Journal of Clinical Oncology 29 (36): controlled trial. 2011; computerized decision aid on mode of delivery - a study 4811–9. protocol. Studies in Health Technology and Informatics 2014; 200 :170–2. {published data only} Zajac 2012 Making Difficult Health Decisions: A Motivated Zajac LE. {published data only} Buhse 2013 Decision Processing Model [PhD thesis] . Pittsburgh: Buhse S, Heller T, Kasper J, Mühlhauser I, Müller UA, University of Pittsburgh, 2012. Lehmann T, Lenz M. An evidence-based shared decision making programme on the prevention of myocardial Zapka 2004 {published data only} infarction in type 2 diabetes: protocol of a randomised- Zapka JG, Lemon SC, Puleo E, Estabrook B, Luckmann controlled trial. 2013; 14 (155):1–8. BMC Family Practice R, Erban S. Patient education for colon cancer screening: a randomized trial of a video mailed before a physical Carroll 2012 {published data only} (9): 141 2004; examination. Annals of Internal Medicine Carroll SL, McGillion M, Stacey D, Healey JS, Browne 683–92. G, Arthur HM, Thabane L. Development and feasibility Zikmund-Fisher 2008 {published data only} testing of decision support for patients who are candidates Zikmund-Fisher BJ, Ubel PA, Smith DM, Derry HA, for a prophylactic implantable defibrillator: a study protoc ol McClure JB, Stark A, et al. Communicating side effect for a pilot randomized controlled trial. Trials 2013; 14 (346): risks in a tamoxifen prophylaxis decision aid: the debiasing 1–9. Patient Education and Counseling influence of pictographs. {published data only} Chambers 2008 (2):209–14. 73 2008; Chambers SK, Ferguson M, Gardiner RA, Nicol D, Gordon Zoffman 2012 {published data only} L, Occhipinti S, et al. ProsCan for men: randomised Zoffman V, Kirkevold M. Realizing empowerment controlled trial of a decision support intervention for men in difficult diabetes care: a guided self-determination with localised prostate cancer. BMC Cancer 2008; 8 :207. 54 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

58 Coylewright 2012 {published data only} Trials study protocol for a randomized controlled trial. Coylewright M, Shepel K, LeBlanc A, Pencille L, Hess E, (127):1–8. 2013; 14 Shah N, Montori VM, Ting HH. Shared decision making Mann 2012 {published data only} in patients with stable coronary artery disease: PCI choice. Mann DM, Lin JJ. Increasing efficacy of primary care- PLOS One 7 2012; (11):1–8. based counseling for diabetes prevention: rationale and design of the ADAPT (Avoiding Diabetes Thru Action Plan Cuypers 2015 {published data only} Targeting) trial. Implementation Science 2012; 7 :6. Cuypers M, Lamers RE, Kil PJ, van de Poll-Franse LV, de Vries M. Impact of a web-based treatment decision aid NCT00813033 {published data only} for early-stage prostate cancer on shared decision-making NCT00813033. Use of a Patient Decision Aid for and health outcomes: study protocol for a randomized Gastrologic Endoscopy in a Pediatric Setting [Creation and 2015; Trials controlled trial.. :231. 16 Pilot Evaluation of a Patient Decision Aid as an Adjunct {published data only} Den Ouden 2015 to the Consenting Process for Gastrointestinal Endoscopy Den Ouden H, Vos RC, Reidsma C, Rutten GEHM. in a Pediatric Setting]. https://clinicaltrials.gov/show/ Shared decision making in type 2 diabetes with a support NCT00813033 (first received December 19, 2008). decision tool that takes into account clinical factors, the NCT01077037 {published data only} intensity of treatment and patient preferences: design of a NCT01077037. Impact of a Decision Aid on Patient BMC Family Practice cluster randomised (OPTIMAL) trial. Decision Making in Emergency Department Chest Pain :27. 2015; 16 Patients [Impact of a Decision Aid on Patient Participation {published data only} Dirmaier 2013 in Decision Making and Resource Use in Low Risk Chest Dirmaier J, Härter M, Weymann N. A tailored, dialogue- Pain Patients: A Randomized Trial]. clinicaltrials.gov/show/ based health communication application for patients with NCT01077037 (first received February 24, 2010). chronic low back pain: study protocol of a randomised NCT01152294 {published data only} BMC Medical Informatics & Decision controlled trial. NCT01152294. Measuring Quality of Decisions About 13 2013; Making (66):1–9. Treatment of Menopausal Symptoms [Measuring Quality {published data only} Geiger 2011 of Decisions About Treatment of Menopausal Symptoms]. Geiger F, Liethmann K, Hoffmann F, Paschedag J, Kasper J. clinicaltrials.gov/show/NCT01152294 (first received June Investigating a training supporting Shared Decision Making 22, 2010). (IT’S SDM 2011): study protocol for a randomized NCT01152307 {published data only} :232. Trials 2011; 12 controlled trial. NCT01152307. Measuring Quality of Decisions About {published data only} Hersch 2014 Treatment of Depression [Measuring Quality of Decisions Hersch J, Barratt A, Jansen J, Houssami N, Irwig L, Jacklyn About Treatment of Depression]. clinicaltrials.gov/show/ G, et al. The effect of information about overdetectection NCT01152307 (first received June 22, 2010). of breast cancer on women’s decision-making about {published data only} NCT01447186 mammography screening: study protocol for a randomised NCT01447186. Adaptation of the American Cancer 2014; BMJ Open controlled trial. (5):1–10. 4 Society (ACS) Early Detection of Prostate Cancer Patient Hess 2014 {published data only} Decision Aid for Spanish Speaking Men [Adaptation of Hess EP, Wyatt KD, Kharbanda AB, Louie JP, Dayan PS, the American Cancer Society (ACS) Early Detection of Tzimenatos L, et al. Effectiveness of the head CT choice Prostate Cancer Patient Decision Aid for Spanish Speaking decision aid in parents of children with minor head trauma: Men]. clinicaltrials.gov/show/NCT01447186 (first received study protocol for a multicenter randomized trial. Trials October 3, 2011). 14 2014; (253):1–11. {published data only} NCT01618097 {published data only} Jimbo 2012 NCT01618097. Evaluation of DVD and Internet Decision Jimbo M, Kelly-Blake K, Sen A, Hawley ST, Ruffin MT Aids for Hip and Knee Osteoarthritis: Focus on Health 4th. Decision Aid to Technologically Enhance Shared Literacy. clinicaltrials.gov/show/NCT01618097 (first decision making (DATES): study protocol for a randomized received May 29, 2012). (381):1–16. 14 controlled trial. Trials 2013; NCT01713894 {published data only} Layton 2012 {published data only} NCT01713894. Decision Aid - Extreme Prematurity Effects of a web-based decision aid on African American [Utility of a Clinically Relevant Decision Aid, for Parents men’s prostate screening knowledge and behavior. Ongoing Facing Extremely Premature Delivery]. clinicaltrials.gov/ study -. show/NCT01713894 (first received October 22, 2012). {published data only} LeBlanc 2013 NCT01771536 {published data only} LeBlanc A, Bodde AE, Branda ME, Yost KJ, Herrin J, NCT01771536. The PCI Choice Trial: a Pilot Randomized Williams MD, et al. Translating comparative effectiveness Trial of a Decision Aid for Patients With Stable Coronary of depression medications into practice by comparing the Artery Disease. clinicaltrials.gov/show/NCT01771536 depression medication choice decision aid to usual care: (first received December 15, 2012). 55 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

59 NCT01851785 {published data only} {published data only} NCT02248974 NCT01851785. African American Preference for Knee NCT02248974. Development & Testing of a Decision Replacement: A Patient-Centered Intervention (ACTION) Aid for LVAD Placement (VADDA) [Development and [Behavioral & Social Science Research on Understanding User Testing of a Decision Aid for Left Ventricular Assist and Reducing Health Disparities]. clinicaltrials.gov/show/ Device (LVAD) Placement]. clinicaltrials.gov/show/ NCT01851785 (first received May 8, 2013). NCT02248974 (first received September 17, 2014). {published data only} NCT01941186 {published data only} NCT02259699 NCT01941186. A Family Centered Intervention to NCT02259699. Ovarian Cancer Patient-Centered Promote Optimal Child Development [A Family Centered Decision Aid (PCOA) [Ovarian Cancer Patient–Centered Intervention to Promote Optimal Child Development Decision Aid]. clinicaltrials.gov/show/NCT02259699 (first at the Interface of the Health System and Community]. received May 15, 2014). clinicaltrials.gov/show/NCT01941186 (first received NCT02308592 {published data only} September 9, 2013). NCT02308592. Patient Decision Aid for Antidepressant NCT01976325 {published data only} Use in Pregnancy [Patient Decision Aid (PDA) for NCT01976325. Evaluating the Ottawa Malaria Decision Antidepressant Use In Pregnancy]. clinicaltrials.gov/show/ Aid (OMDA) [Incorporation of the ’Ottawa Malaria NCT02308592 (first received December 2, 2014). Decision Aid’ Into the Pre–travel Consultation Process: NCT02319525 {published data only} Assessment of Travelers’ Knowledge, Decisional Conflict, NCT02319525. Individualized Patient Decision Making Preparation for Decision–making and Medication for Treatment Choices Among Minorities With Lupus Adherence Compared to Standard Care]. clinicaltrials.gov/ [Individualized Patient Decision Making for Treatment show/NCT01976325 (first received October 29, 2013). clinicaltrials.gov/ Choices Among Minorities With Lupus]. {published data only} NCT02026102 show/NCT02319525 (first received November 5, 2014). NCT02026102. A Pilot Trial of Patient Decision Aids {published data only} NCT02326597 for Implantable Cardioverter-Defibrillators (ICDs) [A NCT02326597. Decision Aid for Therapeutic Options Pilot Trial of Patient Decision Aids for Implantable In Sickle Cell Disease [Comparative Effectiveness of a Cardioverter–Defibrillators (ICDs)]. clinicaltrials.gov/ Decision Aid for Therapeutic Options in Sickle Cell show/NCT02026102 (first received December 12, 2013). Disease]. clinicaltrials.gov/show/NCT02326597 (first {published data only} NCT02084290 received December 18, 2014). NCT02084290. Evaluating a Shared Decision Making {published data only} NCT02344576 Program for Crohn’s Disease [Evaluating a Prediction Tool NCT02344576. Trial of a Decision Support Intervention and Decision Aid for Patients With Crohn’s Disease]. for Patients and Caregivers Offered Destination Therapy clinicaltrials.gov/show/NCT02084290 (first received Heart Assist Device (DECIDE-LVAD) [A Multicenter January 29, 2014). Trial of a Shared Decision Support Intervention for {published data only} NCT02110979 Patients and Their Caregivers Offered Destination Therapy NCT02110979. Validation of a Patient Decision Aid for for End–Stage Heart Failure]. clinicaltrials.gov/show/ Type 2 Diabetes [Validation of a Patient Decision Aid for NCT02344576 (first received January 16, 2015). Type 2 Diabetes]. clinicaltrials.gov/show/NCT02110979 {published data only} NCT02488317 (first received April 4, 2014). NCT02488317. Empowering Patients On Choices for NCT02145481 {published data only} Renal Replacement Therapy (Aim 3) (EPOCH-RRT) NCT02145481. Decisional Quality for Patients With (EPOCH-RRT) [Empowering Patients On Choices for Coronary Artery Disease (DeQCAD). clinicaltrials.gov/ Renal Replacement Therapy (Aim 3)]. clinicaltrials.gov/ show/NCT02145481 (first received May 15, 2014). show/NCT02488317 (first received June 15, 2015). {published data only} NCT02198690 {published data only} NCT02488603 NCT02198690. Trial of a Mammography Decision NCT02488603. Decision Aids for Tamoxifen Treatment Aid for Women Aged 75 and Older [Randomized Trial in Breast Cancer Patients [Utilization of Decision Aids of a Mammography Decision Aid for Women Aged 75 for Tamoxifen Treatment in Breast Cancer Patients: A and Older]. clinicaltrials.gov/show/NCT02198690 (first Randomized Controlled Trial.]. clinicaltrials.gov/show/ received July 16, 2014). NCT02488603 (first received June 25, 2015). {published data only} NCT02235571 {published data only} NCT02492009 NCT02235571. iChoose Decision Kidney Aid for End- NCT02492009. Patient Decision Aid for Antidepressant Use in Pregnancy [Patient Decision Aid (PDA) for Stage Renal Disease Patients [iChoose Kidney Decision Aid Antidepressant Use In Pregnancy: a Pilot RCT]. for Treatment Options Among End–Stage Renal Disease clinicaltrials.gov/show/NCT02492009 (first received June (ESRD) Patients]. clinicaltrials.gov/show/NCT02235571 22, 2015). (first received September 6, 2014). 56 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

60 Bennett 2010 NCT02503553 {published data only} Bennett C, Graham ID, Kristjansson E, Kearing SA, Clay NCT02503553. Decision Aids in Cerebral Aneurysm KF, O’Connor AM. Validation of a preparation for decision Treatment. clinicaltrials.gov/show/NCT02503553 (first (1): making scale. Patient Education and Counseling 78 2010; received July 17, 2015). 130–33. NCT02516449 {published data only} Brehaut 2003 NCT02516449. Assessment of Shared Decision Making Brehaut JC, O’Connor AM, Wood TJ, Hack TF, Siminoff Aids in Asthma [Utility of Two Patients Decision Aids L, Gordon E, et al. Validation of a decision regret scale. About Asthma Inhaled Controller Medication Use in 2003; 23 Medical Decision Making (4):281–92. Adult Patients With Asthma]. clinicaltrials.gov/show/ Brouwers 2010 NCT02516449 (first received July 2, 2015). Brouwers M, Stacey D, O’Connor A. Knowledge creation: {published data only} NCT02540044 182 2010; synthesis, tools and products. CMAJ (2):E68–72. NCT02540044. Supporting Patient Care With Electronic Brown 2015 Resource (SuPER) (SuPER) [Supporting Patient Care Brown JG, Joyce KE, Stacey D, Thomson MD. Patients or With Electronic Resource (SuPER): Efficacy of an Online volunteers? The impact of motivation for trial participation Decision Aid for Patients Considering Biologic Therapy on the efficacy of patient decision aids: a secondary analysis for Rheumatoid Arthritis]. clinicaltrials.gov/show/ Medical Decision Making of a Cochrane Systematic Review. NCT02540044 (first received September 1, 2015). (4):419–35. 35 2015; {published data only} NCT02611050 Charles 1997 NCT02611050. Treatment Decisions for Multi-vessel CAD Charles C, Gafni A, Whelan T. Shared decision-making in [Treatment Decisions for Multi–vessel Coronary Artery the medical encounter: what does it mean?. Social Science Disease Patients]. clinicaltrials.gov/show/NCT02611050 (5):681–92. and Medicine 1997; 44 (first received Noovember 9, 2015). Charles 2010 Oostendorp 2011 {published data only} Charles C, Gafni A, Freeman E. Implementing shared Oostendorp L, Ottevanger P, van der Graaf W, Stalmeier P. treatment decision making and treatment decision aids: a Assessing the information desire of patients with advanced Psicooncologia 2010; 7 cautionary tale. (2-3):243–55. cancer by providing information with a decision aid, which Clinical Evidence 2013 is evaluated in a randomized trial: a study protocol. BMC Clinical Evidence. How much do we know?. Available Medical Informatics & Decision Making 2011; 11 (9):1–9. y- from: clinicalevidence.bmj.com/x/set/static/cms/efficac Yu 2015 {published data only} categorisations.html (accessed 29 October 2013). Yu CH, Ivers NM, Stacey D, Rezmovitz J, Telner D, Coyne 2013 Thorpe K, et al. Impact of an interprofessional shared Coyne I, O’Mathuna DP, Gibson F, Shields L, Sheaf decision-making and goal-setting decision aid for patients G. Interventions for promoting participation in shared with diabetes on decisional conflict: study protocol for a Cochrane decision-making for children with cancer. (1):286. randomized controlled trial. Trials 2015; 16 Database of Systematic Reviews 2013, Issue 6. [DOI: 10.1002/14651858.CD008970.pub2] Additional references Degner 1992 Degner LF, Sloan JA. Decision making during serious Andrews 2013 Journal of illness: what role do patients really want to play. Andrews JC, Schünemann HJ, Oxman AD, Pottie K, 45 Clinical Epidemiology 1992; (9):941–50. Meerpohl JJ, Coello PA, et al. GRADE guidelines 15: Going from evidence to recommendation - determinants Duncan 2010 of a recommendation’s direction and strength. Journal of Duncan E, Best C, Hagen S. Shared decision making (7):726–35. [DOI: 10.1016/ 66 2013; Clinical Epidemiology interventions for people with mental health conditions. j.jclinepi.2013.02.003] 2010, Issue 1. Cochrane Database of Systematic Reviews [DOI: 10.1002/14651858.CD007297.pub2] Barry 2008 Barry MJ, Wescott PH, Reifler EJ, Chang Y, Moulton BW. Durand 2008 Reactions of potential jurors to a hypothetical malpractice Durand MA, Stiel M, Boivin J, Elwyn G. Where is the suit. Alleging failing to perform a prostate-specific-antigen theory? Evaluating the theoretical frameworks described test. Summer : Journal of Law, Medicine & Ethics 2008; Patient Education and in decision support technologies. 396–402. Counseling 2008; 71 (1):125–35. Durand 2014 Bekker 2003 Durand MA, Carpenter L, Dolan H, Bravo P, Mann M, Bekker HL, Legare F, Stacey D, O’Connor A, Lemyre L. Is Bunn F, et al. Do interventions designed to support shared anxiety a suitable measure of decision aid effectiveness: a decision-making reduce health inequalities? A systematic systematic review. Patient Education and Counselling 2003; review and meta-analysis. PLOS One 2014; 9 (4):1–14. 50 (3):255–62. 57 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

61 Hollinghurst 2010 Elwyn 2005 Hollinghurst S, Emmett C, Peters TJ, Watson H, Fahey T, Elwyn G, Hutchings H, Edwards A, Rapport F, Wensing Murphy DJ, et al. Economic evaluation of the DIAMOND M, Cheung WY, et al. The OPTION scale: measuring the randomized trial: cost and outcomes of 2 decision aids for extent that clinicians involve patients in decision-making mode of delivery among women with previous caesarian Health Expectations tasks. (1):34–42. 8 2005; :453–63. 2010; 30 BMJ section. Elwyn 2006 IPDAS 2005a Elwyn G, O’Connor A, Stacey D, Volk R, Edwards A, International Patient Decision Aid Standards Collaboration. Coulter A, et al. Developing a quality criteria framework for Background Document. 2005. ipdas.ohri.ca/ patient decision aids: online international Delphi consensus IPDAS ̇Background.pdf (accessed 29 Oct 2013). process. 333 (7565):417. BMJ 2006; IPDAS 2005b Elwyn 2013 International Patient Decision Aid Standards Collaboration. Elwyn G, I Scholl, Tietbohl C, Mann M, Edwards IPDAS Voting Document - 2nd Round. 2005. AGK, Clay C, et al. “Many miles to go...”: A systematic ipdas.ohri.ca/IPDAS ̇Second ̇Round.pdf (accessed 29 Oct review of the implementation of patient decision support 2013). BMC: Medical interventions into routine clinical practice. Informatics and Decision Making 2013; 13 (Suppl 2):S14. IPDAS 2013 Volk RJ, Llewellyn-Thomas H, Stacey D, Elwyn G. Ten Gentles 2013 years of the International Patient Decision Aid Standards Gentles SJ, Stacey D, Bennett C, Alshurafa M, Walter SD. collaboration: evolution of the core dimensions for Factors explaining the heterogeneity of effects of patient BMC: Medical assessing the quality of patient decision aids. decision aids on knowledge of outcome probabilities: a (Suppl 2):S1. 13 2013; Informatics and Decision Making systematic review sub-analysis. Systematic Reviews 2013; 2 : Joseph-Williams 2013 95. Joseph-Williams N, Newcombe R, Politi M, Durand MA, GRADEpro GDT [Computer program] Sivell S, Stacey D, et al. Toward minimum standards GRADE Working Group, McMaster University. for certifying patient decision aids: a modified Delphi GRADEpro GDT. Version accessed prior to 21 March (6): Medical Decision Making 2014; 34 consensus process. 2017. Hamilton (ON): GRADE Working Group, 699–710. [DOI: 10.1177/0272989X13501721] McMaster University, 2014. Kiesler 2006 Gravel 2006 Kiesler DJ, Auerbach SM. Optimal matches of patient Gravel K, Legare F, Graham ID. Barriers and facilitators to preferences for information, decision-making and implementing shared decision-making in clinical practice: interpersonal behavior: evidence, models and interventions. a systematic review of health professionals’ perceptions. (3):319–41. 61 Patient Education and Counseling 2006; :16. 1 2006; Implementation Science LeBlanc 2010 Hays 1993 LeBlanc A, Kenny DA, O’Connor AM, Legare F. Decisional Hays RD, Sherbourne CD, Mazel RM. The RAND 36-item conflict in patients and their physicians: a dyadic approach 2 Health Survey 1.0. Health Economics 1993; (3):217–27. to shared decision making. Medical Decision Making 2010; Hibbard 1997 (1):61–8. 29 HIbbard JH, Slovic P, Jewett JJ. Informing consumer Legare 2008b decisions in health care: Implications from decision-making Legare F, Ratte S, Gravel K, Graham ID. Barriers and research. Milbank Quarterly 1997; 75 (3):395–414. facilitators to implementing shared decision-making in Hibbard 2013 clinical practice: update of a systematic review of health Hibbard JH, Greene J. What the evidence shows about Patient Education and Counseling professionals’ perceptions. patient activation: better health outcomes and care 2008; 73 (3):526–35. experiences; fewer data on costs. Health Affairs 32 (2): 2013; Legare 2010 207–14. Legare F, Ratte S, Stacey D, Kryworuchko J, Gravel K, Higgins 2011 Graham ID, et al. Interventions for improving the adoption Higgins JPT, Green S (editors). Cochrane Handbook of shared decision making by healthcare professionals. for Systematic Reviews of Interventions Version 5.1.0 Cochrane Database of Systematic Reviews 2010, Issue 5. [updated March 2011]. The Cochrane Collaboration, [DOI: 10.1002/14651858.CD006732.pub2] 2011. Available from www.cochrane-handbook.org. Legare 2014 Legare F, Stacey D, Turcotte S, Cossi MJ, Kryworuchko J, Hoffman 2015 Graham ID, et al. Interventions for improving the adoption Hoffmann TC, Del Mar C. Patients’ expectations of the of shared decision making by healthcare professionals. benefits and harms of treatments, screening, and tests: Cochrane Database of Systematic Reviews 2014, Issue 9. Journal of the American Medical a systematic review. [DOI: 10.1002/14651858.CD006732.pub3] 175 (2):274–86. Association 2015; 58 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

62 and ‘Summary of findings’ tables. In: Higgins JPT, Makoul 2006 Green S (editors). Cochrane Handbook for Systematic Makoul G, Clayman ML. An integrative model of shared Reviews of Interventions Version 5.1.0 (updated March Patient Education decision making in medical encounters. 2011). The Cochrane Collaboration, 2011. Available from (3):301–12. and Counseling 2006; 60 www.handbook.cochrane.org. Michie 2002 Michie S, Dormandy E, Marteau TM. The multi- Sepucha 2013 dimensional measure of informed choice: a validation study. Sepucha KR, Borkhoff CM, Lally J, Levin CA, Matlock 48 2002; (1):87–91. Patient Education and Counseling DD, Ng CJ, et al. Establishing the effectiveness of patient decision aids: key constructs and measurement instruments. Mulley 1995 BMC: Medical Informatics and Decision Making 13 2013; Mulley A. Outcomes research: implications for policy and (Suppl 2):S12. practice. In: Smith R, Delamother T editor(s). Outcomes in . London: BMJ Publishing Group, 1995. Clinical Practice Spielberger 1970 Spielberger CD, Gorsuch RL, Lushene RE. Manual for the Munro 2016 State-Trait Anxiety Inventory (Self-evaluations questionn . aire) Munro S, Stacey D, Lewis KB, Bansback N. Choosing Palo Alto, CA: Consulting Psychologists Press, 1970. treatment and screening options congruent with values: do decision aids help? Sub-analysis of a systematic review. Stewart 1992 (4):491–500. 99 2016; Patient Education & Counseling Stewart AL, Ware JE Jr (editors). Measuring Functioning NCGC/NICE 2012 . and Well-being: The Medical Outcomes Study Approach National Clinical Guideline Centre. Patient experience in Durham NC: Duke University Press, 1992. adult NHS services: improving the experience of care for Trenaman 2014 people using adult NHS services. 2012. www.nice.org.uk/ Trenaman L, Stirling B, Bansback N. The cost-effectiveness nicemedia/live/13668/58283/58283.pdf. London, UK: of patient decision aids: a systematic review. Healthcare The Author, (accessed prior to 27 March 2017). 2 2014; (4):2510257. O’Connor 1995 Trenaman 2016 O’Connor AM. Validation of a decisional conflict scale. Trenaman L, Selva A, Desroches S, Singh K, Bissonnette J, (1):25–30. Medical Decision Making 1995; 15 Bansback N, et al. A measurement framework for adherence O’Connor 1998b in patient decision aid trials applied in a systematic review O’Connor AM, Tugwell P, Wells GA, Elmslie T, Jolly subanalysis. 77 : 2016; Journal of Clinical Epidemiology E, Hollingworth G, et al. A decision aid for women 15–23. considering hormone therapy after menopause: decision Trikalinos 2014 support framework and evaluation. Patient Education and Trikalinos TA, Wieland LS, Adam GP, Zgodic A, Ntzani EE. 1998; 33 (3):267–79. Counselling Decision Aids for Cancer Screening and Treatment . Rockville, O’Connor 2007 MD: Agency for Healthcare Research and Quality, 2014. O’Connor AM, Wennberg JE, Legare F, Llewellyn-Thomas Washington State 2016 HA, Moulton BW, Sepucha KR, et al. Toward the ’tipping Washington State Health Authority. Patient decision Health point’: decision aids and informed patient choice. aid certification criteria. 2016. www.hca.wa.gov/hw/ 26 2007; Affairs (3):716–25. Documents/sdm ̇cert ̇criteria.pdf (accessed prior to 27 RevMan 2014 [Computer program] March 2017). The Nordic Cochrane Centre, The Cochrane Collaboration. Weston 2001 Review Manager (RevMan). Version 5.3. Copenhagen: Weston WW. Informed and shared decision-making: the The Nordic Cochrane Centre, The Cochrane Collaboration, crux of patient-centered care. 165 2001; CMAJ (4):438–9. 2014. RNAO 2009 References to other published versions of this review Registered Nurses’ Association of Ontario. Decision support for adults living with chronic kidney disease. 2009. O’Connor 1999b rnao.ca/bpg/guidelines/decision-support-adults-living- O’Connor AM, Rostom A, Fiset V, Tetroe J, Entwistle chronic-kidney-disease. Toronto, Ontario: The Author, V, Llewellyn-Thomas H, et al. Decision aids for patients (accessed prior to 27 March 2017). facing health treatment or screening decisions: systematic BMJ review. (7212):731–4. 319 1999; Rothert 1987 Rothert M, Talarcyzk GJ. Patient compliance and the O’Connor 2001 Journal decision making process of clinicians and patients. O’Connor AM, Stacey D, Rovner D, Holmes-Rovner M, 2 1987; (1):55–71. of Compliance in Health Care Tetroe J, Llewellyn-Thomas H, et al. Decision aids for people facing health treatment or screening decisions. Schünemann 2011 2001, Issue 3. Cochrane Database of Systematic Reviews Schünemann HJ, Oxman AD, Higgins JPT, Vist GE, [DOI: 10.1002/14651858.CD001431] Glasziou P, Guyatt GH. Chapter 11: Presenting results 59 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

63 O’Connor 2003 Stacey 2011 O’Connor AM, Stacey D, Rovner D, Holmes-Rovner M, Stacey D, Bennett CL, Barry MJ, Col NF, Eden KB, Tetroe J, Llewellyn-Thomas H, et al. Decision aids for Holmes-Rovner M, et al. Decision aids for people facing people facing health treatment or screening decisions. health treatment or screening decisions. Cochrane Database Cochrane Database of Systematic Reviews 2003, Issue 1. of Systematic Reviews 2011, Issue 10. [DOI: 10.1002/ [DOI: 10.1002/14651858.CD001431] 14651858.CD001431.pub3] Stacey 2014b O’Connor 2009b Stacey D, Legare F, Col NF, Bennett CL, Barry MJ, O’Connor AM, Bennett C, Stacey D, Barry M, Col NF, Eden KB, et al. Decision aids for people facing health Eden KB, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database treatment or screening decisions. Cochrane Database 2014, Issue 1. [DOI: 10.1002/ of Systematic Reviews of Systematic Reviews 2009, Issue 3. [DOI: 10.1002/ 14651858.CD001431.pub4] ∗ 14651858.CD001431.pub2] Indicates the major publication for the study 60 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

64 C H A R A C T E R I S T I C S O F S T U D I E S Characteristics of included studies [ordered by study ID] Allen 2010 Methods Cluster-randomized to decision aid vs usual care Participants 398 + 414 men considering prostate cancer screening in the USA Interventions obabilities, explicit DA: computer tailored programme on clinical problem, outcome pr values clarification, others’ opinion and guidance (step-by-step process for making the decision; interactive computer programme: inherently guided the patient through the decision aid and decision making process), tailored printout gi ven to patients to promote discussion with others (practitioner, significant others) Comparator: no intervention Primary outcomes: decisional status, knowledge, decision sel Outcomes f-efficacy, decisional consis- tency Secondary outcomes: desire for involvement in decision making , decisional conflict, preferred options Outcomes assessed pre- and postintervention - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “Sites were blocked on size and percent of Low risk bias) male employees and randomly assigned by computer-generated random numbers to condition within blocks” (p 2173, Setting) Allocation concealment (selection bias) Unclear risk The study does not address this criterion. Unclear risk The study does not address this criterion. Blinding of participants and personnel (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes measured Low risk were not subjective to interpretation bias) All outcomes Incomplete outcome data (attrition bias) Low risk No missing outcome data and low rate All outcomes of attrition that was consistent between groups Selective reporting (reporting bias) Unclear risk No mention of protocol 61 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

65 Allen 2010 ( Continued) Intervention delivery: mention of money Other bias Low risk incentive to complete paperwork, but was judged to have no effect on outcomes mea- sured (p 2175) Arterburn 2011 Methods Randomized to decision aid vs usual care Participants 75 + 77 participants considering bariatric surgery in the USA Interventions DA: booklet + video on options’ outcomes, clinical problem, out come probabilities, others’ opinion, guidance (list of questions to discuss with cli nician) ical problem) Comparator: usual care (general information pamphlets on clin Primary outcomes: knowledge, values, values concordance Outcomes ecisional self-efficacy, Secondary outcomes: treatment preference, decisional conflict, d proportion undecided Primary outcomes assessed at baseline, postintervention an d 3 months follow-up; sec- ondary outcomes assessed at baseline and postintervention Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “[U]sed computer-assisted, block randomi- bias) sation process to ensure balanced allocation of participants” (p 1670, Participants and randomization) Unclear risk Allocation concealment (selection bias) No mention of allocation concealment and no mention of impact on study Blinding of participants and personnel Unclear risk “[S]tudy was not blinded” (p 1670, Partic- (performance bias) ipants and randomization); no mention of impact on study All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subject to interpre- bias) All outcomes tation Incomplete outcome data (attrition bias) Unclear risk Measures: mentioned 4 choices for treat- All outcomes ment preference (surgery, drug therapy, diet and/or exercise programme and unsure) but onlyreportedonsurgeryandunsure options (p 1671); minimal attrition that was consis- 62 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

66 Arterburn 2011 ( Continued) tent between groups Selective reporting (reporting bias) Unclear risk No mention of study protocol or trial regis- tration; all pre-specified outcomes included Low risk The study appears to be free of other sources Other bias of bias Auvinen 2004 Methods Randomized to decision aid vs usual care Participants 103 + 100 men newly diagnosed with prostate cancer in Finland DA: pamphlet patient decision aid created for study on options Interventions ’ outcomes, outcome probability, guidance Comparator: usual care by clinical guideline Outcomes Primary outcome: uptake of options Secondary outcome: participation in decision making Other outcomes (from Huang 2014): death (5 years), disease-free su rvival (10-years), biochemical failure (serum PSA elevation) (5 years), biochemical f ailure-free survival (5 years), disease progression (5 years), disease progression-fr ee survival (5 years) (data from 104 + 106 men) Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Auvinen 2001, p 2: “randomized centrally, Random sequence generation (selection Low risk bias) using software based on a random number generator”; no blocking used Auvinen 2004, (primary study), p 1: “ran- domized using a computer algorithm based on random numbers” Allocation concealment (selection bias) Unclear risk Auvinen 2001,p 2, Patients and Methods: randomized centrally at the Finnish Cancer Registry Auvinen 2004, (primary study), p 1: ran- domized centrally Comment: central allocation confers low risk Blinding of participants and personnel Auvinen 2001, p 3: “recognized carry-over High risk (performance bias) effect because same physician in charge for All outcomes intervention and control groups, diminish 63 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

67 Auvinen 2004 ( Continued) contrast between groups, as these physicians were more motivated to inform patients than those physicians not participating” Auvinen 2004 (primary study): no blind- ing but primary outcome is choice of treat- ment for prostate, objectively recorded. But unsure how physicians may have influenced decisions Blinding of outcome assessment (detection Low risk No blinding but primary outcome is choice of treatment for prostate, objectively bias) recorded All outcomes Incomplete outcome data (attrition bias) Low risk Auvinen 2001, p 3: flow-chart “Imbalance in the numbers of patients be- All outcomes tween the arms within two hospitals. Not expected to affect the results in any way”; “some participants refused to give informed consent, health deterioration, not seen by urologist” (p 4) Auvinen 2004 (primary study), p 2: flow di- agram and results; low attrition and consis- tent between groups Selective reporting (reporting bias) Unclear risk No indication that trial registered in central trials registry Auvinen 2001, p 2: “The study protocol was approved by an ethical committee in each participating hospital” Auvinen 2004 (primary study), p 1: “The study protocol was approved by the insti- tutional review board at each participating hospital” Low risk Appears to be free of other potential biases Other bias Barry 1997 Methods Randomized to decision aid vs usual care Participants 104 + 123 patients considering benign prostatic hyperplasia treatment in the USA Interventions DA: Health Dialog interactive videodisc on options’ outcomes , clinical problem, out- come probability, others’ opinion cal problem Comparator: usual care using general information on the clini Outcomes Primary outcome: knowledge Secondary outcomes: uptake of option, satisfaction with DM pro cess, satisfaction with decision, interest in DM, general health outcomes, condition s pecific health outcomes 64 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

68 Barry 1997 ( Continued) - Notes Risk of bias Risk of bias Support for judgement Bias Authors’ judgement Low risk “Stratified by study site in concealed blocks Random sequence generation (selection bias) of 10” (p 2) Allocation concealment (selection bias) Low risk Study coordinator opening serially num- bered, opaque, sealed envelopes (p 2) Blinding of participants and personnel Low risk No blinding but phase 1 eliminated risk of (performance bias) contamination All outcomes Blinding of outcome assessment (detection Low risk No blinding but phase 1 eliminated risk of bias) outcome assessor interfering with decision All outcomes Incomplete outcome data (attrition bias) Low risk Patient accrual and follow-up reported; All outcomes post-randomizationwithdrawals couldhave biased the results (more in intervention group) - however they reported no evidence of a differential effect of the study group (p 3) Unclear risk No indication that trial registered in central Selective reporting (reporting bias) trials registry Low risk Other bias Appears to be free of other potential biases Bekker 2004 Methods Randomized to detailed vs routine consultation Participants 59 + 58 pregnant women who have received a maternal serum screen ing positive test result for Down syndrome in the UK Interventions ptions’ outcomes,clinical DA (inconsult): decisionanalysisplusroutine consultationono problem, outcome probability, values clarification, guidance/ coaching Comparator: routine consultation on options’ outcomes, outco me probability Outcomes Primary outcome: anxiety Secondaryoutcomes:uptake of option,knowledge, decisional co nflict,informeddecision making, satisfaction with consultation, consultation length 65 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

69 Bekker 2004 ( Continued) - Notes Risk of bias Risk of bias Support for judgement Bias Authors’ judgement Unclear risk Bekker 2003, p 2 - section 2.3 Sample Random sequence generation (selection bias) and Procedure: “randomly allocated... us- ing previously numbered... envelopes” Bekker 2004 (primary study), p 3: “Partic- ipants were randomly allocated by previ- ously numbered envelopes”; does not men- tion how sequence was generated Allocation concealment (selection bias) Bekker 2003, p 2 - section 2.3 Sample and Low risk Procedure: “Using previously numbered, sealed, opaque envelopes” Bekker 2004 (primary study), p 3: previ- ously numbered, sealed, opaque envelopes Low risk blinded, Blinding of participants and personnel Participants personnel not blinded. Same personnel did (performance bias) control & intervention. Tape recorded ses- All outcomes sions to ensure no bias Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured All outcomes Incomplete outcome data (attrition bias) Unclear risk Bekker 2003 flow diagram indicates pos- trandomization attrition with more attri- All outcomes tion in decision aid group; no discussion on implications of attrition Bekker 2004 (primary study), p 4: results/ flow diagram; baseline characteristics not included Selective reporting (reporting bias) Unclear risk Bekker 2003: the coding frame was devel- oped from literature. Does not mention protocol Bekker 2004 (primary study): no informa- tion provided about central trials registry Other bias Unclear risk Bekker 2003: does not directly address baseline characteristics of participants Bekker 2004 (primary study): appears to be free of other potential biases 66 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

70 Bernstein 1998 Methods Randomized to decision aid vs usual care 65 + 53 patients with coronary artery disease considering reva scularization surgery in Participants the USA DA: Health Dialog video on options’ outcomes, clinical problem , outcome probability, Interventions others’ opinion Comparator: usual care (no information provided) g process Outcomes Primary outcome: satisfaction with decision and decision makin n with care, general health Secondary outcomes: uptake of option, knowledge, satisfactio outcomes, condition specific health outcomes Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection “Randomization was stratified by study site in blocks of 10” (p 3) bias) Allocation concealment (selection bias) Low risk “[R]andomization performed by a study co- ordinator opening opaque, sealed envelopes at study headquarters” (p 3) Unclear risk Neithersubjectsnorstudystaffwere blinded Blinding of participants and personnel to treatment assignment - could lead to dif- (performance bias) ferent satisfaction ratings based on knowing All outcomes the treatment received Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured All outcomes Incomplete outcome data (attrition bias) Low risk Flow diagram (p 3); low attrition of eligible All outcomes participants randomized and consistent be- tween group Unclear risk No information provided indicating trial Selective reporting (reporting bias) was included in central trials registry Other bias Low risk Appears to be free of other potential biases 67 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

71 Berry 2013 Methods Randomized to decision aid vs usual care 266 + 228 men considering prostate cancer treatment in the USA Participants roblem, outcome prob- Interventions DA: interactive web based video on options’ outcomes, clinical p k doctor and automated sum- abilities, others’ opinion, guidance (list of questions to as mary) Comparator: usual care Primary outcome: decisional conflict Outcomes Secondary outcome: preferred/actual treatment choice (pre- and p ost-DA), proportion undecided from Other outcomes (Bosco 2012): choice concordance (6 months post-DA). (Data 239 + 209 men) Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection Methods section- second paragraph, p 3: “Participants were randomized automati- bias) cally by the P3P application to study groups (1:1 using a simple randomization scheme with no blocking)” Allocation concealment (selection bias) Low risk Methods section, p 3: “Participants were randomized automatically by the P3P ap- plication to study groups” Participants were not blinded and study Unclear risk Blinding of participants and personnel does not address the effect on the results (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear whether outcome assessors are blinded, but outcomes are not subject to in- bias) All outcomes terpretation Incomplete outcome data (attrition bias) Low risk Used intention-to-treat analysis and low All outcomes dropout (p 4) Selective reporting (reporting bias) Protocol made available Low risk Other bias Unclear risk Was a multicentre trial which could have lead to contamination, protocol violation and biased questionnaire completion 68 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

72 Bjorklund 2012 Methods Randomized to decision aid vs usual care 236 + 247 women less than 11 weeks pregnant considering Down sy ndrome screening Participants in Sweden DA: linear video on options’ outcomes, clinical problem, outcom e probabilities, others’ Interventions opinion, and guidance (step-by-step process for making the decisi on) Comparator: usual care using pamphlet e of combined ul- Primary outcomes: knowledge (post-DA), attitude (post-DA), uptak Outcomes trasound and biochemical screening (post-DA) ) Secondary outcomes: values congruent with chosen option (post-DA - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Unclear risk Random sequence generation (selection “The midwife allocatedthe participantsran- domly by sealed envelopes” (p 391) but bias) does not state the actual sequence genera- tion method Allocation concealment (selection bias) Low risk Used sealed envelopes, “prepared, sequen- tially coded and distributed to the maternity units by the research group” (p 391) Unclear risk “It was not possible to blind neither [sic] Blinding of participants and personnel (performance bias) the midwives nor the participants due to the All outcomes characteristics of the intervention” (p 395) . The study does not address the effects of this on the results Blinding of outcome assessment (detection Low risk No blinding but outcomes were objectively bias) measured and not subjective to interpreta- tion All outcomes Incomplete outcome data (attrition bias) Unclear risk No mention of why some participants’ data were excluded in Tables 2, 3 and 4 All outcomes Selective reporting (reporting bias) Unclear risk No mention of study protocol Other bias Low risk Appears to be free of other sources of bias 69 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

73 Bozic 2013 Methods Randomized to decision aid vs usual care 95 + 103 participants with hip and/or knee osteoarthritis cons idering hip/knee surgery Participants come probabilities, Interventions DA: DVD and booklet on options’ outcomes, clinical problem, out coaching with health coach explicit values clarification, others’ opinions, and guidance/ Comparator: usual care using pamphlet Outcomes Primary outcomes: informed decision/knowledge (pre, immedia tely post, and 6 weeks follow-up) Secondary outcomes: preferred treatment choice (pre and immedia tely post), patient and provider satisfaction (immediately post), length of consultat ion time Notes Trial registration: NCT01492257 Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “The randomization was blocked with use Random sequence generation (selection Low risk bias) of random permuted blocks in groups of four, six, or eight to help ensure that the groups were balanced” (p 1634) Allocation concealment (selection bias) Low risk “Patients were randomized to either the in- tervention group or the control group with use of the sealed envelop method” (p 1634) Unclear risk “[S]urgeons were not blinded to the inter- Blinding of participants and personnel (performance bias) vention” (p 1635). Knowing the allocation All outcomes of participants, surgeons’ favourable scoring could be due to greater investment in de- cision-making. Insufficient information to make a judgment Blinding of outcome assessment (detection Low risk Outcomes are objectively measured and not bias) subject to interpretation All outcomes Unclear risk 62% (123/198) retention rate therefore Incomplete outcome data (attrition bias) All outcomes high attrition rate - however the attrition was balanced between groups Selective reporting (reporting bias) Low risk Protocol available Other bias Low risk Appears to be free of other sources of bias 70 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

74 Brazell 2014 Methods Randomized to DA + standard counselling vs usual care + standar d counselling elvic organ prolapse 53 + 51 women presenting for the management and treatment of p Participants DA: paper-based or web-based DA on clinical problem, options’ ou tcomes, outcome Interventions probabilities, patient stories and standard counselling Comparator: standard counselling alone tion) Outcomes Primary outcomes: decisional conflict (immediately postconsulta Secondary outcomes: choice (3 months after making decision), decisi onal regret (3 months after making decision) - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “Patients were randomized 1:1 using a ran- dom numbers table in blocks of 6” (p 231) bias) Unclear risk Insufficient information provided to make Allocation concealment (selection bias) judgment Blinding of participants and personnel Unclear risk Insufficient information provided to make (performance bias) judgment All outcomes Blinding of outcome assessment (detection Insufficient information provided to make Unclear risk judgment bias) All outcomes Unclear risk Incomplete outcome data (attrition bias) High attrition but balanced between All outcomes groups: “39 randomized subjects were ei- ther missed by the research assistant at their new patient visit and thus did not receive a DCS questionnaire to complete or they canceled their appointments and did not reschedule a new one” (p 233). There was a 48% (50/104) attrition rate for Decisional Regret measures Selective reporting (reporting bias) Low risk Trial registered Other bias High risk Risk of contamination due to same physi- cians in both groups. Also, outcomes mea- sured after the PtDA and physician consult 71 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

75 Chabrera 2015 Methods Randomized to DA vs usual care 73 + 74 men recently diagnosed with prostate cancer considering t reatment options Participants ons’ outcomes, outcome Interventions DA: 2-part decision support booklet with clinical problem, opti probabilities, patient stories, explicit values clarificatio n, and guidance Comparator: usual care Outcomes Primary outcomes: knowledge, decisional conflict, satisfaction with decision-making process Secondary outcome: coping Outcomes assessed at 3 months postintervention Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “[S]tudy participants were randomized into 1 of 2 arms using a computer-generated random list with bias) unequal blocks” (p E44) Allocation concealment (selection bias) Unclear risk Insufficient information provided to make judg- ment Blinding of participants and personnel Unclear risk Insufficient information provided to make judg- ment (performance bias) All outcomes Blinding of outcome assessment (detection Unclear risk Insufficient information provided to make judg- ment bias) All outcomes Incomplete outcome data (attrition bias) Low risk Balanced attrition in both groups All outcomes Selective reporting (reporting bias) No protocol provided; trial not registered Unclear risk Other bias Unclear risk Prostate cancer in Catalonia is common; however, only 147 were recruited for this trial (p E44) 72 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

76 Chambers 2012 Methods Randomized to DA vs usual care 74 + 77 healthcare workers who did not receive the influenza vaccin e considering receiv- Participants ing the vaccine in Canada DA: web-based DA on options’ outcomes, clinical problem, outcome probabilities, ex- Interventions plicit values clarification and guidance Comparator: usual care using pamphlet Outcomes Primary outcomes: confidence in decision (post-DA) Secondary outcomes: impact on immunization intent (post-DA), pro portion undecided - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “The randomization list was generated using the Low risk Random sequence generation (selection bias) randomization function in Excel 2002 (version 10. 6856.6856 SP3)” (p 199) Allocation concealment (selection bias) Low risk “The list was imported from Excel into a Microsoft SQL Server database. The online application would sequentially assign a random identification number and their decision aid status (seeing the decision aid or not) from the randomization list when users logged into the survey.” (p 199) Unclear risk Not reported whether or not they were blinded dur- Blinding of participants and personnel (performance bias) ing the course of the intervention All outcomes Blinding of outcome assessment (detection Low risk Questionnaire scores are objective and not subject bias) to interpretation All outcomes Incomplete outcome data (attrition bias) High risk 65% completion rate in intervention arm and 77% All outcomes completion rate in control arm: attrition could be different where the respondents and non-respon- dents are different Selective reporting (reporting bias) Low risk Protocol available Other bias Unclear risk Figure 1 numbers for exclusion are not logical 73 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

77 Clancy 1988 Methods Randomized to decision aid vs usual care Participants 753 + 263 health physicians considering Hep B vaccine in the USA Interventions obability, explicit DA: pamphlet on options’ outcomes, clinical problem, outcome pr values clarification (personal decision analysis), guidance/coach ing Comparator: usual care (no information provided) Outcomes Uptake of option Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random numbers table; all incoming resi- Low risk Random sequence generation (selection bias) dents were assigned to Group 2 (non-ran- domized residents identified as subgroup) (p 2) Unclear risk No information provided Allocation concealment (selection bias) Blinding of participants and personnel Unclear risk No blinding of participants or personnel. (performance bias) Did not report on how this may affect their All outcomes findings Blinding of outcome assessment (detection Low risk Unclear blinding but decisions for screen- bias) ing were retrieved from health records (ob- jective data) All outcomes Unclear risk Flow chart not included. Insufficient infor- Incomplete outcome data (attrition bias) All outcomes mation to make a judgment Selective reporting (reporting bias) Unclear risk No information provided Other bias Potential selection bias - non-randomized High risk residents were added to group 2 and there- fore potential unbalanced distribution (p 287) Low response rate among those offered de- cision analysis 74 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

78 Davison 1997 Methods care Randomized to decision aid + audio-taped consultation vs usual 30 + 30 men with prostate cancer considering treatment in Canada Participants nical problem, outcome Interventions DA: written + audiotape consultation of options’ outcomes, cli probability, others’ opinion ical problem) Comparator: usual care (general information pamphlets on clin Primary outcomes: role in decision making Outcomes Secondary outcomes: anxiety, depression Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement “The group to which subjects were assigned Random sequence generation (selection Low risk bias) was predetermined by a block randomiza- tion procedure. This ensured there were an equal number of subjects in both groups for each physician.” (p 5, Data collection) Allocation concealment (selection bias) Unclear risk Not mentioned; group assignment prede- termined by block randomization proce- dure (p 5) Blinding of participants and personnel Unclear risk No blinding; study does not report on how (performance bias) the results could be influenced by lack of blinding All outcomes Unclear risk Unclear blinding and whether outcomes Blinding of outcome assessment (detection bias) could be affected by unblinded assessor All outcomes Incomplete outcome data (attrition bias) No flow diagram; p 12 explains why certain Low risk All outcomes men did not listen to audiotape. All men approached by study investigator agreed to participate; only 1 man refused to complete the second set of questionnaires Selective reporting (reporting bias) Unclear risk Protocol not mentioned Other bias Low risk Appears to be free of other sources of bias; similar baseline characteristics 75 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

79 De Achaval 2012 Methods Randomized to detailed vs simple vs usual care 70 + 70 + 71 patients diagnosed with knee osteoarthritis consi dering treatment in the Participants USA Complex DA: video booklet + interactive joint analysis on opt ions’ outcomes, clinical Interventions problem, outcome probabilities, explicit values clarification , others’ opinion and guid- ance (list of questions) Comparator DA: video booklet on options’ outcomes, clinical pr oblem, outcome prob- abilities, others’ opinion and guidance (list of questions) Comparator: usual care receiving generic booklet Outcomes Decisional conflict (baseline and postintervention) - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Computer generated list with uneven Random sequence generation (selection bias) blocks (p 231) Allocation concealment (selection bias) Low risk Numbered, sealed and opaque envelopes (p 231) Blinding of participants and personnel Likely not blinded, but low threat of bias Low risk in study (p 231) (performance bias) All outcomes Low risk Participants were not blinded but outcome Blinding of outcome assessment (detection bias) was objectively measured (p 231) All outcomes Incomplete outcome data (attrition bias) 3 dropouts; missing data effectsize unlikely Low risk All outcomes to have significant impact on study out- come Selective reporting (reporting bias) Unclear risk Protocol not available Other bias Low risk Appears to be free of other sources of bias 76 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

80 Dolan 2002 Methods Randomized to decision aid vs usual care 50 + 47 average risk for colorectal cancer considering screening in the USA Participants es, clinical problem, Interventions DA: computer with analytic hierarchy process on options’ outcom outcome probability, explicit values clarification, guidance/co aching roblem Comparator: usual care with information on options, clinical p Outcomes Primary outcomes: uptake of option, decisional conflict Secondary outcomes: role in decision making - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “[R]andomization schedules were created Low risk Random sequence generation (selection bias) using a computer random number genera- tor” (p 2, Study interventions) Allocation concealment (selection bias) Low risk Computer-based (p 2, Study interventions) Blinding of participants and personnel Unclear blinding of participants. All pa- Unclear risk (performance bias) tient interviews in both the experimental and control groups were done by the same All outcomes investigator, unclear on how this could con- tribute to risk of bias Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Low risk See flow diagram - low attrition All outcomes Selective reporting (reporting bias) Unclear risk Nothing specifically mentioned re study protocol Low risk Appears to be free of other sources of bias Other bias Evans 2010 Methods Randomized to online decision aid vs paper decision aid vs ques tionnaire vs usual care Participants 129 + 126 + 127 + 132 men considering PSA screening in Wales 77 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

81 Evans 2010 ( Continued) DA: online programme on options’ outcomes, clinical problem, o utcome probabilities, Interventions explicit values clarification, others’ opinion, guidance (inter active computer programme; summary) clinical problem, out- Comparator: paper version of online DA on options’ outcomes, come probabilities, explicit values clarification, others’ opi nion, guidance (interactive computer programme; summary) Comparator: received a questionnaire Comparator: received nothing Outcomes Primary outcomes: knowledge (post-DA) A testing (post-DA), Secondary outcomes: attitude (post-DA), intention to undergo PS nflict anxiety (post-DA), uptake of PSA test (post-DA), total decisional co - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Unclear risk “[A] random sample of 100 men was se- lected from the list.” “The process ensured bias) individual level randomization” (p 4, Re- cruitment process) Allocation concealment (selection bias) Low risk “[A]ffirmative consent forms from each practice were transferred to the research of- ficer who allocated each participant with a number provided remotely by the trial statistician to ensure concealment” (p 4, Recruitment process) The study does not address this outcome Blinding of participants and personnel Unclear risk (performance bias) All outcomes Low risk Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection bias) tively measured and not subjective to inter- All outcomes pretation Incomplete outcome data (attrition bias) See flow diagram indicating high attrition Low risk All outcomes consistently across groups Selective reporting (reporting bias) Low risk Registered as a trial Other bias Low risk The study appears free of other sources of bias 78 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

82 Fagerlin 2011 Methods Decision aid vs delayed intervention vs control 382 + 159 + 100 women with an elevated 5-year risk of breast cancer considering breast Participants cancer prevention medication in the USA DA: tailored DA on options’ outcomes, clinical problem, outcome probabilities, and Interventions explicit values clarification Comparator 1: given DA after 3-month follow-up Comparator 2: given DA after all outcome measures were taken Outcomes Decisional conflict (post-DA), behavioural intent (post-DA), actual b ehaviour (post- ption of risk (post- DA), proportion undecided, perception of benefits (post-DA), perce DA) Other outcomes: Banegas 2013: decisional conflict (post-DA) (data from 690 + 160 + 16 • 2 women) , proportion undecided (3 months) Korfage 2013: knowledge (immediately post and 3 months post-D • A), attitudes (immediately post and 3 months post-DA), behavioural intent (po st-DA), actual behaviour (3 months post-DA), informed decision defined as “part icipants with ttitudes were sufficient knowledge about chemoprevention behavior, whose a prevention behavior” concordant with their intentions or decisions to engage in chemo (data from 383 + 102 + 100 women). Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Random sequence generation was provided Low risk by the author bias) Low risk Central and web-based allocation Allocation concealment (selection bias) Unclear risk Unclear blinding - using an online deci- Blinding of participants and personnel (performance bias) sion aid would have avoided control par- All outcomes ticipants accessing the decision aid Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Unclear risk Does not report exclusions; inadequate re- All outcomes porting on participant flow through the study to determine risk for attrition bias or incomplete outcome data Selective reporting (reporting bias) Unclear risk No mention of study protocol 79 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

83 Fagerlin 2011 ( Continued) Low risk Appears to be free of other sources of bias Other bias Fraenkel 2007 Randomized to decision aid vs usual care Methods Participants 47 + 40 patients with knee pain considering treatment options in the USA Interventions DA: interactive computer tool options’ outcomes, outcome proba bility, explicit values clarification mation pamphlet Comparator: usual care using the Arthritis Foundation infor Decisional self-efficacy, preparation for decision making Outcomes Primary outcome was not specified Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Computer-generated randomization se- Random sequence generation (selection bias) quence (p 2) Allocation concealment (selection bias) Unclear risk No information provided; computer gener- ated Blinding of participants and personnel No blinding but study does not report if it Unclear risk had an impact on the outcomes measured (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Low risk of attrition bias - outcome data for Low risk All outcomes all 40 controls and 44 of 47 intervention (p 3, Results) Selective reporting (reporting bias) Unclear risk No information provided; no indication of trial was registered centrally Other bias Low risk Appears to be free of other potential biases 80 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

84 Fraenkel 2012 Methods us usual care Cluster-randomized control trial of clinics to decision aid vers 69 + 66 patients with nonvalvular atrial fibrillation conside ring anticoagulation with Participants aspirin or warfarin DA: computer-based tool on options’ outcomes, clinical problem, options’ probabilities, Interventions guidance, explicit values clarification Comparator: control arm (no further information provided) Outcomes Primary outcomes: feeling informed and having clear values (ba seline, immediately post) ccuracy of risk (baseline, Secondary outcomes: knowledge (baseline, immediately post), a immediately post), anxiety (baseline, immediately post), wor ry (baseline, immediately r - DA group only), discus- post), rationale for preferred treatment (during the encounte iotape), change in sion of related outcomes (during the encounter as captured on aud ectations (stroke, bleeding) treatment plan (post intervention), anxiety, accurate risk exp Trial registration NCT00829478 Notes Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Unclear risk Inadequate information on random se- bias) quence generation Allocation concealment (selection bias) Unclear risk inadequate information on allocation con- cealment Blinding of participants and personnel “Toavoidcontamination,participantswere Low risk (performance bias) randomized at the level of the firm so that All outcomes all participants in one firm received the in- tervention, and all participants in the sec- ond firm were included in the control arm” (p 1435) Blinding of outcome assessment (detection Low risk “An interviewer blinded to the participant’s group assignment reassessed the primary bias) All outcomes and secondary outcomes after participant’s primary care visit” (p 1436) Incomplete outcome data (attrition bias) Unclear risk Does not appear to be incomplete outcome data; flow diagram does not report partici- All outcomes pation beyond randomization Selective reporting (reporting bias) Low risk Protocol available Other bias Low risk Does not appear to be any other potential sources of bias 81 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

85 Frosch 2008a Methods Randomized to decision aid vs. decision aid + chronic disease tr ajectory vs chronic disease trajectory vs usual care (Internet information) 155 + 152 + 153 + 151 men considering prostate cancer screening Participants DA: information on options’ outcomes, clinical problem, outcom e probabilities, others’ Interventions opinions Comparator 1: information on options’ outcomes, clinical prob lem, outcome probabil- ies for outcomes associated with ities, others’ opinions, explicit values clarification (utilit prostate cancer) Comparator 2: explicit values clarification (utilities for outco mes associated with prostate cancer) e cancer screening on Comparator 3: usual care using public information on prostat vention websites 2005- American Cancer Society and Centers for Disease Control and Pre 2006 Outcomes Primary outcomes: knowledge, actual option, decisional conflict Secondary outcomes: concern about prostate cancer, treatment pre ference if prostate cancer diagnosed Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk Computer algorithm randomly assigned bias) participants to the 4 study groups Allocation concealment (selection bias) Low risk Revealed after signed consent and com- pleted baseline measures Accessed a secure Internet site that hosted Blinding of participants and personnel Unclear risk all study materials; participants had unlim- (performance bias) ited access to assigned intervention, unclear All outcomes blinding of personnel Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were mea- bias) sured via questionnaires and not subjective to interpretation All outcomes Incomplete outcome data (attrition bias) Low risk Used intention-to-treat analysis; imputed All outcomes missing data for participants who did not complete follow-up assessments; minimal attrition Selective reporting (reporting bias) Unclear risk No indication of published protocol 82 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

86 Frosch 2008a ( Continued) Low risk Appears to be free of other potential biases Other bias Gattellari 2003 Randomized to decision aid vs usual care Methods 126 + 122 men considering PSA testing in Australia Participants obability, explicit Interventions DA: pamphlet on options’ outcomes, clinical problem, outcome pr values clarification est and chances of false- Comparator: usual care using brief information on screening t positive results Outcomes perceptions, perceived Preferred option, knowledge, decisional conflict, accurate risk ability to make an informed choice Primary outcome was not specified Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Unclear risk Pre-randomized code - no further informa- Random sequence generation (selection bias) tion (p 1) Allocation concealment (selection bias) Low risk Pre-randomized code unobtrusively marked on envelopes (p 1) Blinding of participants and personnel Consenting men were blinded to allocation, Unclear risk but unclear if personnel were blinded (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Pre-test characteristics included. Flow chart Unclear risk All outcomes not included and reasons for attrition not mentioned; some attrition but balanced be- tween groups Selective reporting (reporting bias) Unclear risk No information provided Other bias Low risk Appears to be free of other potential biases 83 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

87 Gattellari 2005 Methods al care Randomized to decision aid booklet vs decision aid video vs usu 140 + 141 + 140 men considering PSA testing in Australia Participants obability, explicit Interventions DA: pamphlet on options’ outcomes, clinical problem, outcome pr values clarification others’ opinion Comparator 1: video on clinical problem, outcome probability, Comparator 2: usual care using brief information on screening test and chances of false- positive results Outcomes Preferred option, knowledge, decisional conflict, perceived ab ility to make an informed choice Notes Primary outcome was not specified Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection Unique identification codes assigned to bias) participants according to date and time en- rolled into the interventional component of the study. Block randomization of iden- tification codes then performed via com- puter software (p 2 - 2.3.1) Allocation concealment (selection bias) Low risk “Allocation concealment was ensured as the interviewers, responsible for enrolling par- ticipants onto the trial, were blinded to the randomized study design while one of the authors(MG)wasresponsible forrandomi- sation. Hence, it was not possible for either participants or interviewers to be aware of the randomisation sequence.” (p 2 - 2.3.1) Blinding of participants and personnel Low risk Participants and interviewers were blinded (performance bias) All outcomes Blinding of outcome assessment (detection Low risk At post-test, it was not possible to blind the bias) interviewers but outcomes were objectively measured and not subjective to interpreta- All outcomes tion Incomplete outcome data (attrition bias) Low risk Minimal attrition that is consistent across All outcomes groups (figure 1) Selective reporting (reporting bias) Unclear risk “[S]uccess of study protocol” limitation to protocol: men not confronted with actual 84 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

88 Gattellari 2005 ( Continued) decision to undergo PSA screening; no in- dication that trial registered in central trials registry (p 13, paragraph 5) “[H]igh follow-up rate and allocation con- Other bias Low risk cealment; study not subjected to selection bias” (p 13, paragraph 5). Appears to be free of other sources of bias Green 2001 Methods Randomized to decision aid + counselling vs counselling alone v s usual care 29 + 14 women with a first degree relative with breast cancer inte rested in learning about Participants genetic testing in the USA Interventions em, others’ opin- DA: CD-ROM plus counselling on options’ outcomes, clinical probl ions, guidance/coaching Comparator: counselling Comparator: usual care Outcomes Primary outcome: preferred options Secondary outcome: knowledge Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “[B]lock randomization schedule to one of Low risk three groups in a 2:2:1 ratio” (p 2) bias) Allocation concealment (selection bias) Unclear risk No information provided Unclear risk “[G]enetic counsellor blinded to random- Blinding of participants and personnel ization until just prior to the session” (p 2) (performance bias) , unclear if participants were blinded All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to to in- All outcomes terpretation Incomplete outcome data (attrition bias) Unclear risk “Values do not always add up to the num- ber of participants due to missing data”; All outcomes reasons not mentioned (p 4). “Participants’ baseline knowledge was reflected in the control group’s answers”; participants bal- 85 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

89 Green 2001 ( Continued) anced in study groups Selective reporting (reporting bias) Unclear risk No information provided Appears to be free of other sources of bias Low risk Other bias Hamann 2006 Methods Cluster-randomized trial of decision aid vs usual care Participants 54 + 59 patients with schizophraenia considering treatment op tions (cluster-RCT with 12 wards paired and randomized) in Germany DA: 16-page booklet on options’ outcomes, outcome probabiliti Interventions es, explicit values clari- fication, coaching/guidance Comparator: usual care Outcomes ave me a chance Knowledge, participation in decision making (COMRADE - doctor g to decided which treatment I thought was best for me), uptake of p sycho-education, rehospitalization, adherence, satisfaction with care, sever ity of illness (baseline only), attitudes about drug use, decision making preference Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “[O]ne member of each pair being ran- Unclear risk domly assigned to the control or to the in- bias) terventional condition” (p 266). Sequence generation method was not stated Allocation concealment (selection bias) Unclear risk No mention of allocation concealment Blinding of participants and personnel No information provided Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection No information provided Unclear risk bias) All outcomes Low risk Reasons for attrition mentioned Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Unclear risk No information provided 86 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

90 Hamann 2006 ( Continued) Clustering was not accounted for in the Other bias High risk analysis Hanson 2011 Methods Randomized to decision aid vs usual care g problems considering Participants 127 + 129 patients diagnosed with advanced dementia and eatin long-term feeding tube placement in the USA DA: booklet or audio recording on options’ outcomes, clinical pr oblem, outcome proba- Interventions bilities, explicit values clarification, others’ opinion, gui dance (steps in decision making, worksheet, summary) Comparator: usual care Primary outcomes: decisional conflict (3 months post-DA) Outcomes requency of communica- Secondary outcomes: surrogate knowledge, risk perceptions, f tion with providers (3 months post-DA), feeding treatment use (3 , 6 and 9 months post- DA), participation in decision making, satisfaction with the de cision, decisional regret Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk Computerized random number generation bias) (p 2010, Randomization) Allocation concealment (selection bias) Unclear risk No description of method used to conceal allocation (p 2010, Randomization) Blinding of participants and personnel “Cluster randomization prevented double Unclear risk (performance bias) blinding and may have introduced bias due All outcomes to site effects” (p 2014, Discussion); study authors unsure of effect on study Blinding of outcome assessment (detection “[B]ecause of cluster randomization, data Low risk collectors were not blinded to group assign- bias) ment” (p 2010, Randomization); authors All outcomes believe has little impact on study Incomplete outcome data (attrition bias) Unclear risk Intervention group missing data for 1 par- All outcomes ticipant, reason for omission not reported (table 1) No explanation for number of participants in each group (n = 127) given numbers vary 87 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

91 Hanson 2011 ( Continued) from those in ’recruitment and retention’ figure (table 4) Selective reporting (reporting bias) Registered with clinicaltrials.gov, protocol Low risk on website Appears to be free of other potential biases Other bias Low risk Heller 2008 Randomized to decision aid vs usual care Methods Participants 66 + 67 breast cancer patients eligible for breast reconstructio n in the USA DA: interactive software programme on options’ outcomes, oth ers’ opinions Interventions Comparator: standard patient education Outcomes sfaction with decision-mak- Knowledge, anxiety, satisfaction with treatment choice, sati ing ability Notes Primary outcome was not specified Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “upon study entry, the participants were bias) randomized (computer generated) to one of two groups” (p 2) Not enough information provided Unclear risk Allocation concealment (selection bias) Blinding of participants and personnel Unclear risk No information provided (performance bias) All outcomes Blinding of outcome assessment (detection No information provided Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) Low risk Baseline anxiety and knowledge included in All outcomes graphs. Participant numbers between study groups balanced (p 3). Reasons for incom- plete questionnaires and study withdrawals mentioned Selective reporting (reporting bias) Unclear risk No information provided re protocol 88 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

92 Heller 2008 ( Continued) Low risk Appears to be free of other potential biases Other bias Hess 2012 Randomized to decision aid vs usual care Methods Participants 103 + 105 patients in the the emergency department with primar y symptoms of nontrau- matic chest pain and were being considered of admission to the e mergency department observation unit for monitoring and cardiac stress testing w ithin 24 hours DA (in consultation): 1-page printout on options’ outcomes, clini cal problem, and out- Interventions come probabilities Comparator: usual care Primary outcomes: knowledge Outcomes choice, satisfaction with Secondary outcomes: risk perceptions, decisional conflict, actual decision making process, patient-practitioner communication - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “Patients were randomized to either usual Low risk bias) care or shared decision making through a Web-based, computer-generated allocation sequence in a 1:1 concealed fashion” (p 253) Low risk “Patients were randomized to either usual Allocation concealment (selection bias) care or shared decision making through a Web-based, computer-generated allocation sequence in a 1:1 concealed fashion” (p 253) Blinding of participants and personnel Personnel were blinded, but unclear if pa- Low risk tients were blinded (p 253, Outcome mea- (performance bias) All outcomes sures). However, the primary outcome is unlikely to be biased Blinding of outcome assessment (detection Investigators assessing outcomes were Low risk blinded (p 253, Outcome measures). bias) All outcomes Incomplete outcome data (attrition bias) Unclear risk Some of the numbers of patients reported All outcomes in the results did not match the flow chart Selective reporting (reporting bias) Low risk Protocol is available 89 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

93 Hess 2012 ( Continued) Appears to be free of other biases Other bias Low risk Jibaja-Weiss 2011 Methods Randomized to decision aid vs usual care Participants 51 + 49 women diagnosed with breast cancer considering surgical t reatment in the USA outcome probabili- Interventions DA: computer programme on options’ outcomes, clinical problem, nce (step-by-step process for ties, explicit values clarification, others’ opinion and guida making the decision) Comparator: usual care + breast cancer treatment educational ma terials normally pro- vided to patients Surgical treatment preference (post-DA), breast cancer knowledge (pre, post-DA, post- Outcomes DA andconsult), satisfactionwith surgical decision(post-DA),sat isfactionwith decision- making process (post-DA), decisional conflict (pre, post-DA, post-DA a nd consult), proportion undecided Notes Primary outcome was not specified Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “Patients at each hospital were randomized bias) using permuted blocks” (p 42, Methods sec- tion) Allocation concealment (selection bias) Unclear risk Not addressed in the study Blinding of participants and personnel Unclear risk Not addressed in the study (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Unclear risk There is no way to know if the plots include All outcomes all of the participants’ data since they do not specify what was the number of patients used to obtain these mean scores Selective reporting (reporting bias) Unclear risk No mention of protocol Other bias Low risk Appears to be free of other potential biases 90 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

94 Johnson 2006 Methods Randomized to decision aid vs usual care the USA Participants 32 + 35 patients considering endodontic treatment options in DA (in consultation): decision board on options’ outcomes, clinica l problem, outcome Interventions probability, guidance Comparator: usual care ng process, anxiety Outcomes Primary outcomes: knowledge, satisfaction with decision maki - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “[F]our computerized random generation Low risk Random sequence generation (selection bias) lists to assign to one of two groups” (p 3) Allocation concealment (selection bias) Unclear risk Not for residents: computer-generated ran- domization lists (1 for each resident) were prepared by the PI (p 3-4); therefore resi- dents would have had pre-generated lists; Unclear for patients: “allocation was con- cealed from patients” (p 3) but does not ex- plain how Blinding of participants and personnel Unclear risk Blinding not mentioned. Allocation was (performance bias) concealed from patients only (p 3) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to to in- All outcomes terpretation Flow diagram (p 6); all 40 patients agreed to Low risk Incomplete outcome data (attrition bias) participate in the study, but only 32 ques- All outcomes tionnaires were useable several residents did not understand need for entering data on the envelope and placing matched question- naire in it (p 5) Selective reporting (reporting bias) No indication that the trial was registered Unclear risk in a central trials registry Other bias Unclear risk “[B]aseline data obtained because possi- ble that clinicians training in the EndoDB would alter usual care discussions” (p 5) . Mentions taking baseline characteristics, 91 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

95 Johnson 2006 ( Continued) but not included in article Kasper 2008 Methods Randomized to decision aid vs usual care apy in Germany 150 + 147 multiple sclerosis patients considering immunother Participants em, outcome probabil- DA: booklet and worksheet on options’ outcomes, clinical probl Interventions ities, explicit values clarification (based on IPDAS) s) Comparator: information material on immunotherapy (80 page Outcomes Primary outcomes: role in decision making Secondary outcomes: choice, feeling undecided, helpfulness wit h making a decision, attitudes toward immunotherapy, expectations of side effect s realized at 6 months - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk “[A]llocation using computer generated Random sequence generation (selection bias) random numbers” (p 5) Allocation concealment (selection bias) Unclear risk Randomization was carried out by con- cealed allocation, but method of conceal- ment was not described (p 2, Assignment) Low risk Participants were not told whether the in- Blinding of participants and personnel (performance bias) formation they received was standard infor- mation or the newly developed DA (p 3, All outcomes Masking) Low risk Assessors were not told whether the infor- Blinding of outcome assessment (detection bias) mation they received was standard informa- All outcomes tion or the newly developed DA (p 3, Mask- ing) Incomplete outcome data (attrition bias) Low risk Flow of participants (p 2, Fig 1); baseline data/characteristics included All outcomes Selective reporting (reporting bias) Low risk “The protocol of this study has been pub- lished with the trial registration at http:// controlled-trials.com/ ISRCTN25267500” (p 2) 92 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

96 Kasper 2008 ( Continued) Unclear risk Difference in preferred interaction style be- Other bias tween groups at baseline (P value 0.04) (p 5) Kennedy 2002 Randomized to decision aid + coaching vs decision aid only vs usua l care Methods Participants 215 + 206 + 204 women considering treatment for menorrhagia in the UK DA: video + booklet on options’ outcomes, clinical problem, out come probabilities, Interventions ing explicit values clarification, others’ opinions, guidance/coach y a registered nurse Coaching: ~ 20 minute coaching with explicit values clarification b prior to seeing physician Comparator: usual care Primary outcomes: general quality of life Outcomes Secondary outcomes: uptake of option, satisfaction, menorrha gia severity, cost-effective- ness Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Allocation sequence was generated by com- Low risk bias) puter and stratified by consultant and the age at which the woman left full-time edu- cation (p 3) Low risk “Secure randomization ensured by using Allocation concealment (selection bias) a central telephone randomization system” (p 3) Blinding of participants and personnel Possibility of contamination bias; clinicians Unclear risk could have applied the experience gained (performance bias) All outcomes from consultations with the interventions groups in their consultations with the con- trol group (p 6) Blinding of outcome assessment (detection Low risk Unclear if blinding used but most out- comes were objectively measured and not bias) All outcomes subjective to interpretation Incomplete outcome data (attrition bias) Low risk Table 1 and Figure 1 flow diagram (p 4-5) All outcomes 93 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

97 Kennedy 2002 ( Continued) Unclear risk No information provided Selective reporting (reporting bias) Appears to be free from other risks of bias Other bias Low risk Knops 2014 Methods Randomized to decision aid vs usual care Participants 91 + 87 patients with asymptomatic abdominal aortic aneurys m considering elective surgery vs watchful waiting DA: interactive CD-ROM on options’ outcomes, clinical problem, o utcome probabili- Interventions ties, explicit values clarification Comparator: usual care with regular information Primary outcomes: decisional conflict (baseline, 1, 4, and 10 mont Outcomes hs) Secondary outcomes: patient knowledge (baseline and 1 month), a nxiety (baseline, 1, 4, and 10 months), satisfaction with conversation with the surgeo n (baseline and 1 month) , final treatment choice (10 months), aneurysm rupture (10 months), possible date of surgery (10 months), postoperative morbidity and mortalit y (10 months), physical quality of life (baseline, 1, 4, and 10 months) Notes Trial registration: NTR1524 Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Low risk “Computer-generated bias) randomisation ALEA v.2.2, NKI-AVL, the Netherlands) was performed by the investi- gators” (p 2) Low risk “Computer-generated Allocation concealment (selection bias) randomisation ALEA v.2.2, NKI-AVL, the Netherlands) was performed by the investi- gators” (p 2) Blinding of participants and personnel Low risk “Patients and investigators could not be (performance bias) blinded after group assignment, a factor All outcomes which is inherent to the decision aid and the design of the study. Surgeons and nurses in- volved in the outpatient care of the partic- ipants were blinded to the patient’s alloca- tion group, although patients were not pro- hibited from sharing their allocation with them.” (p 3) 94 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

98 Knops 2014 ( Continued) Outcome measurement is not likely to be Blinding of outcome assessment (detection Low risk bias) influenced by lack of blinding as all out- All outcomes comes were measured objectively using val- idated scales and data retrieved from medial records Low risk Appears to have similar attrition between Incomplete outcome data (attrition bias) All outcomes groups. The proportion of values missing varied from 2% to 9% per outcome mea- sure. Missing values were completed by multiple imputation analysis. If one of the outcome measures had more than 25% missing values, that outcome measure for that patient was excluded from analysis. Therefore, missing data have been handled appropriately (p 3) Unclear risk Selective reporting (reporting bias) Insufficient information to make judgment Other bias High risk “Considerable number of patients could not be included, were not asked to participa- tion, or declined to participate. Selection bias may have occured in patients that were not included” (p 6) “Both patients and surgeons were aware of the aim and subject of the study and could not be blinded to the allocation. It is pos- sible that surgeons in the contributing cen- tres offered more than average information to their patients” (p 6). Performance bias may have been introduced in terms of al- tered communication style Krist 2007 Randomized to decision aid booklet vs decision aid web-based vs usual care Methods Participants 196 + 226 + 75 patients considering prostate cancer screening in t he USA Interventions DA: 4 page pamphlet with options’ outcomes, clinical problem, o utcome probability Comparator: web-site with same information as paper based DA Comparator: usual care Outcomes Primary outcomes: role in decision making Secondary outcomes: knowledge, decisional conflict, time spent d iscussing screening, choice (PSA test ordered) Notes - 95 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

99 Krist 2007 ( Continued) Risk of bias Risk of bias Support for judgement Bias Authors’ judgement Low risk Random sequence generation (selection “[C]oordinator referred to pre-generated bias) randomisation tables to inform the partic- ipant to which arm he was randomised” (p 2) Low risk Atthe time of enrolment, the allocationwas Allocation concealment (selection bias) concealed from the coordinator (p 2) Blinding of participants and personnel Physicians were not blinded - could af- High risk (performance bias) fect decision making process and uptake of All outcomes screening Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) p 3, Results; p 4, Flow diagram Low risk All outcomes Selective reporting (reporting bias) No information provided Unclear risk Unclear risk Other bias Uneven groups but done intentionally, ra- tion of 1:3:3 but appears to be free of other potential biases Kupke 2013 Cluster-randomized trial of 2 groups of dental students to de Methods cision board group and non-decision board group. Patients randomized to students in either group Participants 57 + 36 patients with defect in posterior tooth (Class II defect) considering 6 treatment options, including no therapy Interventions DA (in consultation): options’ outcomes, outcome probabilities Comparator: usual care with discussion of the treatment optio ns Outcomes Knowledge (costs/self-payment,survival rate, characteristics andtreatmenttime)(postin- tervention); overall satisfaction with consultation (postint ervention) Notes Primary outcome not specified Risk of bias Risk of bias 96 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

100 Kupke 2013 ( Continued) Authors’ judgement Bias Support for judgement Randomly assigned by a dice (selection of Random sequence generation (selection Low risk students and patient allocation) (p 20) bias) High risk “The patients were assigned to the students Allocation concealment (selection bias) according to common standards of the uni- versity independently and without know- ing which group the student belonged to.” (p 20) Blinding of participants and personnel Low risk “Patients were assigned to the students in- (performance bias) dependently and without knowing which group the students belonged to” (p 20) All outcomes Unclear risk Insufficient information to judge if blind- Blinding of outcome assessment (detection bias) ing of outcome assessment occurred All outcomes Low risk Similar attribution in both groups; “miss- Incomplete outcome data (attrition bias) All outcomes ing answers were treated as incorrect an- swers, while illegible answers were treated as missing values” (p 22) Selective reporting (reporting bias) No mention of study protocol or trial reg- Unclear risk istration. No way to ensure the outcomes they intended to measure are fully reported High risk Did not adjust for clustering in analysis Other bias Kuppermann 2014 Methods Randomized to decision aid vs usual care Participants 375 + 369 11-week pregnant women who had not yet undergone pren atal screening or diagnostic testing Interventions DA: describes clinical condition, options, outcome probabiliti es, values clarification Comparator: usual care Primary outcomes: invasive prenatal diagnostic testing (3 to 6 months) Outcomes Secondary outcomes: testing strategy undergone (3 to 6 months), knowledge (3 to 6 months), accurate risk perception (procedure related miscarriage , DS affected fetus) (3 to 6 months), decisional conflict (3 to 6 months), decisional regret (3 to 6 months) Notes - 97 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

101 Kuppermann 2014 ( Continued) Risk of bias Risk of bias Support for judgement Bias Authors’ judgement Low risk “A computer generated random allocation Random sequence generation (selection bias) sequence assigned participants to experi- mental groups within permuted blocks of random size, with a 1:1 allocation ratio, stratified by age, clinical site, parity, and in- terviewer” (p 1211) Low risk “The randomization code was not available Allocation concealment (selection bias) to any study-related personnel until data analysis was complete” (p 1211) Low risk “Different research associates facilitated Blinding of participants and personnel (performance bias) baseline and follow-up interviews and med- ical record review to ensure blinding to the All outcomes randomization assignment” (p 1211) Low risk “Different research associates facilitated Blinding of outcome assessment (detection bias) baseline and follow-up interviews and med- All outcomes ical record review to ensure blinding to the randomization assignment” (p 1211) Incomplete outcome data (attrition bias) Similar attrition in both groups. “[A]ll re- Low risk ported analyses were based on a modified All outcomes intention-to-treat sample” (p 1211) Selective reporting (reporting bias) Trial registered Low risk Low risk Other bias Appears to be free of other sources of bias Lam 2013 Methods Randomized to decision aid or standard information booklet a fter initial consultation Participants 138 + 138 women considering breast cancer surgery for early-stag e breast cancer Interventions utcome probabilities, DA: take-home booklet on clinical problem, options’ outcomes, o guidance, explicit values clarification Comparator: standard information booklet Outcomes Primary outcomes: treatment decision making difficulties and d ecisional conflict scale at 1 week post consultation, knowledge at 1-week postconsultat ion, decision regret at 1 month after surgery Secondary outcomes: postoperative psychological distress (anx iety and depression) at 98 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

102 Lam 2013 ( Continued) nths after surgery, 1, 4, and 10 months after surgery, decision regret at 4 and 10 mo treatment decision Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Low risk “Patient assignment to treatment and con- trol arms was performed using a prior bias) computer-generated random-number se- quence” (p 2880) Low risk Allocation concealment (selection bias) “A serially labeled, opaque, sealed-envelope method was used for block randomization” (p 2880) Blinding of participants and personnel Low risk “Two research staff members - one respon- (performance bias) sible for preintervention assessment and blockallocationandthe otherforpostinter- All outcomes vention assessments - ensured that the re- searcher performing follow-up assessments was blinded regarding women’s allocation status.” “Blinding surgeons to allocation status proved impractical.” (p 2880) Blinding of outcome assessment (detection Unclear risk 1 research staff member was responsible for bias) postintervention assessments to ensure that All outcomes the researcher performing follow-up assess- ments was blinded regarding women’s allo- cation status (p 2880) Incomplete outcome data (attrition bias) Does not appear to be missing any outcome Low risk All outcomes data; similar attrition in both groups Selective reporting (reporting bias) Low risk Study protocol available online with pub- lished study Other bias Low risk Does not appear to be subject to other sources of bias 99 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

103 Langston 2010 Methods Randomized to decision aid + coaching vs usual care 114 + 108 women pregnant women in their first trimester conside ring use of contracep- Participants tives in the USA DA: double-sided flip chart on clinical problem, outcome probabil Interventions ities, guidance (ad- ministered by a research assistant), coaching (structured, stand ardized, non-directive contraceptive counselling) + usual care Comparator: usual care Primary outcomes: proportion of participants choosing very ef fective contraceptive Outcomes method (post-DA and consult) sult), adher- Secondary outcomes: actual choice on day of procedure (post-DA and con ence of very effective and/or effective methods at 3 months and a t 6 months (post-DA and consult) Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk “Using a random-number table, we deter- Random sequence generation (selection bias) mined the sequence for 1:1 allocation con- strained by blocks of 10” (p 363, Methods- study procedures) Allocation concealment (selection bias) Low risk “Randomization assignments were sealed inside numbered, opaque envelopes” (p 363, Methods-study procedures) Low risk “No blinding of participants or coordina- Blinding of participants and personnel (performance bias) tors was feasible due to the nature of the All outcomes intervention. Physician-providers did not knowthe participant’sallocationgroup, did not discuss the study with patients, and were asked not to change their counselling” (p 363, Methods-study procedures) Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Unclear risk For “method initiation on the day of All outcomes the procedure” it is only said that the “[p]articipants in the intervention group were not more likely to initiate the re- quested method immediately compared to those in the usual care group”; possible that 100 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

104 Langston 2010 ( Continued) the results contradicted the hypothesis and were excluded for this reason No mention of study protocol; not enough Selective reporting (reporting bias) Unclear risk information to permit judgement Low risk Other bias Appears to be free of other potential biases Laupacis 2006 Methods Randomized to decision aid vs usual care 60 + 60 patients undergoing elective open heart surgery consid ering pre-operative au- Participants tologous blood donation in Canada Interventions outcome probability, DA: audiotape booklet on options’ outcomes, clinical problem, Framework) explicit values clarification, guidance (Ottawa Decision Support Comparator: usual care Outcomes Primary outcomes: knowledge, decisional conflict Secondary outcomes: uptake of option, satisfaction with decisi on making process, satis- faction with decision, accurate risk perceptions Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement “Randomization envelopes were prepared Random sequence generation (selection Low risk bias) centrally by a statistician” (p 2) Low risk Allocation concealment (selection bias) “The envelopes were labeled with identifi- cation numbers and contained a card spec- ifying the patient’s group assignment. The envelopes were opened by the interviewer after completion of the baseline interview.” (p 2) Blinding of participants and personnel No information provided Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation 101 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

105 Laupacis 2006 ( Continued) Low risk Incomplete outcome data (attrition bias) Results, p 4; fig 1, flow diagram All outcomes No information provided Unclear risk Selective reporting (reporting bias) Other bias Low risk Appears to be free of other potential biases LeBlanc 2015 Methods Randomized to decision aid vs individualized score only vs usu al care 32 + 33 + 14 women over 50 years diagnosed with osteopenia or os Participants teoporosis not taking biphosphonates or other prescription medication DA (in consultation): clinical problem, individualized risk of co Interventions ndition, options’ out- comes, guidance Comparator 1: individualized risk Comparator 2: usual care Outcomes flict (immediately post) Primary outcomes: knowledge (immediately post), decisional con , participation in decision-making process (immediately post), d ecision to start (immedi- ately post), adherence (6 months), acceptability (timing not specifi ed), satisfaction with the decision-making process (not specified), quality of life (not sp ecified), time (review of video consultation) Secondary outcome: decision quality (not reported) - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk “Patients were allocated using a computer- Random sequence generation (selection bias) generated sequence that randomized them 1:1:1 in a concealed fashion” (p 5) Allocation concealment (selection bias) Low risk “Patients were allocated using a computer- generated sequence that randomized them 1:1:1 in a concealed fashion” (p 5) Blinding of participants and personnel Low risk “Patients and clinicians were aware of (performance bias) the overall objective, presented as im- provement in communication between pa- All outcomes tients and clinicians during the clinical en- counter, but remained blinded to the spe- cific aims” (p 5) 102 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

106 LeBlanc 2015 ( Continued) “After randomization, only data analysts re- Blinding of outcome assessment (detection Low risk bias) mained blind to allocation” (p 5) All outcomes Low risk Incomplete outcome data (attrition bias) Used intention-to-treat analysis; similar at- All outcomes trition in both groups Selective reporting (reporting bias) Unclear risk Trial registered; Checklists available for CONSORT and protocol. Sample size originally calculated based on adherence but re-calculated for decisional conflict given inability to reach original target High risk Other bias “Possible contamination at the clinician level (i.e. clinician who, having used the decision aid with a prior patient, recreates elements of the decision aid with a sub- sequent patient allocated to receive FRAX alone or usual care) was monitored by a de- tailed review of the available video recorded encounters” (p 5) Legare 2008a Methods Randomized to decision aid vs usual care Participants 45 + 45 women considering use of natural health products for man aging menopausal symptoms Interventions lem, explicit values DA: booklet with worksheet on options’ outcomes, clinical prob ork) clarification, guidance/coaching (Ottawa Decision Support Framew lem (did not address risks Comparator: general information brochure on the clinical prob and benefits) Outcomes Primary outcomes: decisional conflict Secondaryoutcomes:knowledge of natural health productsinge neral (notspecificoption outcomes), preferred choice, values-choice agreement, proportio n undecided Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection The randomization scheme was carried out Low risk bias) by a biostatistician using computer-gener- ated unequal blocks 103 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

107 Legare 2008a ( Continued) Low risk Allocation concealment (selection bias) Sealed opaque envelopes containing 1 or the otherdocuments(aPDA inthe intervention group and a general information brochure in the control group) were prepared by an- other individual, external to the study Blinding of participants and personnel Unclear risk The investigators were blinded but no men- tion of blinding of participants (performance bias) All outcomes Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to to in- All outcomes terpretation See Figure 1 for flow diagram, reason for Incomplete outcome data (attrition bias) Low risk All outcomes loss to follow-up was described Low risk Trial registration Selective reporting (reporting bias) is identifier NCT00325923 Other bias Low risk No statistically significant difference in women’scharacteristicsbetweengroups(Ta- ble 1) Legare 2011 Methods Cluster-randomized to decision aid vs usual care 245 + 214 patients with non-emergent acute respiratory infecti Participants ons considering using antibiotics in Canada Interventions DA (in consultation): pamphlet on options’ outcomes, clinical pro blem, outcome prob- abilities, explicit values clarification, guidance and coaching Comparator: delayed intervention Primary outcomes: Outcomes • Patient outcomes: actual choice (pre and post-DA), perceived decisio n quality (pre and post-DA), decisional conflict (pre and post-DA), decision regr et (pre and post-DA), general health outcomes • ecisional conflict Practitioner outcomes: decision, perceived decision quality, d Secondary outcomes: • Patient outcomes: intention to engage in future SDM (pre and po st-DA), participation in decision making • Practitioner outcomes: intention to engage in future SDM and co mply with clinical practice guidelines Notes - 104 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

108 Legare 2011 ( Continued) Risk of bias Risk of bias Support for judgement Authors’ judgement Bias Low risk “A biostatistician simultaneously ran- Random sequence generation (selection bias) domised all FMGs and allocated them to groups using Internet-based software” (p 99) Allocation concealment (selection bias) Low risk “Using Internet-based software” (p 99) Unclear blinding of participants and per- Blinding of participants and personnel Unclear risk (performance bias) sonnel: only biostatistician was blinded (p 99) All outcomes Low risk Biostatistician who assesses the outcomes Blinding of outcome assessment (detection bias) is blinded, outcomes were objectively mea- sured (p 99) All outcomes Incomplete outcome data (attrition bias) Low risk There appear to be no missing data All outcomes Selective reporting (reporting bias) Low risk No missing pre-specified outcomes Other bias Low risk Appears to be free of other sources of bias Legare 2012 Methods re Cluster-randomized controlled trial to decision aid vs usual ca 239+210 adults and children with with a diagnosis of acute resp Participants iratory infection (e.g., bronchitis, otitis media, pharyngitis, rhinosinusitis) DA (in consultation): pamphlet on options’ outcomes, clinical pro blem, outcome proba- Interventions rticipating physicians also bilities, explicit values clarification, guidance and coaching (pa received training in the form of a 2-hour online tutorial and a 2 -hour on-site interactive workshop) Comparator: usual care Outcomes Primary outcome: use of antibiotics (immediately post consulta tion) Secondary outcomes: decisional conflict (immediately post), contro l preference scale (immediately post), quality of decision (immediately post), adh erence to the decision (2 weeks post), repeat consultation (2 weeks post), decisional regr et (2 weeks post), quality of life (2 weeks post) and intention to engage in SDM in future co nsultations regarding antibiotics for acute respiratory infections (2 weeks post) Notes - 105 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

109 Legare 2012 ( Continued) Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection “A biostatistician used internet-based soft- Low risk bias) ware to simultaneously randomize all 12 family practice teaching units to either the intervention group or control group. The teaching units were stratified according to rural or urban location” (p E728) Low risk “A biostatistician used internet-based soft- Allocation concealment (selection bias) ware to simultaneously randomize all 12 family practice teaching units to either the intervention group or control group. The teaching units were stratified according to rural or urban location” (p E728) “Patients with symptoms suggestive of an Blinding of participants and personnel Low risk acute respiratory infection were initially re- (performance bias) cruited by a RA in the waiting room before All outcomes consultation with a physician” (p E728) Blinding of outcome assessment (detection Low risk “The biostatistician was unaware of group bias) allocation, the researchers and research as- sistants who recruited patients and col- All outcomes lected data were not” and “Statistical analy- sis was performed by a statistician who was unaware of the teaching unit allocations” (p E729) Low risk Incomplete outcome data (attrition bias) No missing outcome data All outcomes Selective reporting (reporting bias) Low risk Protocol registered and published Other bias Low risk “To avoid contamination bias, access to the online tutorial was denied to providers in the control groupduringthe trial”(pE728) Leighl 2011 Methods Randomized to DA + usual care vs usual care Participants 107 + 100 patients diagnosed with metastatic CRC considering advanced chemotherapy in Australia and Canada 106 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

110 Leighl 2011 ( Continued) DA: booklet and audiotape on option’ outcomes, clinical proble m, outcome probabili- Interventions ties, explicit values clarification and guidance (steps in decisio n making + worksheet) Comparator: usual care Primary outcomes: knowledge (post-DA), satisfaction with decisi on (post-DA) Outcomes Secondary outcomes: anxiety (pre and post-DA), satisfaction with consultation (post- t of their in- DA), choice leaning (post-DA), decisional conflict (post-DA). achievemen formation preference (post-DA), participation in decision makin g (post-DA), accept- ability (post-DA), quality of life (post-DA) Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Computer generated randomized lists (p Low risk Random sequence generation (selection bias) 2078, Study design) Allocation concealment (selection bias) Low risk Code concealed in sealed envelopes until time of random assignment (p 2078, Study design) Blinding of participants and personnel Patients not blinded and subjective out- Unclear risk (performance bias) comes may be affected by them knowing their assignment All outcomes All outcomes are not subjected to interpre- Low risk Blinding of outcome assessment (detection bias) tation All outcomes Unclear risk 31% dropout rate, but similar losses across Incomplete outcome data (attrition bias) All outcomes all groups Selective reporting (reporting bias) Unclear risk Protocol not available Other bias Appears to be free of other sources of bias Low risk Lepore 2012 Methods Randomized to decision support intervention (decision coaching by telephone + educa- tional pamphlet) vs control Participants 244 + 246 African American men aged 45-70 in the USA 107 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

111 Lepore 2012 ( Continued) DA: condition-specific educational pamphlet on prostate cancer scr eening and tailored Interventions telephone education on options’ outcomes, explicit values clar ification, others’ opinions, and guidance (decision coaching) le consumption) Comparator: attention control (education on fruit and vegetab postrandomization), Outcomes Primary outcomes: knowledge (pretest and post-test at 8 months (post-test), adherence as decisional conflict (posttest), physician visit to discuss testing congruence between testing intentions and behaviors (post-tes t) sk ratio of testing (post- Secondary outcomes: testing intention (post-test), benefit-to-ri test), PSA screening (post-test), anxiety (pretest and post-test) Notes Trial registration NCT01415375 Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “The principal investigator used a com- Random sequence generation (selection Low risk bias) puter-generated randomization schedule to randomize the participant.” (p 322) Allocation concealment (selection bias) Unclear risk “The principal investigator used a com- puter-generated randomization schedule to randomize the participant and emailed the randomization assignment to the interven- tionist.” (p 322) Blinding of participants and personnel Unclear risk Interventionists were not blind to condi- tion. We can assume that patients were (performance bias) blinded as the study design was a tele- All outcomes phone call for both intervention and con- trol groups (p 322) Blinding of outcome assessment (detection Low risk “Data collectors were blind to condition bias) but the interventionists were not” (p 322) All outcomes Low risk Does not appear to be missing any outcome Incomplete outcome data (attrition bias) All outcomes data Selective reporting (reporting bias) Low risk Appears to have reported on all pre-speci- fied outcomes (protocol) Other bias Low risk Appears to be free of other potential sources of bias 108 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

112 Lerman 1997 Methods Randomized to decision aid vs waiting list control Participants 122 + 114 + 164 women considering BRCA1 gene testing in the USA Interventions blem, outcome prob- DA: education and counselling on options’ outcomes, clinical pro ability, explicit values clarification, others’ opinions, gui dance/coaching Comparator: no intervention Outcomes Primary outcome: preferred option ived personal risk/ben- Secondary outcomes: knowledge, accurate risk perceptions, perce efits/limitations, agreement between values and choice Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias No information provided Unclear risk Random sequence generation (selection bias) Allocation concealment (selection bias) Unclear risk No information provided Blinding of participants and personnel Unclear risk Unclear blinding (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation Of 440 women, 400 completed 1-month Incomplete outcome data (attrition bias) Unclear risk All outcomes follow-up interviews; no reasons provided; baseline data/characteristics included (p 2) Selective reporting (reporting bias) No information provided Unclear risk Other bias Low risk Appears to be free of other potential biases Lewis 2010 Methods Cluster-randomized to decision aid vs usual care Participants SA 211 + 232 patients considering colorectal cancer screening in the U Interventions DA: web-based, DVD and VHS videotape formats + stage targeted brochures (and booster kit if patients had not been screened) on options’ out comes, clinical problem, outcome probabilities, others’ opinion, guidance (encouraged patients to communicate 109 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

113 Lewis 2010 ( Continued) ferences; summary) with their practitioners by asking questions and sharing pre Comparator: usual care using Aetna annual reminders to obtai n CRC screening pletion of col- Knowledge of the age at which screening should begin (post-DA), com Outcomes ost-DA), interest in orectal cancer screening (pre, post-DA), intrusive thoughts (pre, p CRC screening (pre, post-DA), intent to ask provider about screen ing (pre, post-DA), readiness to be screened (pre, post-DA), perceived risk of colon can cer (pre, post-DA), general beliefs about colon cancer (pre, post-DA), fears about colo rectal cancer screening (pre, post-DA), perceptions about whether participants had enou gh information (post- c screening tests (post- DA), whether participants had enough information about specifi ticipate in medical decision DA), willingness to pay for screening tests (post), desire to par (post) Practice level measures: assess CRC screening practices (pre, post-DA), referrals (pre, post-DA), quality improvement initiatives Notes Primary outcome was not specified Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Low risk “Randomisation was done using matched bias) pairs and a blocking procedure.” (p 2, Prac- tice recruitment and randomization sec- tion) “Thus, purposive assignment to treatment Allocation concealment (selection bias) Unclear risk group was used, resulting in a hybrid ran- domisation” (p 3, Practice recruitment and randomization section). There is no men- tion of the effect of this purposive assign- ment on the study Unclear risk As mentioned above, staff used purposive Blinding of participants and personnel assignment and were therefore not blinded, (performance bias) All outcomes but there is no mention of the effect on the study Blinding of outcome assessment (detection Low risk The study did not address this outcome, bias) but outcomes were objectively measured All outcomes Incomplete outcome data (attrition bias) Low risk There appear to be no missing outcome data All outcomes Selective reporting (reporting bias) Unclear risk No mention of study protocol Other bias High risk Unadjusted cluster analysis 110 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

114 Loh 2007 Methods Cluster-randomized to decision aid vs usual care 263 + 142 patients with physician diagnosed depression (cluste r RCT with 30 general Participants practitioners randomized) in Germany DA (in consultation): options’ outcomes, clinical problem, expli cit values clarification, Interventions guidance/coaching Comparator: usual care Outcomes Participation in decision making, adherence, satisfaction wit h clinical care, depression severity, consultation length Primary outcome was not specified Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection “[T]wo-thirds of the general practitioners bias) were randomly assigned to the intervention groupbydrawingblindedlotsunderthe su- pervision of the principal investigator and two researchers” (p 3) Allocation concealment (selection bias) Low risk Drawing blinded lots (p 3 - 2.1) Blinding of participants and personnel Unclear risk Unclear blinding (performance bias) All outcomes Blinding of outcome assessment (detection Unclear risk Unclear blinding, not enough information provided to assess whether this contributes bias) to bias on outcomes not measured by us- All outcomes ing a scale (e.g. consultation time was doc- umented in minutes by the physicians fol- lowing each consultation) Incomplete outcome data (attrition bias) Unclear risk “Further results resting on the baseline All outcomes phase of this trial were already presented elsewhere” (p 5, fig); “unequal distribu- tion of physicians was due to possibility of higher dropout rate in intervention group because of additional time and effort” (p 3) Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry 111 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

115 Loh 2007 ( Continued) Low risk Appears to be free of other potential bi- Other bias ases (p 5-6, details pt and physician base- line characteristics). Statistically significant differences were controlled for in outcome analyses Man-Son-Hing 1999 Methods Randomized to decision aid vs usual care tinuing on aspirin vs change 139 + 148 patients on atrial fibrillation trial considering con Participants to Warfarin in Canada DA: audiotape booklet on options’ outcomes, clinical problem, outcome probability, ex- Interventions plicit values clarification, others’ opinions, guidance (Ottawa Decision Support Frame- work) Comparator: usual care Primary outcomes: uptake of options, adherence Outcomes Secondary outcomes: help with making a decision, knowledge, accu rate risk perceptions, ole in decision making decisional conflict, satisfaction with decision making process, r - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Computer-generated scheme (p 2) Low risk bias) Allocation concealment (selection bias) Administered from a central location (p 2) Low risk High risk Unclear blinding however, “contamination, Blinding of participants and personnel (performance bias) physicians may have provided DA informa- tion to patients receiving usual care” (p 7) All outcomes Low risk Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Unclear risk P 4, fig 2 flow chart. Reasons for attrition not mentioned. Baseline data not included All outcomes Selective reporting (reporting bias) Unclear risk No information provided Other bias Low risk No other potential risks of bias 112 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

116 Mann D 2010 Methods Randomized to decision aid vs usual care 80 + 70 participants diagnosed with diabetes considering the u se of statins to reduce Participants coronary risk DA (in consultation): healthcare provider led discussion using d eveloped tool (Statin Interventions Choice) on options’ outcomes,outcome probabilities, guidance (step-by-step process for making the decision; administered by the physician in the consu ltation) Comparator: usual primary care visit + pamphlet Knowledge (postconsult and post-DA), decisional conflict (postconsu lt and post-DA), Outcomes risk estimation (postconsult and post-DA), beliefs (postconsult and post-DA), adherence (3 and 6 months postconsult and post-DA) Primary outcome was not specified Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Unclear risk Participantswere randomizedbutthere isno Random sequence generation (selection bias) mention of method used (p 138, Methods section) Allocation concealment (selection bias) Unclear risk Not reported Blinding of participants and personnel Unclear risk Not reported (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were not Low risk bias) subjective to interpretation All outcomes Incomplete outcome data (attrition bias) Low risk Baseline data was provided All outcomes Selective reporting (reporting bias) Unclear risk Only reports on improvement (i.e. deci- sional conflict scale); does not present out- come data to fullest (no numerical data on knowledge results between groups, only de- scribes in words) Other bias Unclear risk “We did not adjust the clustering of effects given that few participants received care by the same clinicians” (p 139, Analysis sec- tion). No mention of magnitude in change of data due to this choice 113 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

117 Mann E 2010 Methods Randomized to decision aid vs usual care 278 + 139 participants considering diabetes screening in the UK Participants ities and explicit values Interventions DA: screening invitation on clinical problem, outcome probabil clarification roblem Comparator: usual care using screening invitation on clinical p Outcomes Primary outcomes: preferred option (post-DA) Secondary outcomes: whether invitation type impacts on intent ion (post-DA), impact on knowledge (post-DA), impact on attitude (post-DA), risk percepti on - Notes Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Unclear risk “Invitation taken from the top of a ran- domly ordered pile (either standard or one bias) of two versions of an informed decision choice invitation). The materials were or- dered in a way that the invitation type was hidden until the recruitment process was completed” (p 2-3, Methods, Participants section). Unclear how invitation type was hidden Allocation concealment (selection bias) Low risk “Invitation taken from the top of a ran- domly ordered pile; materials were ordered in a way that the invitation type was hid- den until the recruitment process was com- pleted” (p 2-3, Methods, Participants sec- tion) Low risk Interviewers were not aware of the direction Blinding of participants and personnel of anticipated effect of materials, and ma- (performance bias) All outcomes terials were dummy-coded so that no sense of intervention or control would have been communicated to interviewers or partici- pants (p 3, Methods, Participants section) Blinding of outcome assessment (detection Low risk Studydidnotaddressthisoutcome,butout- comes were objectively measured and not bias) All outcomes subject to interpretation Incomplete outcome data (attrition bias) Low risk No missing outcome data All outcomes 114 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

118 Mann E 2010 ( Continued) Unclear risk No mention of protocol; insufficient infor- Selective reporting (reporting bias) mation to permit judgment Unclear risk “Present sample was ... not necessarily rep- Other bias resentative of the highest risk individuals in this age group”; “£5 incentive might have also added a selection bias”; “Lack of anonymitywith verballydeliveredquestion- naire might encourage socially desirable re- sponding” (p 6, Discussion section) Marteau 2010 Randomized to decision aid vs usual care Methods 633 + 639 patients considering diabetes screening in England Participants DA: screening invitation on clinical problem, outcome probabil Interventions ities and explicit values clarification Comparator: usual care using screening invitation on clinical p roblem Outcomes Primary outcome: attendance for screening (post-DA and consult) Secondary outcomes: intention to make changes to lifestyle (pos t-DA and consult), satisfaction with decisions made among attenders (post-DA and co nsult) Notes - Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Low risk “[G]enerated simultaneously in a batch by random numbers using Excel spreadsheet bias) software, stratifying by number of partici- pants in household” (p 2, Randomization section) Allocation concealment (selection bias) Low risk “Randomisation ... was undertaken by the study statistician from a central site” (p 2, Randomization section) Low risk Personnel were blinded and appears that pa- Blinding of participants and personnel (performance bias) tients were unaware which arm they were in (members of the same household received All outcomes the same intervention) (p 2, Randomization section) 115 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

119 Marteau 2010 ( Continued) Clinical and trial staff taking measurements Blinding of outcome assessment (detection Low risk bias) and entering data were unaware of the study All outcomes arm to which participants had been assigned (p 2, Randomization section) Incomplete outcome data (attrition bias) Low risk No missing outcome data All outcomes Selective reporting (reporting bias) Low risk Published protocol (p 2, Methods) Low risk Other bias Appears free of other potential biases Mathers 2012 Cluster-randomized controlled trial of 49 general practices in Methods the UK to decision aid, healthcare professional training workshop and use of PDA in co nsultation, or usual care 95 + 80 participants with type 2 diabetes considering adding or changing to insulin Participants therapy Interventions DA: booklet about clinical problem, treatment options, optio ns’ outcomes, outcome probabilities, explicit values clarification, structured guid ance Comparator: usual care Outcomes Primary outcomes: decisional conflict (immediately postinterve ntion), glycaemic control (glycosolated haemoglobin, HbA1c) at 6 months Secondary outcomes: knowledge (immediately post), realistic e xpectations (immediately post), preference option (immediately post), proportion undeci ded (immediately post), onths), adherence with participation in decision-making (immediately post), regret (6 m chosen option (6 months) Notes Trial registration: ISRCTN14842077 Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk “All eligible and willing practices were ran- domly allocated by a computer” (p 3) bias) Allocation concealment (selection bias) Low risk “A statistician generated the random allo- cation sequence while a secretary who was not involved in the research study assigned participants to either the intervention or control groups” (p 3) 116 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

120 Mathers 2012 ( Continued) “Blinding of the intervention and assess- Blinding of participants and personnel Unclear risk (performance bias) ment of the process measures were not fea- All outcomes sible in view of the nature of the interven- tion studied” (p 3) Blinding of outcome assessment (detection Unclear risk “Blinding of the intervention and assess- bias) ment of the process measures were not fea- sible in view of the nature of the interven- All outcomes tion studied” (p 3) Incomplete outcome data (attrition bias) Low risk Does not appear to be missing any outcome data All outcomes Low risk Selective reporting (reporting bias) Trial registered Other bias Unclear risk Cannot make a judgment with informa- tion provided regarding cessation of re- cruitment at 175 (yet 320 required to allow detection of 0.5% difference in HbA1c) Mathieu 2007 Methods Randomized to decision aid versus usual care Participants 367 + 367 women aged 70 to 71 years and considering a subsequent screening mam- mography in Australia Interventions bability, explicit val- DA: booklet on options’ outcomes, clinical problem, outcome pro ttawa Decision Support ues clarification, others’ opinions, guidance with worksheet (O Framework) Comparator: BreastScreen NSW brochure - includes information for women 70 + but no numeric information about the outcomes of screening Primary outcomes: actual decision, informed choice Outcomes perceptions), anxiety, Secondary outcomes: knowledge (includes 5 questions about risk decisional conflict, breast cancer worry, preference/intension, attitudes about screening, relationship between objective and perceived risk of breast ca ncer Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Computer programme, which assigned al- Low risk bias) locations in accordance with a simple ran- domization schedule (p 2, Methods) 117 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

121 Mathieu 2007 ( Continued) Low risk Randomized by interview staff who ac- Allocation concealment (selection bias) cessed a previously concealed computer programme (p 2, Methods) Blinding of participants and personnel Unclear risk Unclear blinding (performance bias) All outcomes Interviewers [at follow-up] were blinded, Blinding of outcome assessment (detection Low risk bias) outcomes were objectively measured and All outcomes not subjective to to interpretation Fig 1 flow diagram (p 2) Incomplete outcome data (attrition bias) Low risk All outcomes Low risk “The trial wasregisteredwith the Australian Selective reporting (reporting bias) Clinical TrialsRegistryandthe Clinical Tri- als Registration System” (p 5) Other bias Low risk Appears to be free of other potential biases Mathieu 2010 Methods Randomized to decision aid vs usual care Participants 189 + 223 women considering mammography screening DA: Internet programme + worksheet on options’ outcomes, clin Interventions ical problem, outcome probabilities, explicit values clarification, others’ opinio ns, guidance (worksheet with questions relevant to decision making process; one or more que stions that asked patients to clarify their preferences; summary) Comparator: delayed intervention Outcomes Primary outcomes: knowledge (post-DA), risk perception Secondary outcomes: intention (post-DA), values (post-DA), inform ed choice (post- DA), proportion undecided Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “[C]omputer generated simple randomiza- Low risk bias) tion schedule” (p 66, Randomization and baseline questions section) 118 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

122 Mathieu 2010 ( Continued) Unclear risk “[R]andomization was conducted in a con- Allocation concealment (selection bias) cealed manner” (p 66). Method of alloca- tion concealment not stated Blinding of participants and personnel Unclear risk Not reported (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were not subjective to interpretation bias) All outcomes Incomplete outcome data (attrition bias) Low risk All outcomes mentioned in Outcome mea- sures section were reported in the results sec- All outcomes tion (p 68, Table 2; information for inten- tion as well as anxiety and acceptability can be found in text format in the secondary outcomes section on pg.67-68) Unclear risk No mention of protocol Selective reporting (reporting bias) Low risk Appears to be free of other potential sources Other bias of bias McAlister 2005 Methods Cluster-randomized to decision aid vs usual care Participants 219 + 215 patients considering antithrombotic therapy for no nvalvular atrial fibrillation a (cluster-RCT with 102 primary care practices randomized) in Canad DA: audiotape booklet on options’ outcomes, clinical problem, outcome probabilities, Interventions wa Decision Support Frame- explicit values clarification, others’ opinions, guidance (Otta work) Comparator: usual care Outcomes Primary outcomes: uptake of (appropriate) option Secondary outcomes: knowledge, decisional conflict, accurate risk perceptions Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “[C]luster randomization at level of pri- Low risk bias) mary care practice to minimize contami- nation; randomization was done centrally 119 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

123 McAlister 2005 ( Continued) to preserve allocation concealment using a computer generated sequence” (p 2) Randomization was done centrally to pre- Allocation concealment (selection bias) Low risk serve allocation concealment (p 2, Meth- ods) Blinding of participants and personnel Unclear risk Not blinded, but not sure whether the lack of blinding would affect the outcomes (performance bias) All outcomes Outcome assessors blinded Low risk Blinding of outcome assessment (detection bias) All outcomes Results and Fig 1 - flow diagram (p 3) Incomplete outcome data (attrition bias) Low risk All outcomes Low risk DAAFI trial protocol, including copies of Selective reporting (reporting bias) the various questionnaires we employed, has been published (p 1, Methods) Other bias Low risk Appears to be free of other potential biases McBride 2002 Methods Randomized to decision aid vs usual care Participants 289 + 292 perimenopausal women considering hormone replaceme nt therapy in the USA Interventions values clarification, oth- DA: options’ outcomes, clinical problem, outcome probability, ers’ opinions, guidance/coaching Comparator: delayed intervention Outcomes Primary outcome: accurate risk perceptions Secondary outcomes: satisfaction with decision, confidence with k nowledge and making/ discussing decision Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Unclear risk No information provided; Bastian 2002, no bias) information provided - Study design is de- scribed elsewhere (p 4) 120 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

124 McBride 2002 ( Continued) Unclear risk No information provided; Bastian 2002, no Allocation concealment (selection bias) information provided - Study design is de- scribed elsewhere (p 4) Blinding of participants and personnel Unclear risk Unclear blinding (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Complete data are available for 520 (90%) Unclear risk of the women (p 2). Reasons why not men- All outcomes tioned(Bastian2002, p5,Results; p6,Base- line characteristics/data included) Unclear risk No indication that the trial was registered Selective reporting (reporting bias) in a central trials registry Low risk Appears to be free of other potential bi- Other bias ases; Bastian2002, p8-Eligible participants were willing to consider HRT and this may have favoured recruitment of women with higher SES and those who had prior expe- rience with HRT McCaffery 2010 Methods peat smear Randomized to decision aid + informed choice vs HPV testing vs re Participants 104 + 104 + 106 women screened as HPV indeterminate considering HPV testing in Australia Interventions obabilities, explicit DA: pamphlet on options’ outcomes, clinical problem, outcome pr values clarification, others’ opinion and guidance (worksheet) Comparator 1: no decision support, received immediate HPV tes ting Comparator 2: no decision support, received a repeat cervical sm ear at 6 months Outcomes Primary outcomes: quality of life (post-DA) Secondary outcomes: waiting time anxiety (post-DA), , perceived r isk (post-DA), per- ceived seriousness of cancer (post-DA), worriedness (post-DA), intr usive thoughts (post- DA), satisfaction with care (post-DA), anxiety (post-DA), distress a nd concerns (post- DA), self-esteem (post-DA), effect on sexual behaviour (post-DA), he lp seeking be- haviour (post-DA), knowledge (post-DA) Notes - 121 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

125 McCaffery 2010 ( Continued) Risk of bias Risk of bias Support for judgement Bias Authors’ judgement “Participants were randomised centrally by Random sequence generation (selection Low risk bias) the research team within each clinic in blocks of three” (p 2, Design) Low risk “Participants were randomised centrally by Allocation concealment (selection bias) the research team within each clinic in blocks of three” (p 2, Design) Unclear risk Patients and staff were unblinded, but ob- Blinding of participants and personnel (performance bias) jective outcomes were used All outcomes Blinding of outcome assessment (detection All outcomes are on questionnaires; not Low risk subject to interpretation bias) All outcomes Incomplete outcome data (attrition bias) Figure 3: sensitivity analysis was done to Low risk All outcomes include most of the patients Low risk Selective reporting (reporting bias) Protocol available Other bias Low risk Appears to be free of other sources of bias Miller 2005 Methods Randomized to decision aid vs usual care Participants 279 women considering BRCA1-BRCA2 gene testing in the USA Interventions DA: educational intervention on options’ outcomes, personal family cancer history; clin- ical problem, outcome probability, explicit values clarificatio n, others’ opinions, guid- ance/coaching sk Comparator: provision of general information about cancer ri Outcomes Preferred option, knowledge, perceived risk, satisfaction Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement 122 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

126 Miller 2005 ( Continued) Low risk Random sequence generation (selection “[R]andomized by the CATI system” (p 4) bias) after self-initiated telephone contact “[C]omputerized assisted telephone inter- Low risk Allocation concealment (selection bias) view system (CATI)” (p 4) Unclear risk Blinding was not addressed Blinding of participants and personnel (performance bias) All outcomes Unclear blinding but outcomes were objec- Low risk Blinding of outcome assessment (detection bias) tively measured and not subjective to inter- pretation All outcomes Reasons stated for initial drop-out of study Incomplete outcome data (attrition bias) Low risk All outcomes participants (p 8). Patients contacted of- fered reasons for dropping out. Study pro- tocol allowed patients to be reached up to 13 times at follow-up; but still not able to be reached Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry Other bias Low risk Appears to be free of other sources of bias Miller 2011 Methods Decision aid vs attention placebo Participants 132 + 132 participants considering colon cancer screening in the US A DA: computer-based web programme on options’ outcomes, clinical Interventions problem, outcome probabilities, others’ opinion, guidance (encourages patien t-practitioner communica- tion, summary) Comparator: computer-based web programme on prescription dru g refills and safety Outcomes Primary outcomes: receipt of CRC screening (post-DA) Secondary outcomes: ability to state a preference, change in rea diness to receive screening (pre and post-DA), CRC test ordering (post-DA), proportion undecid ed Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement 123 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

127 Miller 2011 ( Continued) Low risk Random sequence generation (selection Block-randomized, stratified by literacy level (p bias) 609, Methods) Study does not address this domain Unclear risk Allocation concealment (selection bias) Low risk Health care providers were not notified of patients’ Blinding of participants and personnel (performance bias) enrolmentinthe studyatanytime (p609,Methods) All outcomes RAs that administered post-DA questionnaire were not blinded but believed to be a low risk of bias (p 613, Discussion) Low risk “[C]linical outcome assessors were [blinded]” (p Blinding of outcome assessment (detection bias) 613, Discussion) All outcomes Incomplete outcome data (attrition bias) Low risk No missing outcome data All outcomes Selective reporting (reporting bias) Low risk Protocol on ClinicalTrials.gov Other bias Unclear risk USD 10 gift card for participation could affect par- ticipant pool Montgomery 2003 Randomized to decision aid + decision analysis vs decision anal Methods ysis vs decision aid vs usual care Participants 51 + 52 + 55 + 59 newly diagnosed hypertensive patients conside ring drug therapy for blood pressure in the UK DA: decision analysis plus information video and leaflet on op tions’ outcomes, clinical Interventions problem, outcome probability, explicit values clarification Comparator: decision analysis on options’ outcomes, outcome p robability, explicit values clarification Comparator: video and leaflet on options’ outcomes, clinical pr oblem Comparator: usual care Primary outcomes: decisional conflict Outcomes Secondary outcomes: uptake of option, knowledge, anxiety Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement 124 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

128 Montgomery 2003 ( Continued) Random sequence generation (selection Low risk Allocation schedule was computer-gener- bias) ated by an individual not involved in the study (p 2) Allocation concealment (selection bias) “[A]llocation was concealed to the author Low risk in advance by the nature of the minimiza- tion procedure” (p 2) Not blinded - unclear if this would intro- Blinding of participants and personnel Unclear risk duce bias to outcome assessed (performance bias) All outcomes Low risk Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Low risk Flow diagram (p 5) All outcomes Selective reporting (reporting bias) Unclear risk No information provided Other bias Low risk Appears to be free of other potential biases Montgomery 2007 Methods aid without values Randomized to decision aid with values clarification vs decision clarification vs usual care Participants 245 + 250 + 247 women with previous caesarean section in the UK Interventions DA: options’ outcomes, clinical problem, outcome probability, explicit values clarifica- tion Comparator: options’ outcomes, clinical problem, outcome prob ability Comparator: usual care Outcomes Primary outcomes: decisional conflict Secondary outcomes: choice, anxiety, knowledge, satisfaction w ith decision Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Blocked by using randomly permuted and Low risk bias) selected blocks of sizes 6, 9, 12, and 15 generated by computer (p 2 Methods, Ran- 125 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

129 Montgomery 2007 ( Continued) domization) Allocation concealment (selection bias) Low risk 1 member of the study team generated the randomization sequence by computer, and another member of staff with no other in- volvement in the trial performed the allo- cation (p 2 Methods, Randomization) Unclear risk Unclear blinding Blinding of participants and personnel (performance bias) All outcomes Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to to in- terpretation All outcomes See flow of women through the study Incomplete outcome data (attrition bias) Low risk All outcomes Low risk Trials registry ISRCTN84367722 Selective reporting (reporting bias) Low risk Recruited more than planned to account Other bias for lost data (p 4, Sample size); baseline characteristics were balanced Montori 2011 Methods Randomized to decision aid vs usual care + booklet 52 + 48 women with low bone mass or osteoporosis considering ta Participants king bisphosphonates in the USA Interventions DA (in consultation): worksheet on options’ outcomes, clinical pr oblem, outcome prob- abilities, guidance (administered by physician) oporosis Comparator: usual care + general information booklet on oste Outcomes Patient knowledge (post-DA), satisfaction with knowledge tran sfer (post-DA), decisional conflict (post-DA), patient-clinician communication (OPTION), trust wi th physician (during intervention), clinician’s perception of decision quali ty (post-DA), clinician’s satisfaction with knowledge transfer (post-DA), uptake (post-DA ), adherence (post-DA) , fidelity (post-DA), contamination (post-DA), risk perception Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement 126 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

130 Montori 2011 ( Continued) Random sequence generation (selection Low risk “computer generated allocation” (p 551, bias) Randomization) Patients randomized “in a concealed fash- Low risk Allocation concealment (selection bias) ion (using a secure study website)” (p 551, Randomization) Blinding of participants and personnel Unclear risk No mention of participants being blinded (performance bias) to their allocation; only mention of data collectors and analysts blinding (p 551, All outcomes Randomization) Low risk “After randomization, data collectors and Blinding of outcome assessment (detection bias) data analysts were blind to allocation” (p 551, Randomization); Outcomes were not All outcomes subject to interpretation Low risk No missing outcome data Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Low risk “The protocol for this trial has been re- ported in full” (p 550, Design) Other bias Unclear risk Appears to be free of other potential biases Morgan 2000 Methods Randomized to decision aid vs usual care Participants 120 + 120 patients with ischaemic heart disease considering re vascularization surgery in Canada Interventions , clinical problem, out- DA: Health Dialog interactive videodisc on options’ outcomes come probability, others’ opinions Comparator: usual care Outcomes Primary outcome: satisfaction with the decision making process Secondary outcomes: uptake of option, knowledge Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Morgan 1997, p 29: all randomization en- Low risk bias) rolment was performed by telephone at which time the participant was assigned 127 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

131 Morgan 2000 ( Continued) Morgan 2000 (primary study), p 2, Meth- ods, Patient Population: “Only the statis- tician was privy to the two randomisation schedules and blocking factor used” Morgan 1997, p 29: only the statistician was Allocation concealment (selection bias) Low risk privy to the two randomization schedules and blocking factor; Morgan 2000, (primary study), p 2, Meth- ods, Patient Population: “only the statisti- cian was privy to the two randomisation schedules and blocking factor used. All ran- domization enrolment was performed by telephone” Unclear risk “[D]ue to nature of trial, neither patients Blinding of participants and personnel or investigators were blinded to the study” - (performance bias) may introduce bias to subjective outcomes All outcomes such as satisfaction Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation Morgan 1997, p 39, Patient accrual and fol- Incomplete outcome data (attrition bias) Low risk All outcomes low-up: baseline characteristics included Morgan 2000 (primary study): 78% com- pleted follow-up (90 of 120 in the interven- tion; 97 of 120 in the control). reasons for attrition were provided Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry Other bias Unclear risk Morgan 1997, p 56: significant number of patients were lost to follow-up (25%); Mor- gan 2000 (primary study): baseline data im- balance (high school grad, income, no. of diseased arteries). Dropout group reported lower incomes, may have affected results. (discussion par. 6) “Selection bias was mini- mized by enrolling available consecutive pa- tients” 128 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

132 Mott 2014 Methods ual care Randomized to shared decision-making process with DA versus us 13 +14 military veterans in USA diagnosed with PTSD and had se rved in Iraq or Participants Afghanistan DA: booklet on clinical problem, options’ outcomes, structured guidance Interventions Comparator: usual care Outcomes Satisfaction with SDM qualitatively (postintervention), per ceived advantages and dis- t preferences (4 months), advantages of SDM qualitative (postintervention), treatmen adherence using treatment engagement (4 months) Not reported as registered in trials database; no primary ou tcome reported Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “Participants were randomized to SDM or Low risk Random sequence generation (selection bias) UC using a computer-generated random- ization sequence” (p 146) Allocation concealment (selection bias) Low risk “[R]andomizationenvelopeswere prepared by the study statistician to ensure that study staffremainedmaskedtorandomizationse- quence” (p 146) Blinding of participants and personnel Insufficient information provided to make Unclear risk judgment (performance bias) All outcomes Study staff not blinded but because out- Blinding of outcome assessment (detection Low risk bias) comes were taken from medical records. “At 4-month follow-up, study staff re- All outcomes viewed participants’ medical records to ex- tract information on treatment preferences and engagement. Medical-record reviews were conducted by a single rater trained in use of the dataextraction form. A second rater, masked to initial ratings, reextracted data from 20% of patients” (p 146) Incomplete outcome data (attrition bias) High risk 27 participants were consented and en- All outcomes rolled , yet only 20 (UC = 11; SMD = 9) completed the study (p 146-147). Only 5 participants in the SDM arm completed the exit interview. No mention of missing data 129 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

133 Mott 2014 ( Continued) Low risk No protocol available but all expected out- Selective reporting (reporting bias) comes reported on Low risk Does not appear to be any other sources of Other bias bias Mullan 2009 Methods Cluster-randomized to decision aid vs usual care tions (cluster RCT with 48 + 37 patients with type 2 diabetes considering treatment op Participants 40 clinicians randomized) in the USA Interventions , outcomes, outcome DA (in consultation): decision cards with information on options probability, explicit values clarification ions Compare: 12-page pamphlet on oral antihyperglycaemic medicat Knowledge, decisional conflict, participation in decision makin g, acceptability of the Outcomes rust in physician, OP- information, change in medication, adherence, HbA1C levels, t TION to analyse audio-taped encounters Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk Computer generated bias) Central allocation Low risk Allocation concealment (selection bias) Blinding of participants and personnel Unclear risk Patients were blinded, the clinicians were not, but each session was recorded (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation Incomplete outcome data (attrition bias) Unclear risk Reasons for attrition not included All outcomes Selective reporting (reporting bias) Low risk Trial registration no. at clinicaltrials.gov re- ported Other bias Low risk Appears to be free of other sources of bias 130 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

134 Murray 2001a Methods Randomized to decision aid vs usual care 57 + 55 men considering treatment for benign prostatic hypert rophy in the UK Participants , clinical problem, out- Interventions DA: Health Dialog interactive videodisc on options, outcomes come probability, others’ opinions Comparator: usual care Primary outcomes: uptake of option, prostate symptoms, costs , anxiety Outcomes Secondary outcomes: decisional conflict, role in decision making, general health status, utility - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk “[R]andomisation schedule, stratified ac- cording to recruitment centre, was gener- bias) ated by computer” (p 4) Allocation concealment (selection bias) Low risk “Allocation were sealed in opaque num- bered envelopes, opened by the study nurse” (p 4) Not blinded but not sure how this would Unclear risk Blinding of participants and personnel introduce bias (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- tively measured and not subjective to to in- bias) terpretation All outcomes Incomplete outcome data (attrition bias) Low risk Flow diagram (p 5); baseline data/charac- All outcomes teristics included and balanced Selective reporting (reporting bias) No indication that the trial was registered Unclear risk in a central trials registry Other bias Low risk Appears to be free of other sources of bias 131 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

135 Murray 2001b Methods Randomized to decision aid vs usual care 102 + 102 women considering hormone replacement therapy in the UK Participants clinical problem, outcome Interventions DA: Health Dialoginteractive videodiscon options outcomes, probability, other’s opinion Comparator: usual care Outcomes Primary outcomes: preferred option Secondary outcomes: help with making a decision, decisional confl ict, role in decision making anxiety, menopausal symptoms, costs, utility, general heal th status - Notes Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Low risk “[R]andomisation schedule, stratified ac- cording to recruitment centre, was gener- bias) ated by computer” (p 3 Methods, Random- ization) Allocation concealment (selection bias) Low risk “Allocations were sealed in opaque num- bered envelopes, opened by the study nurse after collection of the baseline data” (p 3 Methods, Randomization) Blinding of participants and personnel Unclear blinding Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to to in- All outcomes terpretation Incomplete outcome data (attrition bias) Low risk See page 3 figure for Progress of patients All outcomes through trial Selective reporting (reporting bias) Protocol is not mentioned Unclear risk Other bias Low risk Similar baseline characteristics, appears to be free of other potential biases. Educational achievement was higher in control group. Quote “Subsequent analysis showed that educational level not related to use of HRT nor was there an interaction between edu- cational attainment and the intervention” 132 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

136 Nagle 2008 Methods Cluster-randomized to decision aid vs usual care 167 + 172 women in early pregnancy considering genetic testing (26 + 29 general Participants physicians) (cluster RCT with 60 general practitioners randomi zed) in Australia Interventions ks, test limitations, DA: 24-page booklet and worksheet on options, benefits and ris outcomes; clinical problem, outcome probability, explicit valu es clarification, opinions of others’, guidance (Ottawa Decision Support Framework) Comparator: standard pamphlet on prenatal testing Primary outcomes: informed choice, decisional conflict Outcomes Secondary outcomes: anxiety, depression, attitudes toward p regnancy, acceptability of the intervention, choice - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection Computer-generated random numbers (p bias) 3) Allocation concealment (selection bias) Low risk Computer-generated random numbers by an independent statistician; allocation con- cealment was achieved (p 3) Blinding of participants and personnel Unclear risk “Due to the nature of the intervention, it (performance bias) was not possible to blind women, GP’s or researchers” (p 3); unclear if this would in- All outcomes troduce bias to outcome assessed Low risk Researchers were not blinded but outcomes Blinding of outcome assessment (detection bias) were objectively measured and not subjec- All outcomes tive to interpretation Incomplete outcome data (attrition bias) Results, p 4; Fig 1 - flow diagram, p 5 Low risk All outcomes Low risk Trial Registration - The ADEPT trial was Selective reporting (reporting bias) registered in the UK with Current Con- trolled Trials [ISRCTN22532458] and with the Australian Clinical Trials Registry (No: 012606000234516) (p 4) Other bias Low risk Appears to be free of other potential biases (p 8); selection bias but was adjusted for in analysis 133 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

137 Nassar 2007 Methods Randomized to decision aid vs usual care 102 + 98 women diagnosed with a breech presentation from 34 wee ks gestation consid- Participants ering external cephalic version in Australia DA: 24-page booklet, 30-minute audio-CD and worksheet; clinical problem, outcome Interventions probability, explicit values clarification, opinions of other s’, guidance (Ottawa Decision Support Framework) Comparator: usual care counselling and information on the man agement of breech presentation Outcomes Primary outcomes: knowledge, decisional conflict, anxiety, sat isfaction with the decision, red choice Secondary outcomes: preferred role in decision making, prefer - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Low risk Random sequence generation (selection “[R]andomly generated using computer bias) and stratified by parity and center using ran- dom variable block sizes” (p 2) Allocation concealment (selection bias) Low risk “[P]articipants were randomized by tele- phoning a remote, central location” (p 2) Blinding of participants and personnel Unclear risk Womens were not blinded - unclear if this (performance bias) would introduce bias to outcome assessed All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subjective to inter- bias) All outcomes pretation Incomplete outcome data (attrition bias) Low risk Loss to follow-up because of onset of labour All outcomes or incomplete data forms (p 3). Baseline characteristics are included and equal. Min- imumof 84participantsineach studygroup achieved; p 4 - flow diagram Selective reporting (reporting bias) Low risk ISRCTN14570598 Other bias Low risk “Maternal characteristics and baseline mea- sures of cognitive and affective outcomes were comparable between groups” (p 3 Re- sults, Table 1) “Blinding clinicians and employment of a research midwife to interact with women” 134 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

138 Nassar 2007 ( Continued) (p 6) Oakley 2006 Methods Randomized to decision aid vs usual care reatment options to 16 + 17 postmenopausal women with osteoporosis considering t Participants prevent further bone loss in the UK Interventions outcome probability, ex- DA: audiotape booklet on options’ outcomes, clinical problem, plicit values clarification, others’ opinions, guidance (Ottawa Decision Support Frame- work) Comparator: usual care Outcomes Satisfaction with information, decisional conflict (interventi on group only), improve- ment in adherence Notes Primary outcome was not specified Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Unclear risk No information provided bias) Allocation concealment (selection bias) Low risk Group allocation was done by a third party, unconnected to the study and blinded to the identity of the patients (p 1) Unclear risk Unclear blinding Blinding of participants and personnel (performance bias) All outcomes Blinding of outcome assessment (detection Unclear risk Unclear blinding, some outcomes were as- bias) sessed by open-ended questions, do not know whether this contributes to risk of bias All outcomes Incomplete outcome data (attrition bias) Sample characteristics not included; base- Unclear risk All outcomes line satisfaction score included. “No evalu- ation was carried out to determine the rea- sons for non-participation” (p 2) Selective reporting (reporting bias) Unclear risk No information provided Other bias Unclear risk No baseline characteristics (p 2). Only 16 patients in intervention group and 17 in control group; small sample size 135 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

139 Ozanne 2007 Methods (standard counselling) Randomized to decision aid + standard counselling vs usual care 15 + 15 women considering breast cancer prevention in the USA Participants s outcomes, outcome Interventions DA (in consultation): interactive computer decision aid on option probability Comparator: standard counselling Primary outcomes: consultation length Outcomes Secondary outcomes: knowledge, decisional conflict, satisfactio n with the decision, ac- e consultation ceptability of the decision aid, physician satisfaction with th Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Unclear risk Patients were randomized evenly between groups; no information provided about bias) generation (p 149) Allocation concealment (selection bias) Unclear risk No information provided (p 149) Blinding of participants and personnel Unclear risk Unclear blinding (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to to in- terpretation All outcomes Incomplete outcome data (attrition bias) Low risk Demographic data included; reasons for at- All outcomes trition mentioned Selective reporting (reporting bias) No reference to study protocol Unclear risk Other bias Unclear risk Small sample size, does not say how many physicians participated in study, mentions that there were observed changes in physi- cian behaviour (based on doing both inter- vention and control) 136 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

140 Partin 2004 Methods Randomizedtodecisionaidwith others’ opinionsvsdecision a id withoutothers’ opinions vs usual care 384 + 384 + 384 men considering PSA testing in the USA Participants DA: Health Dialog video on options’ outcomes, clinical problem , outcome probability, Interventions others’ opinions Comparator 1: pamphlet on options’ outcomes, clinical problem , outcome probability Comparator 2: usual care Primary outcomes: knowledge Outcomes Secondary outcomes: preferred option, help with making a decis ion, decisional conflict - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Low risk Using a computer-generated algorithm (p 2) bias) Unclear risk No information provided Allocation concealment (selection bias) Blinding of participants and personnel Low risk “[P]roviders were blinded to the fact that (performance bias) their patients were participating in a trial” All outcomes “coordinator did not have direct contact with subjects” (p 5) Blinding of outcome assessment (detection Low risk “[F]ollow-up interviewers blinded, statisti- bias) cians were not”. Outcomes were objectively All outcomes measured and not subjective to to interpre- tation Incomplete outcome data (attrition bias) Low risk Flow diagram (p 2); reasons for attri- All outcomes tion mentioned and participants balanced across study groups. Sample characteristics included Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry Other bias Low risk Appears to be free of other potential biases 137 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

141 Pignone 2000 Methods Randomized to decision aid vs usual care Participants 125 + 124 adults considering colon cancer screening in the USA Interventions ion DA: video of options’ outcomes, clinical problem, others’ opin Comparator: video on car safety Outcomes Primary outcome: uptake of options Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk “[C]omputerized random number genera- tor” (p 2, Methods, Group assignment) bias) Low risk “[R]andomization was performed centrally Allocation concealment (selection bias) and was not balanced among centers. As- signments were placed in sealed, opaque, sequentially numbered envelopes and were distributed to the three sites” (p 2, Methods, Group assignment) Blinding of participants and personnel Unclear risk “The providers and staff were not blinded (performance bias) to intervention status” “3 to 6 months af- All outcomes ter, different RA blinded to participant in- tervention examined clinic records” (p 2) Does not mention whether patients were blinded; unclear if lack of blinding con- tributed to potential risk of bias Blinding of outcome assessment (detection A different research assistant who was Low risk bias) blinded to participants’ intervention status All outcomes examined participants’ clinic records in a standardized and validated manner to deter- mine whether colon cancer screening tests were actually completed within 3 months of the index visit Incomplete outcome data (attrition bias) Unclear risk Because of an administrative error, 18 con- trols did not complete the second and third All outcomes questionnaires (p 4) Selective reporting (reporting bias) Unclear risk Protocol was not mentioned 138 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

142 Pignone 2000 ( Continued) Low risk Baseline characteristics similar, appear to be Other bias no other potential sources of biases. Mini- mized bias from repeated measurements by administering the same questionnaires to the intervention and control participants Protheroe 2007 Randomized to decision aid vs usual care Methods Participants 60 + 56 women considering treatment options for menorrhagia i n the UK DA: interactive computerized DA on options’ outcomes, clinical p roblem, outcome Interventions probability, explicit values clarification, guidance Comparator: information leaflet Primary outcomes: decisional conflict Outcomes lth outcomes, treatment Secondary outcomes: knowledge, anxiety, condition specific hea preference, undecided - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Computer generated randomization, strat- Low risk bias) ified by practice and minimized according to age (p 2, Methods) Allocation concealment (selection bias) Random allocation was concealed from the Unclear risk individual who was making judgments of eligibility, but the method of concealment was not stated (p 2, Methods) Unclear risk Unclear blinding Blinding of participants and personnel (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subjective to to in- bias) All outcomes terpretation Incomplete outcome data (attrition bias) Low risk Fig 6 flow diagram (p 5); baseline data/char- All outcomes acteristics included and balanced (p 4) Selective reporting (reporting bias) Low risk ISRCTN72253427 139 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

143 Protheroe 2007 ( Continued) Low risk Appears to be free of other potential biases Other bias Rubel 2010 Randomized to pretest + decision aid + post-test vs decision aid + post-test vs pretest + Methods posttest vs posttest A Participants 50 + 50 + 50 + 50 men considering prostate cancer screening in the US babilities, others’ opin- DA: booklet on options’ outcomes, clinical problem, outcome pro Interventions ions + pretest and post-test utcome probabilities, Comparator : booklet on options’ outcomes, clinical problem, o others’ opinions + post-test Comparator: pretest + post-test Comparator: post-test Outcomes l conflict (post-DA), Knowledge (pre, post-DA), decisional anxiety (post-DA), decisiona participation in decision making (pre, post-DA), schema for PSA te sting (pre, post-DA) (post-DA) , perception of quality and interpretation of recommendation Primary outcome was not specified Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Electronically generated random number Low risk bias) sequence (p 309, Study design section) Low risk They were given sealed, sequentially num- Allocation concealment (selection bias) bered packets (p 309, Study design section) Unclear risk Blinding of participants and personnel No mention of blinding (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding, but the outcomes were bias) objectively measured and not subject to in- terpretation All outcomes Incomplete outcome data (attrition bias) Low risk No missing outcome data All outcomes Selective reporting (reporting bias) Low risk Protocol followed CONSORT checklist (p 310, Study design section) Other bias Low risk Appears to be free of other potential biases 140 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

144 Ruffin 2007 Methods Randomized to decision aid vs usual care 87 + 87 community dwelling adults not previously screened for C RC in the USA Participants s, clinical problem, out- Interventions DA: interactive website with information on options’ outcome on, guidance come probability, explicit values clarification, others’ opini Comparator: non-interactive website with information on clin ical problem Primary outcome: uptake of option Outcomes Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Low risk “A block randomisation process pro- Random sequence generation (selection bias) grammed by the study computer support staff and verified by a statistician was used including two strata, race and gender” (p 3) Allocation concealment (selection bias) Unclear risk No information provided Blinding of participants and personnel Both blinded Low risk (performance bias) All outcomes Low risk The investigators, data collectors, data en- Blinding of outcome assessment (detection bias) try, and data analyst were all blinded to study arm assignment All outcomes Incomplete outcome data (attrition bias) Low risk Flow diagram (p 3) All outcomes Selective reporting (reporting bias) Unclear risk No information provided Other bias Low risk Appears to be free of other potential biases Sawka 2012 Randomized to decision aid vs usual care Methods Participants 37 + 37 individuals with early-stage papillary thyroid cancer Interventions DA: web-based decision aid with clinical problem, options’ outco mes, outcome proba- bilities, guidance, printout summary Comparator: usual care (consultation with a specialized head an d neck surgeon, and with 1 or more medical specialist) 141 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

145 Sawka 2012 Continued) ( ntervention) Outcomes Primary outcomes: knowledge (baseline and immediately post i cision (baseline, im- Secondary outcomes: decisional conflict, undecided, treatment de mediately post intervention, 6 to12 months), individual pri marily responsible for the treatment decision (6 to 12 months) Notes Trial registration: NCT01083550 Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Low risk Random sequence generation (selection “Central computerized randomization in a 1:1 ratio was performed at a patient level by bias) using variable block sizes of 2 and 4 (allo- cation designed by a study statistician)” (p 2908) Allocation concealment (selection bias) Low risk “Before the random assignment/testing visit, neither the participant, study staff, investigators, nor treating physicians were aware of the allocation, because it had not yet been assigned” (p 2908) Blinding of participants and personnel Low risk “There was no blinding of participants, (performance bias) study staff, or treating physicians after ran- All outcomes dom assignment was completed” (p 2908) , yet it is unlikely that the outcomes are af- fected by the lack of blinding Unclear risk “There was no blinding of participants, Blinding of outcome assessment (detection bias) study staff, or treating physicians after ran- All outcomes dom assignment was completed. However, the statistician was blinded to the allocation of groups at the time of data analysis.” (p 2908) Incomplete outcome data (attrition bias) There does not appear to be any missing Low risk All outcomes outcome data Selective reporting (reporting bias) Unclear risk Authors state the trial is registered, but no link to trial number Other bias Low risk Appears to be free of other potential sources of bias 142 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

146 Schroy 2011 Methods Randomized to detailed vs simple decision aid vs control in the USA Participants 223 + 212 + 231 average-risk patients considering CRC screening Detailed DA: CRC risk assessment + web-based interactive audi o-visual DA on options’ Interventions outcomes, clinical problem, outcome probabilities, others’ op inion and guidance Comparator 1: web-based decision aid only Comparator 2: usual care using pamphlet Knowledge (pre and post-DA), satisfaction with decision making p rocess (pre and post- Outcomes DA), preferred choice (pre and post-DA) Notes Primary outcome was not specified Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Unclear risk No mention of randomization process bias) Unclear risk No mention of allocation concealment Allocation concealment (selection bias) Blinding of participants and personnel Unclear risk Providers were not blinded, subjective out- (performance bias) comes such as satisfaction with decision- All outcomes making process could have been affected, unclear if participants were blinded Blinding of outcome assessment (detection Assessors not blinded but outcome mea- Low risk sures not believed to be influenced by it bias) All outcomes Incomplete outcome data (attrition bias) Low risk No data appears to be missing All outcomes Selective reporting (reporting bias) Unclear risk No mention of examination of selective outcome reporting or study protocol Other bias Appears to be free of other sources of bias Low risk Schwalm 2012 Methods Randomized to decision aid vs usual care Participants 76 + 74 patients undergoing coronary angiography 143 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

147 Schwalm 2012 ( Continued) DA: booklet on options’ outcomes, clinical problem, outcome pro babilities, explicit Interventions values clarification and guidance Comparator: usual care Outcomes Primary outcomes: decisional conflict Secondary outcomes: knowledge, risk perception, value congrue nt with chosen option Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Computerized random number generator Low risk (p 261, Study design) bias) Low risk Sealed envelopes (p 261, Study design) Allocation concealment (selection bias) Unclear risk Patients and physicians were not blinded to Blinding of participants and personnel (performance bias) the allocation (p 261, Study design) All outcomes Blinding of outcome assessment (detection Unclear if DCS score assessed by unblinded Low risk bias) individuals, but outcomes were objectively measured and not subjective to interpreta- All outcomes tion Incomplete outcome data (attrition bias) Low risk Did not seem to have incomplete data All outcomes Low risk Protocol is available Selective reporting (reporting bias) Low risk Appeared to be free of other biases Other bias Schwartz 2001 Methods Randomized to decision aid vs usual care Participants 181 + 190 Ashkenazi Jewish women considering genetic testing in the USA Interventions , clinical problem DA: 16-page booklet on genetic testing with options’ outcomes Comparator: general information on breast cancer, Understanding Breast Changes: A Health Guide for all Women, published by the National Cancer Institute Outcomes Primary outcome: preferred option Secondary outcomes: knowledge, accurate risk perceptions Notes - 144 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

148 Schwartz 2001 ( Continued) Risk of bias Risk of bias Support for judgement Authors’ judgement Bias Random sequence generation (selection Low risk Computer-generated (p 3) bias) Allocation concealment (selection bias) Unclear risk No information provided Unclear blinding Blinding of participants and personnel Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to inter- All outcomes pretation High retention rate, baseline data and rea- Low risk Incomplete outcome data (attrition bias) sons for lost to follow-up were provided (p All outcomes 2, Participants section) Selective reporting (reporting bias) Unclear risk No information provided Other bias Low risk Appears to be free of other potential biases Schwartz 2009a Methods nselling alone Randomized to decision aid + genetic counselling vs genetic cou 100 + 114 women considering prophylactic mastectomy for being B Participants RCA1/2 mutation carriers in the USA DA: CD-Rom on options’ outcomes, clinical problem, risk communica tion with in- Interventions n, others’ opinion; guidance/ dividually tailored risk graphs, explicit values clarificatio counselling - genetic counselling as usual care (Ottawa Decision Support Framework) Comparator: genetic counselling on benefits and risks of test ing, clinical problem (risk assessment, cancer risks associated with mutations, process of testing and interpretation of results) plus written letter outlining all guidelines an d recommendations Outcomes Primary outcomes: decisional conflict, satisfaction with decisio n, actual choice (risk reduction mastectomy) Secondary outcomes: remaining undecided Notes - Risk of bias Risk of bias 145 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

149 Schwartz 2009a ( Continued) Authors’ judgement Bias Support for judgement Low risk Randomized via computer-generated ran- Random sequence generation (selection bias) dom number in a 1:1 ratio (p 3, Procedure) Allocation concealment (selection bias) Unclear risk No information provided Unclear risk Unclear blinding Blinding of participants and personnel (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to inter- All outcomes pretation Fig. 1 - flow diagram (p 3) Incomplete outcome data (attrition bias) Low risk All outcomes Selective reporting (reporting bias) Unclear risk Protocol not mentioned Other bias Low risk Appears to be free of other sources of bias (p 8) “when variable for not watching DA cd was considered in multivariate models, the results did not change substantively (data not shown)” Sheridan 2006 Methods Randomized to decision aid vs usual care (list of risk factors) 49 + 38 adults with no history of cardiovascular disease in the U SA Participants DA: computerized decision aid on options’ outcomes, outcome pro babilities Interventions Comparator: list of CHD risk factors to present to doctor Outcomes Patient-practitioner communication (e.g. discussion with doctor , specific plan to reduce risk discussed with doctor) Notes Primary outcome was not specified Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk “[C]omputerized random number genera- bias) tor” (p 2) 146 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

150 Sheridan 2006 ( Continued) Low risk “[S]ealed in security envelopes” (p 2) Allocation concealment (selection bias) Participants were blinded but the doctors Blinding of participants and personnel Unclear risk who saw both groups were not (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcome was patient reported bias) All outcomes Results (p 5); Flow diagram (p 10); Baseline Low risk Incomplete outcome data (attrition bias) characteristics/data included All outcomes Low risk Selective reporting (reporting bias) ClinicalTrials.gov NCT00315978 Other bias Low risk Appears to have no other potential risk of bias Sheridan 2011 Randomized to decision aid + tailored messages vs usual care Methods Participants 81 + 79 patients with moderate or high risk for CHD considering CHD prevention strategies in the USA Interventions DA: web-based decision aid on options’ outcomes, clinical proble m, outcome probabil- ities, explicit values clarification and guidance Comparator: usual care using computer programme Preferred choice (post-DA), adherence Outcomes (post-DA), patient Other outcomes (Sheridan 2014): patient-provider communication participation (post-DA), patients perceptions of discussions an d the health care visit patients) (post-DA), preferred choice (baseline and post-DA) (data from 81 +79 Primary outcome was not specified Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “Patients were randomised by study staff Unclear risk bias) who accessed an online randomised sched- ule”(p2).Sequence generationmethodnot stated Allocation concealment (selection bias) Low risk “Patients were randomised by study staff who accessed an online randomised sched- ule” (p 2) 147 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

151 Sheridan 2011 ( Continued) Blinding of participants and personnel Patients blinded and physicians unblinded Low risk but objective outcomes are not likely af- (performance bias) fected by lack of blinding All outcomes Blinding of outcome assessment (detection Low risk Outcomes deemed objective therefore lack of blinding did not influence assessment bias) All outcomes There appears to be no missing data Low risk Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Protocol made available Low risk Low risk Other bias Appears to be free of other sources of bias Shorten 2005 Randomized to decision aid vs usual care Methods Participants ection considering 85 + 84 pregnant women who have experienced previous cesarean s birthing options in Australia Interventions DA: decision aid booklet on options’ outcomes, clinical problem , outcome probabilities, explicit values clarification, guidance (Ottawa Decision Support Framework) Comparator: usual care Outcomes Primary outcomes: knowledge, decisional conflict ion Secondary outcomes: preferred option, help with making a decis - Notes Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Random sequence generation (selection Low risk Computer-based randomized generation (p bias) 3, Procedure) Allocation concealment (selection bias) Low risk “[O]paque envelopes containing a random allocation for each participant code num- ber” (p 3) Blinding of participants and personnel Unclear risk Participants/midwives/ (performance bias) doctors were blinded to patients’ allocation. However, women who used the decision aid All outcomes as specified and in a process of consultation with their midwife or doctor would have negated the blinding of their clinicians, and 148 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

152 Shorten 2005 ( Continued) perhaps of the women themselves. For the intervention group, this may have affected the level and type of information exchanged between them and their caregivers Low risk Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection bias) tively measured and not subjective to to in- terpretation All outcomes Incomplete outcome data (attrition bias) Unclear risk 16 women were lost to follow-up from the intervention group and 18 from the control All outcomes group (no reasons listed) (p 4, Results) Selective reporting (reporting bias) Reference to published protocol Low risk Low risk Other bias Appears to be free of other potential biases Shourie 2013 Cluster-randomized controlled trial of GP practices to web-base Methods d MMR DA + usual care, MMR leaflet + usual care, versus usual care Participants 50 + 93 + 77 parents’ of children facing their first dose MMR vaccina tion Interventions Web-based DA: clinical problem, options’ outcomes, explicit val ues clarification, guid- ance MMR leaflet: Health Scotland leaflet, ’MMR: your questions ans wered’ Comparator: usual care Outcomes ntervention) Primary outcomes: decisional conflict (baseline and 2 weeks posti Secondary outcomes: choice uptake of first dose MMR (when child was 1 5 months), knowledge (baseline and 2 weeks; results not provided), MMR im munization cognitions (baseline and 2 weeks post; results not provided), immunizati on trade-off beliefs (baseline and 2 weeks post; results not provided), anxiety (baseline and 2 weeks post; results not provided), use of the intervention (baseline and 2 weeks post) Notes Trial registration: UK Clinical Research Network - UKCRN ID 48 11 Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection “Simple randomisation using a computer- Low risk bias) generated random list allocated GP prac- tices on a 1:1:1 basis” (p 3) 149 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

153 Shourie 2013 ( Continued) Low risk “An independent researcher who had no Allocation concealment (selection bias) contact with participants generated the al- location sequence and assigned the GP practices to their allocated arm” (p 3) Blinding of participants and personnel Unclear risk “Onreceiptof the completedbaseline ques- (performance bias) tionnaire and consent form, the appropri- All outcomes ate intervention was delivered. At this point the researchers and participants were no longer blind to allocation” (p 3). We don’t know if receiving the intervention had an effect on the ultimate decision that was made Blinding of outcome assessment (detection Outcome data assessment does not depend Low risk bias) on the assessor. It is an objective question- All outcomes naire No missing primary outcome data. Low risk Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Unclear risk Protocol registered. Primary outcome re- ported as stated. Secondary outcomes are not reported (p 3) Other bias Unclear risk Difference in allocation to groups (50 + 93 + 77). Unclear what effect this difference had on the results Smith 2010 Methods Randomized to detailed vs simple decision aid vs usual care 196 + 188 + 188 socioeconomically disadvantaged participants di agnosed with average Participants l cancer screening in or slightly above average risk of bowel cancer considering bowe Australia Interventions DA: booklet + DVD + worksheet + question prompt list on option s’ outcomes, clini- cal problem, outcome probabilities, explicit values clarificati on, guidance (step-by-step process for making the decision; worksheet; encourages patien ts to communicate with practitioners by asking questions; summary) Comparator: booklet + DVD + worksheet on options’ outcomes, cl inical problem, out- come probabilities, explicit values clarification, guidance (ste p-by-step process for mak- ing the decision; worksheet; encourages patients to communicat e with practitioners by asking questions; summary) Comparator: usual care using standard information booklet 150 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

154 Smith 2010 ( Continued) Primary outcomes: values congruent with chosen option (post-DA), participation in Outcomes decision making (pre, post-DA) oice (post-DA), de- Secondary outcomes: knowledge (pre, post-DA), attitude, actual ch in decision mak- cisional conflict (post-DA), decision satisfaction (post-DA), confidence ing (post-DA), general anxiety (post-DA), worry about developing bowel cancer (pre, post-DA), risk perception rticipants) Other outcomes (Smith 2014): screening participation (357 + 173 pa Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement “Participants who verbally consented to Random sequence generation (selection Low risk bias) take part were then randomised to one of the three groups using random permutated blocks of size 6 and 9 for each sex stratum” (p 3, Participants and recruitment section) Low risk Central allocation; “interviewers respon- Allocation concealment (selection bias) sible for recruiting participants were not aware of the randomization sequence or al- location and therefore did not know which intervention respondents would receive” (p 3, Participants and recruitment section) Blinding of participants and personnel “It was not possible for the reviewers to Low risk be blinded to the group allocation. How- (performance bias) All outcomes ever, all questions used standardised word- ing with pre-coded responses and were asked within a supervised environment, where interviewer performances were regu- larly monitored to ensure scripts were read as written” (p 3, Outcome measures sec- tion) Blinding of outcome assessment (detection “[A]nalyses were by intention to treat and Low risk carried out blinded to intervention” (p 5, bias) All outcomes Statistical analysis section); outcomes mea- sured were not subject to interpretation Incomplete outcome data (attrition bias) Low risk Explanation for the missing data reported All outcomes at base of tables Selective reporting (reporting bias) Low risk Study protocol available (ClinicalTri- als.gov NCT00765869 and Australian New Zealand Clinical Trials Registry 151 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

155 Smith 2010 ( Continued) 12608000011381) Other bias Low risk Appears to be free of other potential sources of bias Stacey 2014a Methods Randomized to decision aid vs usual care g joint replacement in Participants 71 + 71 adults diagnosed with knee osteoarthritis considerin Canada DA: DVD + booklet + worksheet on options’ outcomes, clinical pro blem, outcome Interventions n, guidance (1 page summary for probabilities, explicit values clarification, others’ opinio the surgeon) Comparator: usual care Primary outcomes: feasibility (including recruitment, data col Outcomes lection), preliminary ef- fectiveness Secondary outcomes: knowledge (post-DA, pre-surgeon consult), in formed values-con- gruent with chosen option (post-DA, pre-surgeon consult), uptake of chosen option at 1 year; decisional conflict (SURE test), preparation for decision m aking (4 items), wait times Notes Trial registration: NCT00743951 Risk of bias Risk of bias Bias Authors’ judgement Support for judgement “The allocation schedule was computer- Random sequence generation (selection Low risk bias) generated centrally by a statistician using a permuted block design with randomly vary- ing block lengths of 4, 6, or 8.” (p 3) Allocation concealment (selection bias) Low risk “Allocations were concealed in numbered opaque sealed envelopes” (p 3) Blinding of participants and personnel “Patients were not informed of the interven- Low risk tion characteristics” (p 3) (performance bias) All outcomes Blinding of outcome assessment (detection Low risk “Although the research assistant was not bias) blindedtogroupallocation,studyoutcomes All outcomes foreffectivenesswere objective andobtained from clinic data (e.g. date of surgery or wait list status)” (p 3) 152 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

156 Stacey 2014a ( Continued) Low risk Incomplete outcome data (attrition bias) No missing outcome data All outcomes Protocol registered on ClinicalTrials.gov Low risk Selective reporting (reporting bias) Other bias Low risk Appears to be free of other potential sources of bias Steckelberg 2011 Methods Randomized to decision aid vs usual care 785 + 792 patients with no CRC history considering CRC screenin g in Germany Participants DA: brochure on options’ outcomes, clinical problem, and outcome Interventions probabilities Comparator: usual care using pamphlet Outcomes Primary outcomes: values congruent with chosen option (post-DA) Secondary outcomes: knowledge (post-DA), combination of actual an d planned uptake (post-DA), risk perception Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Computer generated sequence (p 2, Ran- Low risk bias) domization and blinding) Allocation was concealed. Identity numbers Low risk Allocation concealment (selection bias) were independent of allocation, and study members did not have access to the data. (p 2, Randomization and blinding) Blinding of participants and personnel Low risk Trial staff who sent out questionnaires and reminders were not aware of study arm, un- (performance bias) clear if participants were blinded (p 2, Ran- All outcomes domization and blinding) Blinding of outcome assessment (detection Low risk Trial staff and statistician who entered data bias) were blinded (p 2, Randomization and All outcomes blinding) Incomplete outcome data (attrition bias) Low risk 12% missing one or both questionnaires All outcomes in intervention group vs 9.2% in control; judged to have low impact on study out- 153 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

157 Steckelberg 2011 ( Continued) come (p 2) Selective reporting (reporting bias) Low risk Protocol available Participants who completed the trial do not Unclear risk Other bias add up Taylor 2006 l Methods Randomized to print DA versus video DA versus wait list contro to consider 98 + 95 + 92 African American men with no history of prostate cancer Participants prostate cancer screening Print DA: clinical problem; outcome probabilities; guidance (li Interventions st of questions to ask at next appointment); others’ opinions Video DA: clinical problem; others’ opinions Wait list comparator: no information provided until 1 month p ostrandomization (base- line assessment for this group coincided with 1-month assessme nt of print and video arms) Prostate cancer screening intention (baseline and 1 month; not r Outcomes eported), prostate screen- ing uptake (1 year; not included because wait list received inter vention before 1 year) process variables including use and perception of the interven tion materials (1 month) , prostate cancer knowledge (baseline and 1 month post), decision al conflict (baseline and 1 month post), satisfaction with screening decision (baselin e and 1 month post) Notes No primary outcome reported; not found in trials registry Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Unclear risk Insufficient information related to random sequence generation bias) Unclear risk Insufficientinformationtojudge allocation Allocation concealment (selection bias) concealment Blinding of participants and personnel Unclear risk Insufficient information to judge blinding; (performance bias) however, participants were requested to not All outcomes share intervention materials with others to prevent contamination between groups (p 2180) Blinding of outcome assessment (detection Insufficient information to judge blinding Unclear risk bias) of outcome assessment All outcomes 154 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

158 Taylor 2006 ( Continued) Low risk Incomplete outcome data (attrition bias) Does not appear to be missing any outcome All outcomes data No protocol registered or published Unclear risk Selective reporting (reporting bias) Other bias “All participants were mailed $25 for their Unclear risk participation following completion of the 1-month interview” (p 2181) “Men who reported that they had not yet had a chance to read/watch the materials were given an additional week to do so and called again to complete the follow-up as- sessment” (p 2181) Thomson 2007 Methods Randomized to decision aid vs usual care by clinical guidelines Participants 69 + 67 patients with atrial fibrillation considering treatme nt options in the UK DA (in consultation): computerized decision on options’ outcomes Interventions , clinical problem, outcome probabilities, explicit values clarification, guidance /coaching by physician Comparator: guidelines applied as direct advice Outcomes Primary outcome: decisional conflict Secondary outcomes: anxiety, knowledge, resource use, choice, h ealth outcomes (stroke, transient ischaemic attack, bleeding events) Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias “[E]lectronically-generated random per- Random sequence generation (selection Low risk bias) muted blocks via a web-based randomisa- tion service” (p 2, Recruitment and ran- domization) Allocation concealment (selection bias) Low risk “[E]lectronically-generated random per- muted blocks via a web-based randomisa- tion service” (p 2, Recruitment and ran- domization) Blinding of participants and personnel Physicianswere blinded. Unclearifpatients Unclear risk (performance bias) are blinded and how that may affect the All outcomes outcome 155 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

159 Thomson 2007 ( Continued) Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to inter- All outcomes pretation See flow diagram Incomplete outcome data (attrition bias) Low risk All outcomes Selective reporting (reporting bias) Low risk ISRCTN24808514 Low risk Baseline characteristics similar, sample size Other bias similar, not stopped early Trevena 2008 Randomized to decision aid vs usual care by consumer guidelines Methods 157 + 157 patients not previously screened for colorectal cancer i n Australia Participants DA: age-gender-family history specific DA booklet with informa tion on options, out- Interventions sonal worksheet with steps come probabilities, explicit values clarification, guidance (per in decision making) (Theory of planned behaviour) Comparator: consumer guidelines recommending faecal occult blo od testing Outcomes Primary outcome: informed choice Secondary outcomes: knowledge, values, screening intention (ch oice); test uptake, anx- iety, acceptability of the intervention, satisfaction with th e decision - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Low risk Random sequence generation (selection “Sequential ID numbers were randomly assigned by computer program to DA or bias) Guidelines (G) in blocks of four” (p 3) Allocation concealment (selection bias) Low risk “Allocation was concealed via the pass- word-protected program” (p 3) Blinding of participants and personnel Unclear risk Participants were blinded to the interven- tion type - not sure about GPs (performance bias) All outcomes Blinding of outcome assessment (detection Researchers were blinded to allocation for Low risk bias) all telephone interviews, outcomes were All outcomes objectively measured 156 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

160 Trevena 2008 ( Continued) Unclear risk Incomplete outcome data (attrition bias) Baseline characteristicsincluded(p3).Fig2 All outcomes flow chart (p 5). Reasons for loss to follow- up not mentioned Low risk ClinicalTrials.gov - NCT00148226 Selective reporting (reporting bias) Low risk Appears to be free of other potential biases Other bias Van Peperstraten 2010 Methods Randomized to decision aid vs usual care 152 + 156 infertile women on wait list for in vitro fertilizat ion in the Netherlands Participants DA: self-administered booklet on options’ outcomes, clinical p Interventions roblem, outcome prob- process for making deci- abilities, explicit values clarification, guidance (step-by-step rocess; 1 or more questions sion, worksheet with questions relevant to decision-making p that asked patients to clarify their preferences; summary to b e shared with practitioner) , coaching (by trained in vitro fertilization nurse) + standard in vitro fertilization care Comparator: standard in vitro fertilization care, including a session in which the number of embryos transferred was discussed Outcomes Primary outcomes: actual choice (postintervention and consult) Secondary outcomes: knowledge (pre, post-DA and consult), empowe rment (pre, post- DA and consult), participation in decision making, decisional con flict (post-DA and n (pre, post-DA and consult), levels of anxiety (pre, post-DA and consult), depressio nsult), condition- consult), cost evaluation of empowerment strategy (post-DA and co specific health outcomes (pregnancies) (post-DA and consult) Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk Computer generated list (p 2, Methods sec- bias) tion) Allocation concealment (selection bias) Central allocation (p 2, Methods section) Low risk Blinding of participants and personnel Low risk “Because of the nature of the intervention (performance bias) it was not possible to blind the participants or in vitro fertilisation doctors to the al- All outcomes location. Participation in our trial did not change the normal in vitro routine.” (p 2, Methods section) 157 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

161 Van Peperstraten 2010 ( Continued) Unclear blinding but outcomes assessed Blinding of outcome assessment (detection Low risk bias) were not subjective to interpretation All outcomes Unclear risk Incomplete outcome data (attrition bias) There are categories in each column of ta- All outcomes ble 1 (p 3) where the denominators do not match the number of people in the group and no reason was given to explain why this would be or if this affects the study Low risk Outcomes same as those registered with Selective reporting (reporting bias) ClinicalTrials.gov Other bias The study appear to be free of other sources Low risk of bias Vandemheen 2009 Randomized to decision aid vs usual care Methods 70 + 79 patients with cystic fibrosis considering referral for l ung transplantation in Participants Canada Interventions DA: self-administered booklet with clinical problem, outcome p robability, explicit values clarification, guidance (Ottawa Decision Support Framework) Comparator: blank pages Outcomes Primary outcomes: knowledge, accurate risk perceptions, decisi onal conflict rability of decision, Secondary outcomes: preparation for decision making, choice, du undecided Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Low risk “[C]omputer-generated random listing of bias) two treatment allocations blocked in blocks of 2 or 4, stratified by site and infection sta- tus of Burkholderia cepacia” (p 2) Low risk Central allocation (p 2) Allocation concealment (selection bias) Blinding of participants and personnel Single blinded RCT; patients and re- Unclear risk (performance bias) searchers were blinded but physicians were All outcomes notbecause theywere involvedwith patients before being randomized 158 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

162 Vandemheen 2009 ( Continued) Research staff, who were blinded to treat- Blinding of outcome assessment (detection Low risk bias) ment allocation, telephoned each patient All outcomes and had them complete a follow-up ques- tionnaire; other outcomes reported are ob- jectively measured Low risk Baseline characteristics included (Flow dia- Incomplete outcome data (attrition bias) All outcomes gram, p 2) Low risk Selective reporting (reporting bias) Clinical trial registered with www.clinical- trials.gov (NCT00345449) Low risk Other bias Appears to be free of other potential biases Vodermaier 2009 Methods Randomized to decision aid vs usual care Participants n Germany 74 + 78 women with breast cancer considering treatment options i Interventions DA: Decision board administered by research psychologists and booklet on options’ outcomes, clinical problem, outcome probability Comparator: booklet on clinical problem Outcomes Primary outcome: decisional conflict n with decision making, Secondary outcomes: choice, length of consultation, satisfactio participation in decision making Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Unclear risk “Randomisation after the patient gave writ- bias) ten informed consent” “Random assign- ment was performed by means of numbered cards in envelopes” “stratified by age group” (p 2) Allocation concealment (selection bias) Low risk “[N]umbered cards in envelopes” (p 2) Blinding of participants and personnel Not blinded - unclear if this would intro- Unclear risk (performance bias) duce bias to outcome assessed All outcomes 159 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

163 Vodermaier 2009 ( Continued) Not blinded but outcomes were objectively Blinding of outcome assessment (detection Low risk bias) measured and not subjective to interpreta- All outcomes tion Flow diagram, p 5; baseline characteristics Incomplete outcome data (attrition bias) Unclear risk All outcomes not included Selective reporting (reporting bias) Unclear risk No information provided Low risk Appears to be free of other potential biases Other bias Volk 1999 Methods Randomized to decision aid vs usual care 80 + 80 men considering PSA testing in the USA Participants DA: Health Dialog videotape and brochure on options’ outcomes , clinical problem, Interventions outcome probability, others’ opinion Comparator: usual care Primary outcomes: knowledge, preferred/uptake of option Outcomes Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Volk 1999 (primary Low risk study), p 3: “[r]andomization by permuted bias) blocks” “Each block included the numbers 1 through 4”; , p 2, Methods: Randomization Volk 2003 by permuted blocks was used to balance the number of subjects in each arm of the study Allocation concealment (selection bias) Unclear risk Volk 1999 (primary study): no information provided Volk 2003, p 2: “[d]etails of the study pro- cedures, subjects, and 2-week follow-up re- sults can be found elsewhere” Blinding of participants and personnel Low risk Participants were not blinded to the treat- (performance bias) ment assignment, but the physicians were; All outcomes therefore outcomes were unlikely to be bi- ased 160 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

164 Volk 1999 ( Continued) Interviewers were not blinded but outcomes Blinding of outcome assessment (detection Low risk bias) were objectively measured and not subjec- All outcomes tive to interpretation Volk 1999 (primary study), p 2, Procedures: Incomplete outcome data (attrition bias) Low risk All outcomes baseline values included Volk 2003, p 4 Fig 1 - flow diagram; baseline data not included Unclear risk No information provided Selective reporting (reporting bias) Low risk Other bias Volk 1999 (primary study): appears to be free of other potential biases Volk 2003: appears to be free of other sources of bias Vuorma 2003 Randomized to decision aid vs usual care Methods Participants and 184 + 179 women considering treatment for menorrhagia in Finl Interventions DA: booklet on options’ outcomes, clinical problem, outcome pro bability Comparator: usual care Outcomes Primary outcomes: uptake of option ed, anxiety, satisfaction, Secondary outcomes: knowledge, proportion remaining undecid health outcomes, use and cost of healthcare services Notes - Risk of bias Risk of bias Bias Support for judgement Authors’ judgement Random sequence generation (selection Low risk Vuorma 2003 (primary study), p 2, Ran- bias) domization: computer-generated; done by a researcher who did not participate in the planning or concealment procedures “[D]one in STAKES, by researcher sepa- rately for each hospital in computer-gener- ated varying clusters”(p 2) Vuorma 2004: no information provided Allocation concealment (selection bias) Low risk Vuorma 2003 (primary study), p 2 “se- quentially numbered, opaque and sealed en- velopes” Vuorma 2004, p 2 “sequentially numbered, 161 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

165 Vuorma 2003 ( Continued) opaque, sealed envelopes” Blinding of participants and personnel Unclear risk No blinding, unclear if measurements could be influenced by lack of blinding (performance bias) All outcomes Low risk Study staff were not blinded but outcomes Blinding of outcome assessment (detection were objectively measured and not subjec- bias) All outcomes tive to interpretation Incomplete outcome data (attrition bias) Vuorma 2003 (primary study): flow chart Low risk balanced. All outcomes Reasonsfornon-eligibility. “One womenon HRT was randomized by mistake and in- cluded in analyses.” Baseline characteristics included and balanced across groups (p 4- 5) Vuorma 2004, flow diagram (p 3) Selective reporting (reporting bias) Unclear risk Vuorma 2003 (primary study): no mention of study protocol Vuorma 2004: no information provided Other bias Vuorma 2003 (primary study), p 7: “in- Low risk crease in knowledge in both study groups, carry-over effect; change in decision-mak- ing process of intervention group may have altered physician’s negotiation with pa- tients” appears to be free of other potential biases Vuorma 2004, p 5: “comparison of the base- line characteristics presented elsewhere” In the pre-trial group compared with the con- trol group, there was a greater increase in the dimensions of physical role functioning and emotional role functioning of the RAND- 36 Watson 2006 Methods Randomized to decision aid vs usual care Participants 475 + 522 men considering prostate cancer screening in the UK Interventions DA: leaflet on options’ outcomes, clinical problem, outcome prob ability Comparator: usual care 162 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

166 Watson 2006 ( Continued) Primary outcomes: knowledge, screening intention, attitude s Outcomes Secondary outcomes: preferred role in decision making - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Low risk “[R]andom numbers generated centrally by Random sequence generation (selection Stata v8.2” (p 3) bias) Allocation concealment (selection bias) Low risk “[R]andom numbers generated centrally by Stata v8.2” (p 3) Blinding of participants and personnel Unclear risk No information provided (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk tively measured and not subjective to inter- bias) pretation All outcomes Incomplete outcome data (attrition bias) Flow diagram (p 2); reason for exclusion Low risk All outcomes from analysis mentioned. Sample character- istics of risk included Selective reporting (reporting bias) No information provided Unclear risk Other bias Unclear risk “Adjustment for multiple testing was not accounted for and hence a degree of caution with interpretation is required, particularly in relation to findings with a P-value close to 0.05” (p 3) Weymiller 2007 Methods Cluster-randomized to decision aid vs usual care Participants 51 + 46 patients with type 2 diabetes in the USA Interventions cal problem, tailored DA (in consultation): 1-page decision aid options’ outcomes, clini outcome probability, guidance/coaching Comparator: booklet on cholesterol management Outcomes Primary outcomes: knowledge, decisional conflict Secondary outcomes: consultation length, acceptability of the i ntervention, adherence, 163 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

167 Weymiller 2007 ( Continued) estimated personal risk, trust, patient participation (OPTI ON), choice Notes - Risk of bias Risk of bias Support for judgement Bias Authors’ judgement Low risk Computer-generated allocation sequence Random sequence generation (selection (p 2) bias) Nannenga 2009: no information provided Low risk Computer-generated allocation sequence, Allocation concealment (selection bias) unavailable to personnel enrolling patients. “[W]ith concealed allocation” (Abstract); “maintained allocation concealment” (p 5) ; randomized by concealed central alloca- tion (Nannenga 2009, p 2) Low risk Participants and clinicians blinded to the Blinding of participants and personnel study objectives, providers and patients (performance bias) All outcomes were naive to this study objective Blinding of outcome assessment (detection Low risk Data analysts and statisticians blinded to bias) allocation; intervention and outcomes; ad- All outcomes equate blinding wherever possible Flow diagram (p 3); reasons for attrition Incomplete outcome data (attrition bias) Low risk All outcomes mentioned(p4);baseline characteristicsin- cluded; flow diagram Nannenga 2009, p 3: reasons for attri- tion mentioned and study groups balanced; baseline characteristics included Selective reporting (reporting bias) Low risk ClinicalTrials.gov identifier: NCT00217061 Other bias Enrollment of patients already receiving Low risk statin therapy and limited statin uptake de- creased the precision of our results; results should best be interpreted as preliminary and requiring verification Nannenga 2009: appears to be free of other potential biases 164 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

168 Whelan 2003 Methods Randomized to decision aid vs usual care ant chemotherapy Participants 82 + 93 women with node negative breast cancer considering adjuv in Canada Interventions DA: decision board and booklet on options’ outcomes, clinical pr oblem, outcome prob- ability, guidance/coaching Comparator: booklet on clinical problem Primary outcomes: knowledge, satisfaction of participant Outcomes Secondary outcomes: preferred option, anxiety, accurate risk p erceptions, participation in decision making - Notes Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Unclear risk No information provided bias) Low risk Randomization, which was performed at a Allocation concealment (selection bias) central location (p 3) Blinding of participants and personnel Unclear risk Unable to blind participants in our trial for (performance bias) practical reasons, measures were taken to All outcomes minimize bias in the design of the study and the assessment of outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation Incomplete outcome data (attrition bias) Unclear risk Flow diagram not included. “[O]ne pa- All outcomes tient excluded from analysis, determined by physician not to be candidate for chemo- therapy” (p 4). Baseline data/characteristics included Selective reporting (reporting bias) Unclear risk Unclear if lack of blinding contributed to potential risk of bias Other bias Low risk Appears to be free of other potential biases 165 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

169 Whelan 2004 Methods Cluster-randomized to decision aid vs usual care 94 + 107 women with Stage 1 or 2 breast cancer considering surgery (cluster-RCT with Participants 27 surgeons randomized) in Canada DA: decision board on options’ outcomes, outcome probability, guidance/coaching Interventions Comparator: usual care Outcomes Primary outcomes: preferred option, knowledge, decisional co nflict, satisfaction Secondary outcomes: accurate risk perceptions, anxiety Notes - Risk of bias Risk of bias Authors’ judgement Bias Support for judgement Doesnotspecifyhowthe sequence wasgen- Random sequence generation (selection Unclear risk bias) erated; a paired cluster randomization pro- cess was used (p 2, Study design and pro- cedures) Allocation concealment (selection bias) Unclear risk Randomly assigned in a concealed fashion, but method of concealment was not stated (p 2, Study design and procedures) Blinding of participants and personnel Unclear risk “[C]hose cluster randomization method to (performance bias) avoid contamination that might have oc- curred if surgeons used decision board for All outcomes some patients and not others” (p 6); un- clear if this would introduce bias Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Baseline characteristics not included; rea- Unclear risk All outcomes sons given for loss of participants Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry Other bias Low risk Appears to be free of other potential biases 166 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

170 Williams 2013 Methods at home or in clinic Randomized to decision aid at home or in clinic versus usual care 134 + 138 + 134 +137 men aged 40-70 years with no history of prost ate cancer who Participants had pre-registered for screening DA: content adapted from the Centers for Disease Control and P revention’s PCS educa- Interventions tional tool. Includes clinical problem, treatment options, ou tcome probabilities, explicit values clarification, others’ stories, summary worksheet Comparator: information booklet. A 3-page fact sheet requiri ng 5 minutes to read. r testing, how to Information presented in a Q&A format on who is recommended fo interpret results, and the limitations of testing Outcomes n with decision Knowledge, decisional conflict, screening outcomes, satisfactio isfaction with decision Outcomes assessed at baseline, 2 months, 13 months, except sat (2 months and 13 months) No primary outcome reported; trial registration not provide d Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Insufficient information to judge random Unclear risk bias) sequence generation Insufficientinformationtojudge allocation Unclear risk Allocation concealment (selection bias) concealment Blinding of participants and personnel Unclear risk Insufficient information to judge blinding of participants and personnel (performance bias) All outcomes Blinding of outcome assessment (detection Unclear risk Insufficient information to judge blinding bias) of outcome assessment All outcomes Low risk There does not appear to be any outcome Incomplete outcome data (attrition bias) All outcomes data missing Selective reporting (reporting bias) Unclear risk No registered or published protocol Other bias Low risk Appears to be free of other potential biases 167 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

171 Wolf 1996 Methods Randomized to decision aid vs usual care Participants 103 + 102 men considering PSA testing in the USA Interventions bility, others’ opinions DA: script of options’ outcomes, clinical problem, outcome proba Comparator: usual care (single sentence) Outcomes Preferred option Notes - Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Unclear risk Wolf 1996 (primary study): no information provided bias) Wolf 1998, p 2: “the methodology of the randomized trial has been reported previ- ously” Allocation concealment (selection bias) Unclear risk Wolf 1996 (primary study): no information provided Wolf 1998, p 2: “The methodology of the randomized trial has been reported previ- ously” Blinding of participants and personnel Unclear blinding Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection Unclear blinding but outcomes were objec- Low risk bias) tively measured and not subjective to inter- All outcomes pretation Wolf 1996 (primary study), p 2: needed a Incomplete outcome data (attrition bias) Low risk All outcomes minimum sample size of 150 participants, and was achieved with total sample size of 205. Reasons for attrition mentioned; base- line characteristics included Wolf 1998: no information provided except that methodology of the randomized trial and the content of the informational inter- vention reported previously (p 2). Baseline characteristics included; flow of participants not included Selective reporting (reporting bias) Unclear risk No indication that the trial was registered in a central trials registry 168 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

172 Wolf 1996 ( Continued) Wolf 1996 (primary study): participant Other bias Low risk population had lower SES therefore external validity of the findings limited, but overall appears to be free of other potential biases Wolf 1998: appears to be free of other po- tential biases Wolf 2000 Randomized to decision aid vs usual care Methods Participants 266 + 133 elderly ( ≥ 65 years) considering CRC screening in the USA DA: script of options’ outcomes, clinical problem, outcome proba bilities Interventions Comparator: usual care (5 sentences) Outcomes Primary outcome: preferred option Secondary outcomes: accurate risk perceptions Notes - Risk of bias Risk of bias Authors’ judgement Support for judgement Bias Random sequence generation (selection Unclear risk “[P]atients were randomised” (p 2); does bias) not indicate how Allocation concealment (selection bias) Unclear risk No information provided Unclear risk Unclear blinding Blinding of participants and personnel (performance bias) All outcomes Unclear blinding but outcomes were objec- Blinding of outcome assessment (detection Low risk bias) tively measured and not subjective to inter- pretation All outcomes Incomplete outcome data (attrition bias) Baseline data not included (p 2, Results) Unclear risk All outcomes Unclear risk Protocol not mentioned Selective reporting (reporting bias) Other bias Low risk Appears to be free of other potential biases 169 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

173 Wong 2006 Methods Randomized to decision aid vs placebo control leaflet Participants 162 + 164 women referred for pregnancy termination in the UK Interventions outcome probability, DA: decision aid leaflet on options’ outcomes, clinical problem, explicit values clarification Comparator: placebo leaflet on contraception use post pregnancy termination flict, anxiety Outcomes Primary outcomes: uptake of option, knowledge, decisional con - Notes Risk of bias Risk of bias Bias Authors’ judgement Support for judgement Random sequence generation (selection Low risk “1:1 ratio, balanced block of 10”; “enve- lope preparation by drawing slips of pa- bias) per labelled either control or intervention”; “the slip determined leaflet placed into en- velope” (p 2) Allocation concealment (selection bias) Low risk Consecutive numbered, opaque trial enve- lope (p 2, Methods) Blinding of participants and personnel Unclear blinding Unclear risk (performance bias) All outcomes Blinding of outcome assessment (detection Low risk Unclear blinding but outcomes were objec- bias) tively measured and not subjective to inter- All outcomes pretation Baseline characteristics not included (p 3) Incomplete outcome data (attrition bias) Unclear risk All outcomes ; reasons for attrition and incompletion mentioned Selective reporting (reporting bias) No information provided Unclear risk Other bias Low risk Appears to be free of other potential biases CHD : coronary heart disease; CRC : colorectal cancer; DA : decision aid; HPV : human papilloma virus; HRT : hormone replacement RCT therapy; : New South Wales; OA : osteoarthritis; PSA : prostate-specific antigen; PTSD : post-traumatic stress disorder; NSW : randomized controlled trial; SES : socioeconomic status. 170 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

174 Characteristics of excluded studies [ordered by study ID] Reason for exclusion Study of the decision aid in one arm of a study Abadie 2009 Study did not evaluate the decision aid (evaluated clinician use only) Hypothetical choice, not at a point of decision making Adab 2003 Study not focused on making a choice; adhering to medications onl y Al Saffar 2008 Alegría 2014 Not a patient decision aid Altiner 2007 Not a patient decision aid Anderson 2011 Not a randomized controlled trial Not a patient decision aid (not including benefits and harms) Arimori 2006 Armstrong 2005 Unable to ascertain whether intervention meets criteria to qu alify as a patient decision aid; additional information requested from author but not provided Arterburn 2013 Not evaluating a patient decision aid Au 2011 Not a randomized controlled trial Bakken 2014 Not a patient decision aid; related to lifestyle choices Becker 2009 Hypothetical choice; not at the point of decision making Not a patient decision aid Belkora 2012 Bellmunt 2010 Not a patient decision aid Bennett 2011 Compares 3 versions of the same patient decision aid Bieber 2006 Study did not evaluate the patient decision aid (evaluated sha red decision-making process); not a patient decision aid Branda 2013 2 patient decision aids with findings aggregated Brenner 2014 Not a patient decision aid Breslin 2008 Not a randomized controlled trial Brown 2004 Not focused on making a choice (no specific decision to be made) Brundage 2001 Not a randomized controlled trial 171 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

175 ( Continued) Not a patient decision aid (general patient education only) Burton 2007 Buzhardt 2011 Not evaluating patient decision making Not evaluating a patient decision aid Campbell 2014 Carling 2008 Hypothetical choice, not at point of decision making Not a patient decision aid Causarano 2015 Chadwick 1991 Not a randomized controlled trial Not a patient decision aid Chan 2011 Not a randomized controlled trial Chewning 1999 Not a randomized controlled trial Chiew 2008 Clouston 2014 Not a patient decision aid Col 2007 Additional information requested from Unable to ascertain characteristics of the patient decision aid. author but not provided (e.g. values clarification) Colella 2004 Not a patient decision aid (describes model of care) Costanza 2011 Not a randomized controlled trial Coulter 2003 Not a randomized controlled trial (editorial) Not a randomized controlled trial Cox 2012 Crang-Svalenius 1996 Not a randomized controlled trial Davison 1999 es clarification) to qualify as a patient Unable to ascertain whether intervention meets criteria (valu decision aid Davison 2007 Not a patient decision aid De Boer 2012 Not a randomized controlled trial De Haan 2013 Not a randomized controlled trial of a patient decision aid Deen 2012 Not a patient decision aid Deinzer 2009 Not a patient decision aid Denig 2014 not a patient decision aid 172 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

176 ( Continued) Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Deschamps 2004 e after 2014 publication) Deyo 2000 Simple versus detailed patient decision aid (excluded in updat Diefenbach 2012 Not a patient decision aid Not focused on making a choice Dobke 2008 Dodin 2001 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Does not report results of the randomized controlled trial; d Donovan 2012 escriptive article offering techniques of provision of information Not a patient decision aid Driscoll 2008 Quasi-RCT: randomization was by days of the week Dunn 1998 Not a decision aid (no decision support) Eaton 2011 Hypothetical choice, not at point of decision making Eden 2009 Eden 2014 The educational brochure (control group) provided information about the options, benefits, and harms making it a simple patient decision aid Eden 2015 Not a treatment or screening decision Hypothetical choice, not a randomized controlled trial Edwards 2012 End of life decision El-Jawahri 2010 Not a randomized controlled trial (Quasi-experimental design ); unclear whether at point of decision Ellison 2008 making Elwyn 2004 No difference in intervention between arms; risk communicatio n did not have values clarification Emery 2007 Not a patient decision aid Emmett 2007 Not a randomized controlled trial Feldman-Stewart 2006 Hypothetical choice, not at point of decision making Feldman-Stewart 2012 Same patient decision aid with vs without values clarification Fiks 2013a Not patient decision making (uptake of vaccine) Flood 1996 Non-randomized allocation; wait list control 173 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

177 ( Continued) Not a patient decision aid Francis 2009 in informed consent for surgery Fraval 2015 Not a patient decision aid; general education material to obta Not a randomized controlled trial Frosch 2001 Frosch 2003 booklet Same decision aid delivered on the Internet versus on DVD plus Frosch 2008b Not a randomized controlled trial Not a patient decision aid Frosch 2011 Qualitative study for an included RCT Frost 2009 Not a patient decision aid and aims to increase screening rates Fujiwara 2015 Hypothetical decision Garvelink 2013 Not a patient decision aid Genz 2012 Not a patient decision aid Giordano 2014 Goel 2001 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Graham 2000 Not a patient decision aid (general information) Gray 2009 Hypothetical choice, not at the point of decision making Not a patient decision aid (educational intervention) Green 2001b Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Green 2004 Not focused on making a choice (intervention to increase patient i Greenfield 1985 nvolvement in care) Griffith 2008a Hypothetical choice, not at the point of decision making Griffith 2008b Not a randomized controlled trial Gruppen 1994 Not a patient decision aid Gummersbach 2015 Not a patient decision aid and a hypothetical decision Hacking 2013 Not a patient decision aid Hall 2007 Not about evaluating a patient decision aid Hall 2011 Not a patient decision aid 174 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

178 ( Continued) not a patient decision aid Hamann 2014 Harmsen 2014 Not a patient decision aid Not a randomized controlled trial Harwood 2011 Healton 1999 Not a patient decision aid (education to promote compliance) Not a treatment or screening decision Henderson 2013 Herrera 1983 Quasi-RCT: assigned to 1 of 2 alternating groups Conjoint analysis for values clarification without informati on on options, pros and cons Hess 2015 Not a patient decision aid; no values clarification Hewison 2001 Not a randomized controlled trial Heyn 2013 Not a randomized controlled trial (letter) Hickish 1995 Not a patient decision aid (general information; no values clar Hochlehnert 2006 ification) Hofbauer 2008 Not a randomized controlled trial Hoffman 2009 Not a patient decision aid Holbrook 2007 Hypothetical choice, not at the point of decision making Hollen 2013 Not a treatment or screening decision Not focused on making a choice (promotes complying with a recommend Holloway 2003 ed option) Holmes-Rovner 2011 Not a randomized controlled trial Study does not evaluate a decision aid; evaluation of spiritu Holt 2009 al versus non-spiritual framework Hope 2010 Same content Huijbregts 2013 Not a patient decision aid Hunt 2005 Not focused on making a choice (promotes complying with a recommend ed option) Hunter 1999 Not focused on making a choice (no specific decision) Hunter 2005 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Huyghe 2009 Hypothetical choice, not at point of decision making for all part icipants 175 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

179 ( Continued) No difference in content of interventions - testing mode of del ivery Ilic 2008 2008 Urologia Internationalis ) Isebaert 2007 Not a randomized controlled trial (English paper published in Not a patient decision aid Jackson 2011 Jerant 2007 Not focused on making a choice - adherence to screening No comparison outcome data provided (only presents data for int Jibaja-Weiss 2006 ervention group) Joosten 2009 Not a patient decision aid Not a patient decision aid Joosten 2011 Jorm 2003 Hypothetical choice, not at point of decision making - community s ample asked to evaluate information booklet on depression Kakkilaya 2011 Hypothetical choice, not at point of decision making Kaplan 2014a Not a patient decision aid Kaplan 2014b Not randomized controlled trial results; cross-sectional anal ysis of baseline data Kassan 2012 Web arm only, not a randomized controlled trial Hypothetical choice, not at point of decision making Kellar 2008 No specific decision to be made and not a true randomized controll ed trial Kiatpongsan 2014 Not a randomized controlled trial; not a patient decision aid Kobelka 2009 Lifestyle only Koelewijn-van Loon 2009 Krawczyk 2012 Uptake of a recommended option Kripalani 2007 Not a patient decision aid Krones 2008 Not a patient decision aid - no benefits and harms Kuppermann 2009 e after 2014 publication) Simple versus detailed patient decision aid (excluded in updat Kurian 2009 Not a randomized controlled trial; not a patient decision aid Köpke 2009 Not a patient decision aid Köpke 2014 Not a patient decision aid 176 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

180 ( Continued) Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Labrecque 2010 t of pilot study LaCroix 1999 Inadequate comparison outcome data provided, secondary repor Lairson 2011 Not a patient decision aid (to increase uptake of screening) Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Lalonde 2006 Lancaster 2009 Not a patient decision aid Not a patient decision aid Landrey 2013 Not a patient decision aid (no values clarification) Lazcano Ponce 2000 Simple versus detailed patient decision aid (excluded in updat Legare 2003 e after 2014 publication) Leung 2004 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Levin 2011 Not a patient decision aid Lewis 2003 Hypothetical choice, not at the point of decision making Lewis 2012 Uptake of a recommended option Lopez-Jornet 2012 Not a patient decision aid/not at point of decision-making Lukens 2013 Not a patient decision aid. Results in response to clinical vign ettes (hypothetical scenarios) Not a randomized controlled trial (all patients received DA) Lurie 2011 Not a patient decision aid (no values clarification) Maisels 1983 Not about evaluating a patient decision aid Mancini 2006 Manne 2009 ) Not focused on making a choice (about adherence not decision making Manns 2005 Not focused on making a choice (Promotes complying with a recommend ed option) Markham 2003 Not a patient decision aid (review of patient information pamp hlets on pre-operative fasting) Martin 2012 Hypothetical choice, not at the point of decision making Maslin 1998 Insufficient outcome data provided in publication; requested f rom author but not provided Matlock 2014 End of life 177 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

181 ( Continued) Not a patient decision aid - genetic counselling only Matloff 2006 Mazur 1994 Hypothetical choice, not at the point of decision making Not a patient decision aid McCaffery 2007 McGinley 2002 Not a patient decision aid (no values clarification) Not a patient decision aid McGowan 2008 McInerney-Leo 2004 Not a patient decision aid (no risk/benefit information; no val ues clarification) Not a patient decision aid; hypothetical choice, not at point of d Mclaren 2012 ecision making Meropol 2013 Not a patient decision aid Unable to ascertain whether intervention meets criteria (valu Michie 1997 es clarification) to qualify as a patient decision aid; additional information requested but author w as unable to provide the intervention Miller 2014a provider No specific decision; related to increasing visits to healthcare Miller 2014b Aims to increase visits to healthcare providers; interventio n targeted to partners Mishel 2009 Not a patient decision aid (information only) Mohammad 2012 Not a patient decision aid; presents only benefits, not harms Not a randomized controlled trial Molenaar 2001 Not a randomized controlled trial (editorial) Mulley 2006 Simple versus detailed patient decision aid (excluded in updat Myers 2005a e after 2014 publication) Myers 2005b Not a randomized controlled trial (editorial) Myers 2007 Not a patient decision aid Myers 2011 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Myers 2013 Uptake of screening Neubeck 2008 Study protocol, does not appear to be patient decision aid Newton 2001 Not a randomized controlled trial O’Cathain 2002 Suite of 8 decision aids (not an efficacy trial) 178 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

182 ( Continued) Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) O’Connor 1999a O’Connor 1996 No patient decision aid - framing effects O’Connor 1998a e after 2014 publication) Simple versus detailed patient decision aid (excluded in updat O’Connor 2009a Not a patient decision aid Not a patient decision aid O’Connor 2011 Owens 2014A Not an RCT; doctoral dissertation Not a patient decision aid Patanwala 2011 Not an RCT Patel 2014 Not a patient decision aid (focus on provision of information) Pearson 2005 Not a patient decision aid (decision aid only supplies mortalit y risk information; no risk info; no values Peele 2005 clarification) Petty 2014 d Not a randomized controlled trial and not a patient decision ai Philip 2010 Not a randomized controlled trial, not a patient decision aid (p romotes complying with a recommended option) Phillips 1995 Quasi-RCT: alternating order based on patients’ initial app ointment sequence Not a patient decision aid; compared the effect of 3 different va Pignone 2013 lues clarification methods Pinto 2008 About clinical trial entry Not a patient decision aid Powers 2011 Not a patient decision aid (general patient education resource) Proctor 2006 Prunty 2008 About a lifestyle choice - whether or not to have a child or have an other child if I have multiple sclerosis Ranta 2015 Not a patient decision aid; intended to increase guideline adh erence for transient ischaemic attack/stroke Rapley 2006 Not a randomized controlled trial Raynes-Greenow 2009 No difference in intervention content; comparison of presenta tion formats; audio-guided decision aid versus booklet only Raynes-Greenow 2010 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) 179 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

183 ( Continued) Not focused on making a choice (promotes complying with a recommend ed option) Rimer 2001 ed option) Rimer 2002 Not focused on making a choice (promotes complying with a recommend Robinson 2013 Not a patient decision aid Benefits or harms of self-testing are not provided as informat ion on the website; values clarification Ronda 2014 exercise asks users to qualify value statements as benefits or harms Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Rostom 2002 Not a patient decision aid Roter 2012 Simple versus detailed patient decision aid (excluded in updat Rothert 1997 e after 2014 publication) Rovner 2004 Not a randomized controlled trial Not a patient decision aid Rubinstein 2011 Not a patient decision aid Ruddy 2009 Ruehlman 2012 Not a patient decision aid Ruland 2013 No specific decision to be made Ryser 2004 Not focused on making a choice (promotes complying with a recommend ed option) Sassen 2014 Not a patient decision aid evaluation study; healthcare profe ssionals were recruited, not patients Not a patient decision aid - general information; not a specific decision Saver 2007 Sawka 2011 Not a randomized controlled trial Scaffidi 2014 Not an RCT Schapira 2000 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Schapira 2007 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Schwartz 2009b Hypothetical choice, not at the point of decision making Sears 2007 About do not resuscitate versus initiating cardiopulmonary r esuscitation decision Sequist 2011 Not a patient decision aid (promotes complying with a recommende d option) Shah 2012 Not a patient decision aid, lifestyle choices 180 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

184 ( Continued) Not a randomized controlled trial Sheppard 2012 Sheridan 2004 Not a randomized controlled trial Hypothetical choice, not at point of decision making Sheridan 2010 Sheridan 2012 Not a patient decision aid - no benefits and harms Sherman 2014 Not a randomized controlled trial Shirai 2012 Not a patient decision aid Silver 2012 Hypothetical choice, not at point of decision making Siminoff 2006 Not a patient decision aid (no discussion of harms) Simon 2012a Not a patient decision aid Not a patient decision aid Simon 2012b Smith 2011a No decision regarding treatment or screening to be made (decisio n regarding full disclosure) Smith 2011b Not a patient decision aid, not an RCT Solberg 2010 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Sorenson 2004 Not a randomized controlled trial Sparano 2006 Not a patient decision aid Not a randomized controlled trial (about instrument developm Stalmeier 2009 ent) Starosta 2015 Not a patient decision aid - benefits and harms of screening are m issing Stein 2013 End of life Steiner 2003 Not a patient decision aid (only effectiveness not cons of option s; not at point of decision making) Stephens 2008 Not a randomized controlled trial Stiggelbout 2008 Not a patient decision aid Stirling 2012 Not a treatment or screening decision Street 1995 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Street 1998 Not focused on making a choice (promotes complying with a recommend ed option) 181 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

185 ( Continued) Hypothetical choice, not at the point of decision making, not a ra ndomized controlled trial Sundaresan 2011 Tabak 1995 Not a randomized controlled trial Taylor 2013 Not a patient decision aid - benefits and harms of screening not i ncluded Not a patient decision aid; about stopping medication use Ten 2008 Thomas 2013 Not a patient decision aid Not a randomized controlled trial; not at point of decision mak ing Thomson 2006 Unable to ascertain whether intervention meets criteria to qu Thornton 1995 alify as a patient decision aid; additional information requested from author but not provided Tiller 2006 e after 2014 publication) Simple versus detailed patient decision aid (excluded in updat Not a patient decision aid Tinsel 2013 Not a patient decision aid - benefits and harms of screening are m issing Tomko 2015 Ukoli 2013 Not an RCT Valdez 2001 Not a randomized controlled trial; not focused on making a choice (complying with a recommended option) Van der Krieke 2013 Not a patient decision aid, no benefits/harms Simple versus detailed patient decision aid (excluded in updat Van Roosmalen 2004 e after 2014 publication) Van Steenkiste 2008 Not a randomized controlled trial Van Til 2009 Hypothetical choice, not at the point of decision making Van Tol-Geerdink 2013 Not a randomized controlled trial; insufficient information t o judge random sequence generation, allocation concealment, and blinding Veroff 2012 Not a patient decision aid Volandes 2009 Advanced care planning options Volandes 2011 Hypothetical choice, end-of-life decision Volandes 2013 Advanced care planning Volk 2008 Simple versus detailed patient decision aid (excluded in updat e after 2014 publication) 182 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

186 ( Continued) Not a randomized controlled trial (commentary) Von Wagner 2011 Wagner 1995 Not a randomized controlled trial e after 2014 publication) Simple versus detailed patient decision aid (excluded in updat Wakefield 2008a Wakefield 2008b e after 2014 publication) Simple versus detailed patient decision aid (excluded in updat Wakefield 2008c simple versus detailed patient decision aid (excluded in updat e after 2014 publication) Wallston 1991 Not a patient decision aid - patient preference study Wang 2004 te genetic counselling process, no focused Not a patient decision aid - intent of intervention to facilita decision Warner 2015 Not a treatment or screening decision Simple versus detailed patient decision aid Watts 2014 Welschen 2012 Not a patient decision aid Wennberg 2010 Same decision aid in both groups Westermann 2013 Not a patient decision aid Weymann 2015 Patients not at the point of decision making Not a patient decision aid Wilhelm 2009 Unable to ascertain characteristics of the patient decision aid. Additional information requested from Wilkes 2013 author but not provided (e.g. values clarification) Wilkie 2013 Not treatment or screening decision Wilkins 2006 Not a randomized controlled trial Willemsen 2006 Lifestyle change Williams-Piehota 2008 Not a randomized controlled trial Williamson 2014 Lifestyle decision - not treatment or screening Woltmann 2011 Not a patient decision aid Wroe 2005 Not focused on making a choice - promotes complying with a recommen ded option Yee 2014 Not a patient decision aid 183 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

187 ( Continued) End-of-life decision Yun 2011 Zajac 2012 Hypothetical dation Not focused on making a choice - promotes complying with a recommen Zapka 2004 Zikmund-Fisher 2008 tion of probabilities No difference in intervention content - comparison of presenta Zoffman 2012 Not a randomized controlled trial, not a patient decision aid [ordered by study ID] Characteristics of ongoing studies ACTRN12615000523505 Trial name or title rthritis increases knowledge and reduces The motherhood choices decision aid for women with rheumatoid a decisional conflict: a randomized controlled study Methods RCT 130 women diagnosed with rheumatoid arthritis and currently under the care of a rheumatologist Participants Interventions Patient decision aid vs usual care Outcomes Decisional conflict, knowledge, anxiety, depression, self-effi cacy Starting date May 2015 Tanya Meade; School of Social Science and Psychology University o Contact information f Western Sydney; Sydney, Australia Notes Trial #: ACTRN12615000523505 ACTRN12615000843550 Trial name or title harms of antibiotic use for coughs and colds in Evaluation of decision aids for parents about the benefits and children Methods Pilot RCT Participants 108 adult parents or primary caregivers of a child Interventions Patient decision aid vs usual care Outcomes Informed choice, knowledge, attitudes towards antibiotic us e, intention to use antibiotic, decisional conflict, confidence in decision-making, usability and accessibility of the written materials Starting date August 2015 184 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

188 ACTRN12615000843550 ( Continued) Mr Peter D Coxeter; [email protected]; Bond University, Queensland, Australia Contact information Notes ACTRN12615000843550 Al-Itejawi 2015 Trial name or title tients with prostate cancer (Cost-)effectiveness and implementation of a decision aid for pa Methods Stepped wedge cluster RCT Newly diagnosed adult participants with localized prostate ca Participants ncer Interventions Patient decision aid vs usual care Decisional conflict,qualityof life, treatmentpreferences, pa rticipationindecisionmaking,knowledge, patient- Outcomes provider communication May 2015 Starting date Contact information Hoda Al-Itejawi; Afdeling Urologie, Amsterdam, the Netherl ands Notes Trial #: NTR5177 Anderson 2014 Trial name or title n Choice Trial (CPC) Shared decision making in the emergency department: Chest Pai RCT Methods Participants Presenting to the emergency department with chest pain Interventions Chest Pain Choice decision aid vs usual care Outcomes Knowledge, patient engagement, decisional conflict, satisfact ion, adverse events, admissions, healthcare uti- lization Starting date October 2013 Contact information Erik P Hess, Mayo Clinic Notes NCT01969240; verified September 2014, estimated study compl etion March 2016 185 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

189 Aslani 2014 Trial name or title Computerized decision aid on mode of delivery Cluster RCT Methods Pregnant Iranian women Participants Interventions Computerized decision aid Outcomes Decisional conflict, knowledge Starting date Not reported Azam Aslani, Mashhad University, Iran Contact information - Notes Buhse 2013 Trial name or title Efficacy of an evidence-based informed shared decision making pro gram for prevention of myocardial infarc- tion in type 2 diabetes Methods RCT Participants 154 patients with type 2 diabetes Interventions Shared decision-making programme consisting of a decision aid b ooklet and a curriculum for group coun- selling vs placebo counselling Outcomes ividual treatment goals, achievement of treatment Knowledge, sustainability of knowledge, achievement of ind e goals prioritized by individual patients, medication uptak Starting date March 2013 Contact information Matthias Lenz, University of Hamburg Notes ISRCTN84636255 Carroll 2012 Trial name or title Development of and feasibility testing of decision support f or patients who are candidates for an implantable defibrillator RCT Methods Participants Referred for consideration of an implantable cardioverter-de fibrillators (non-cardiac resynchronization ther- apy) for a primary prevention indication 186 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

190 Carroll 2012 ( Continued) Patient decision aid provided prior to the consultation with t he physician, which provides a lay summary that Interventions outlines the facts, risks, benefits (including probabilities), specific to the option of an implantable defibrillator or the option of medical management vs usual care Outcomes d decision quality, sure test, reparation for Decision aid development and evaluation, decisional conflict an decision-making scale, medical outcomes trust short form (SF-36v2 ) June 2012 Starting date Contact information Sandra Carroll, McMaster University Trial #: NCT01876173 Notes Chambers 2008 ProsCan for Men: randomized controlled trial of a decision sup port intervention for men with localised Trial name or title prostate cancer Methods RCT Participants 700 men newly diagnosed with localized prostate cancer Interventions A tele-based nurse delivered 5-session decision support/psych osocial intervention vs usual care Outcomes Cancer threat appraisal; decision-related distress and bothe r from treatment side effects; involvement in ization; use of healthcare resources; and a return decision making; satisfaction with health care; heathcare util to previous activities Starting date Not yet assessed Contact information Suzanne K Chambers, Griffith University Notes Trials #: ACTRN012607000233426 Coylewright 2012 Trial name or title Shared decision making in patients with stable coronary arter y disease: PCI Choice Methods RCT Participants - Interventions - Outcomes - Starting date - 187 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

191 Coylewright 2012 ( Continued) Megan Coylewright, Mayo Clinic Contact information Notes Upcoming RCT Cuypers 2015 Prostate cancer patient-centered care: impact of a treatment decis ion aid in a pragmatic, cluster randomized Trial name or title controlled trial Methods Pragmatic RCT Participants 400 men newly diagnosed with early stage prostate cancer Interventions Decision aid (online) vs usual care Outcomes , decision making role, knowledge, satisfaction Decisional conflict, decisional regret, treatment satisfaction , health-related quality of life, personality with decision-making process, preparation for decision-making (anxiety, depression, optimism), skills measures (self-efficacy , health literacy, numeracy) Starting date May 2014 Contact information M Cuypers; [email protected]; Tilburg University Social and Behavioral Sciences Tilburg, the Netherlands Notes NTR4554 Den Ouden 2015 Trial name or title on tool that takes into account clinical factors, Shared decision-making in type 2 diabetes with a support decisi the intensity of treatment and patient preferences Methods Cluster RCT Participants 150 adults with type 2 diabetes mellitus for 8-15 years Interventions Patient decision aid with training vs usual care Outcomes Achievement of diabetes-specific health goals, satisfaction wi th treatment, quality of life, well-being, coping, evidence of shared decision-making March 2012 Starting date Contact information [email protected]; Henk den Ouden; Julius Cntre for H ealth Sciences and Primary Care, University Medical Centre, Utrecht, the Netherlands Notes Trial #: NCT02285881 188 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

192 Dirmaier 2013 Trial name or title r patients with chronic low back pain Tailored, dialogue-based health communication application fo RCT Methods Participants 414 patients with self-reported chronic low back pain s control (standard info) Web-based interactive health communication application (IHCA) v Interventions Outcomes on for decision making, decisional conflict Knowledge, patient empowerment, website usage, preparati Starting date 2012 Martin Härter, University Medical Center Hamburg-Eppendorf Contact information International Clinical Trials Registry DRKS00003322 Notes Geiger 2011 Trial name or title Investigating a training supporting Shared Decision Making (IT’S SDM 2011): study protocol for a random- ized controlled trial RCT Methods Participants 40 physicians that contribute a sequence of 4 medical consultatio ns including a diagnostic or treatment decision Interventions A training curriculum for the doctors - intend to stimulate effo rts to involve their patients in the decision- making process Outcomes Physician-patient communication, effect of SDM on perceived qual ity of the decision process and on the elaboration of the decision, decisional conflict Starting date Not yet assessed Contact information Friedemann Geiger, University Medical Center Schleswig - Hol stein Notes Trials #: ISRCTN78716079 Hersch 2014 Effect of information about over detection of breast cancer on wo men’s decision-making about mammography Trial name or title screening Methods RCT Participants 970 women aged 48-50 189 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

193 Hersch 2014 ( Continued) Intervention (evidence-based information booklet including o ver detection, breast cancer mortality reduction Interventions and false positives) vs control information booklet (includin g mortality reduction and false positives only) Outcomes Knowledge, consistency between attitudes and intentions, de cision conflict, confidence, regret, anxiety, per- ceived risk, quality of life Starting date June 2014 Contact information Kirsten McCaffery, University of Sydney Australian New Zealand Clinical Trials Registry ACTRN12613 Notes 001035718 Hess 2014 Shared decision making in parents of children with head trauma Trial name or title : head CT choice Methods RCT Participants uma above the eyebrows and is positive for 1004 parent-child dyad, seeking care for a child who had blunt tra at least 1 PECARN clinical prediction rules Interventions Patient decision aid vs usual care Outcomes Knowledge, engagement in decision-making process, decisional conflict, trust in the physician, satisfaction with the decision-making process, choice, healthcare utilizatio n 7-days post ER visit, rate of clinically impor- tant traumatic brain injury Starting date April 2014 Contact information Erik Hess; Mayo Clinic; Rochester, MN Trial #: NCT02063087 Notes Jimbo 2012 Trial name or title Decision aid to technologically enhance shared decision making Methods RCT Participants Patients who are not current with colorectal cancer screening Interventions isk assessment) vs web based decision aid Web based decision aid + interactive component (preferences and r only Outcomes Uptake of screening on patient determinants/preference/int ention before the patient-physician encounter, and on shared decision making, concordance and patient intention du ring/after the patient-physician encounter 190 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

194 Jimbo 2012 ( Continued) May 2012 Starting date Contact information Masahito Jimbo, University of Michigan study completion October 2014 Trial # :NCT01514786; last updated December 2013, estimated Notes Layton 2012 Trial name or title Effects of a web-based decision aid on African American men’s pros tate screening knowledge and behavior - Methods 128 African American men Participants - Interventions Outcomes - Starting date - Contact information Beverly Layton, Walden University Notes Unpublished thesis LeBlanc 2013 Trial name or title Translating comparative effectiveness of depression medicat ions into practice by comparing the depression andomized controlled trial medication choice decision aid to usual care: study protocol for a r Methods RCT 300 patients Participants Interventions Use of the Depression Medication Choice decision aid by patient s and their primary care clinician during the clinical encounter vs usual care Outcomes Decisional conflict, knowledge, satisfaction, preference in deci sion making style, patient involvement in decision making, depression outcomes, medication adherence December 2011 Starting date Contact information Victor Montori, Mayo Clinic, USA Notes NCT01502891 191 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

195 Mann 2012 Trial name or title s prevention: rationale and design of the Increasing efficacy of primary care-based counselling for diabete ADAPT (Avoiding Diabetes Thru Action Plan Targeting) trial Methods RCT Participants Primary care providers Interventions Using the ADAPT (Avoiding Diabetes Thru Action Plan Targeting ) system to enhance providers’ effectiveness to counsel about lifestyle behaviour changes Outcome measurements are designed to detect changes in patient behaviours that are most likely to result Outcomes from the use of ADAPT tool: difference between intervention a nd control patients in the change in mean rence of differences in haemoglobin A1C and steps per day at baseline and after 6 months, and 6 month diffe self-reported diet between the 2 groups Starting date Not yet assessed Devin Mann, Boston University School of Medicine Contact information Notes Trial #: NCT01473654 NCT00813033 Trial name or title Use of a patient decision aid for gastrologic endoscopy in a pae diatric setting Methods Interventional efficacy study 80 parents considering gastro-endoscopy for child Participants Not yet assessed Interventions Knowledge, expectations of outcomes, clarity of values, decisi on, decision conflict Outcomes December 2008 Starting date Contact information Nancy Neilan, Children’s Mercy Hospital, Kansas City Notes Trials #: NCT00813033; completed March 2011 NCT01077037 Trial name or title Pain Choice Trial Shared decision making in the emergency department: the Chest Methods RCT Participants 1500 adults admitted to the emergency department for chest pai n, being considered by the treating clinical for admission for cardiac testing 192 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

196 NCT01077037 ( Continued) Patient decision aid vs usual care Interventions on, rate of cardiac testing, etc), patient engagement Outcomes Knowledge, healthcare utilization (rate of hospital admissi ician, satisfaction with decision, safety (major in decision-making process, decisional conflict, trust in the phys adverse cardiac events within 30 days) October 2013 Starting date Contact information [email protected]; Mayo Clinic, Rochester, Minnesota, USA Trial #: NCT01969240 Notes NCT01152294 Trial name or title l symptoms Measuring quality of decisions about treatment of menopausa RCT Methods Participants e menopause or seriously considered taking Patients talked with healthcare provider about ways to manag medicine or supplement to manage menopause Interventions Decision aid (DVD/booklet) vs usual care Outcomes Knowledge, value concordance Starting date June 2010 Karen R Sepucha, Massachusetts General Hospital Contact information Notes NCT01152294; completed, study results on clinicaltrials.gov NCT01152307 Measuring quality of decisions about treatment of depressio n Trial name or title Methods RCT Participants Patients that talked to a healthcare provider about starting or stopping a treatment (prescription medicine for depression or counselling) Decision aid (DVD/booklet) vs usual care Interventions Outcomes Knowledge, value concordance Starting date June 2010 193 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

197 NCT01152307 ( Continued) Karen R Sepucha, Massachusetts General Hospital Contact information NCT01152307; completed, study results on clinicaltrials.gov Notes NCT01447186 Trial name or title Informed decisions about lung cancer screening Methods RCT 500 adults between 55 and 77 years olds who are currently smoki Participants ng or quit within the past 15 years Interventions Patient decision aid vs standard educational information Outcomes Decisional conflict: value subscale and informed subscale Starting date March 2015 Contact information MD Anderson Cancer Center; USA Notes Trial #: NCT02286713 NCT01618097 Trial name or title steoarthritis: focus on health literacy Evaluation of DVD and Internet decision aids for hip and knee o RCT Methods Participants Osteoarthritis patients Interventions DVD decision aid vs Internet-based decision aid Outcomes Decisional conflict, decision self-efficacy, knowledge Starting date January 2012 Contact information Kelli D Allen, Duke University Notes Trial #: NCT01618097; last updated March 2014, study completi on date January 2014 194 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

198 NCT01713894 Trial name or title g extremely premature delivery Utility of a clinically relevant decision aid, for parents facin RCT Methods ity Participants 300 women who are receiving counselling at the limits of viabil Decision aid vs usual care Interventions Decisional conflict, knowledge Outcomes Starting date May 2013 [email protected] ; Ursula Guillen, Christiana Care Health Systems; Universi Contact information ty of Michigan Notes Trial # NCT01713894 NCT01771536 Trial name or title Study to test use of a decision aid in a clinical visit to help pati ents choose a diabetes medication. Translating EP) Information on Comparative Effectiveness Into Practice (TRIC Methods RCT Participants Type 2 diabetes mellitus patients Interventions Diabetes medication decision aid vs usual care Outcomes Patient satisfaction and knowledge. Physician adoption and s atisfaction with the decision aid January 2011 Starting date Nilay D Shah, Mayo Clinic Contact information NCT01293578; estimated completion date December 2014 Notes NCT01851785 Behavioral and social science research on understanding and red ucing health disparities: African American Trial name or title preference for knee replacement: a patient-centred interventi on (ACTION) Methods RCT Participants ith presence of knee OA African-American participants referred to orthopaedic doctor w Interventions Decision aid video + communication, skill-building interventi on vs educational programme (an NIH-devel- oped booklet) that summarizes how to live with knee OA but doe s not mention joint replacement 195 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

199 NCT01851785 ( Continued) Recommendation and receipt of knee joint replacement Outcomes July 2010 Starting date Contact information Said A Ibrahim, University of Pennsylvania Notes Trial #: NCT01851785; last verified May 2013, estimated compl etion date June 2015 NCT01941186 A family centered intervention to promote optimal child devel opment Trial name or title Methods RCT 64 parent-child dyad in which the child is aged 0-36 months screenin Participants g positive for developmental concern Interventions Patient decision aid vs usual care Outcomes owledge, uncertainty, intervention acceptability, in- Evaluation by early intervention specialist, attitudes, kn tervention feasibility Starting date December 2013 Contact information Children’s Hospital of Philadelphia Philadelphia, PN, USA, 19104 Notes Trial #: NCT01941186 NCT01976325 Trial name or title e-travel consultation process Incorporation of the ’Ottawa Malaria Decision Aid’ into the pr RCT Methods 100 adults attending a travel clinic before travelling to an a rea with known chloroquine-resistant malaria Participants Interventions Decision aid vs usual care Outcomes Knowledge, decisional conflict, preparation for decision-makin g, medication adherence Starting date January 2014 Contact information [email protected]; Anne E McCarthy; Ottawa Hospital Research Institute Notes Trial # NCT01976325 196 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

200 NCT02026102 Trial name or title erter-defibrillators (ICDs) A pilot trial of patient decision aids for implantable cardiov RCT Methods ion implantable cardioverter-defibrillators Participants 60 patients with heart failure referred for primary prevent Decision aid toolkit vs usual care Interventions Intervention acceptability, decision quality (knowledge and v alues concordance), quality of life, depressive Outcomes symptoms, health status, spiritual well-being September 2014 Starting date [email protected]; University of Colorado Hospita Contact information l (UCH) Trial #: NCT02026102 Notes NCT02084290 Evaluating a prediction tool and decision aid for patients wit h Crohn’s disease Trial name or title Methods RCT Participants 300 adults with Crohn’s disease Interventions Patient decision aid and SDM programme vs usual care Preferred choice, actual choice, adherence, cost of care, remission Outcomes , patient on steroids, surgeries, Crohn’s disease related hospitalizations Starting date March 2014 Contact information cock Medical Center [email protected]; Corey A Siegel; Dartmouth-Hitch Notes Trial #: NCT02084290 NCT02110979 Trial name or title Validation of a patient decision aid for type 2 diabetes Methods RCT Participants 200 type 2 diabetes patients Interventions Patient decision aid vs usual care Outcomes Knowledge, decisional conflict 197 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

201 NCT02110979 ( Continued) April 2014 Starting date Contact information EPI-Q Inc, Oak Brook, IL, USA, 60523 www.epi-q.com/our-approach Trial #: NCT02110979 Notes NCT02145481 Trial name or title Decisional quality for patients with stable coronary artery d isease RCT Methods Participants 846 adults with stable coronary artery disease Patient decision aid vs standard education Interventions Quality of the decision-making process, knowledge, communicati on, involvement, treatment preferences Outcomes Starting date May 2014 Contact information R. Adams Dudley; University of California, San Francisco Notes Trial # NCT02145481 NCT02198690 Trial name or title Randomized trial of a mammography decision aid for women aged 75 and older RCT Methods 550 women aged 75-89 years Participants Decision aid vs usual care Interventions Receipt of mammography screening, acceptability, anxiety, deci sion-making role, decisional conflict, home Outcomes safety, home safety discussions, knowledge, preparation fo r decision-making, screening discussions, screening intentions Starting date September 2014 Contact information Mara A Schonberg, MD, MPH; [email protected]; Beth I srael Deaconess Medical Center; Boston, MA, USA Notes NCT02198690 198 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

202 NCT02235571 Trial name or title tage renal disease (ESRD) patients iChoose kidney decision aid for treatment options among end-s RCT Methods r and being evaluated for kidney transplant Participants 450 adults with end-stage renal disease on dialysis for < 1 yea Patient decision aid vs usual care Interventions Outcomes lant, treatment preferences Knowledge, evidence of shared decision-making, access to transp Starting date September 2014 Rachel Patzer; Emory Transplant Center; Atlanta, GA, USA Contact information Trial # NCT02235571 Notes NCT02248974 Trial name or title cular assist device (LVAD) placement Development and user testing of a decision aid for left ventri RCT Methods 144 adults who are candidates for a left ventricular assist dev ice Participants Interventions Patient decision aid vs. standard education Outcomes Knowledge, decisional conflict, control preferences scale, Colla boRATE score, perceived quality of care, satisfaction with decision-making process, decisional regret, satisfaction with life, preparation for decision- making, usability and acceptability of the intervention Starting date February 2014 Jennifer Blumenthal-Barby; Baylor College of Medicine; Hous ton, TX Contact information Trial #: NCT02248974 Notes NCT02259699 Trial name or title Ovarian cancer patient-centered decision aid Methods RCT Participants cancer 221 women with stage III optimally debulked advanced ovarian Interventions Patient decision aid vs usual care Outcomes Satisfaction with decision, evidence of shared decision-making , quality of life, satisfaction with care and satisfaction with cancer treatment 199 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

203 NCT02259699 ( Continued) December 2014 Starting date [email protected]; Lari Wenzel; University of California, Ir vine, USA Contact information Notes Trial #: NCT02259699 NCT02308592 Trial name or title Patient decision aid for antidepressant use in pregnancy Methods RCT Participants 50 women aged 18 years or older planning a pregnancy or <30 week s pregnant Interventions Patient decision aid vs standard resource sheet Outcomes depression, anxiety, decisional conflict, knowledge, interve ntion acceptability, choice, satisfaction with DA Starting date January 2015 Contact information [email protected] Women’s College Hospital, Toronto, Ontario, Canada Notes Trial #: NCT02308592 NCT02319525 Trial name or title Methods Participants Interventions Outcomes Starting date Contact information Notes 200 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

204 NCT02326597 Trial name or title Decision aid for therapeutic options in sickle cell disease RCT Methods Participants 120 individuals with sickle cell disease ages 8 to 80 years Decision aid vs usual care Interventions Outcomes expectations, preparation for decision-making, Knowledge, self-efficacy, decisional conflict, values, realistic choice predisposition, stage of decision-making, decisional re gret September 2014 Starting date Contact information [email protected]; principal investigator Lakshmanan Krishnamurti; Emory University, Atlanta, GA, USA Notes Trial # NCT02326597 NCT02344576 A multicentertrial of ashareddecisionsupportintervention forpatients andtheircaregiversoffereddestination Trial name or title therapy for end-stage heart failure Methods RCT Participants 400 adults advanced heart failure and are being evaluated for destination left ventricular assist device Interventions Patient decision aid vs usual care Knowledge, values, decisional conflict, decisional regret, str Outcomes ess, anxiety, depression, quality of life, control preferences scale, illness acceptance, health status May 2015 Starting date [email protected] ; University of Colorado, Denver Contact information Notes Trial #: NCT02344576 NCT02488317 Trial name or title Empowering patients on choices for renal replacement therapy Methods RCT Participants 150 adults with kidney disease Interventions Patient decision aid vs usual care 201 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

205 NCT02488317 ( Continued) Preference forshareddecision-making(CPS), decisional conflict,d ecisionself-efficacy, knowledge, preparation Outcomes for decision making May 2015 Starting date Contact information Francesca Tentori; Arbor Research Collaborative for Health; A nn Arbor, MI Notes Trial #: NCT02488317 NCT02488603 Utilization of decision aids for tamoxifen treatment in brea Trial name or title st cancer patients Methods RCT 360 breast cancer patients referred for tamoxifen treatment Participants Interventions Patient decision aid vs usual care Knowledge, decisional conflict scale, satisfaction with decision , quality of life Outcomes Starting date August 2015 Contact information Eun Sook Lee; National Cancer Center, Korea Notes Trial # NCT02488603 NCT02492009 Trial name or title Patient decision aid for antidepressant use in pregnancy Methods RCT 50 women aged 18 years or older planning a pregnancy or < 30 week s pregnant Participants Interventions Patient decision aid vs standard resource sheet Outcomes Depression, anxiety, decisional conflict, knowledge, interve ntion acceptability, choice Starting date June 2015 Contact information [email protected] or [email protected] Section of Women’s Mental Health, King’s College London Notes Trial #: NCT02492009 202 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

206 NCT02503553 Trial name or title Decision aids in cerebral aneurysm treatment RCT Methods Participants 60 patients undergoing treatment for cerebral aneurysm Patient decision aid vs usual care Interventions Outcomes vels, patient satisfaction level Participation in the shared-decision making process; stress le Starting date August 2015 Kimon Bekelis; Dartmouth-Hitchcock Medical Center; New Hampshi Contact information re, USA Notes Trial #: NCT02503553 NCT02516449 Trial name or title Assessment of shared decision making aids in asthma Methods RCT Participants 51 adults with mild to severe asthma Interventions Patient decision aid vs usual care Outcomes Knowledge, decisional conflict, treatment adherence, asthma con trol Starting date March 2013 Centre de recherche de l’Institut universitaire de cardiologi Contact information e et de pneumologie de Québec, Québec, Canada, G1V 4G5 Trial # NCT02516449 Notes NCT02540044 Trial name or title Supporting patient care with electronic resource (SuPER): efficacy of an online decision aid for patients considering biologic therapy for rheumatoid arthritis Methods RCT Participants s have recommended initiating a biologic/subse- 144 adults with rheumatoid arthritis whose rheumatologist quent entry biologic or switching to another biologic agent Interventions Online patient decision aid vs online standard information 203 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

207 NCT02540044 ( Continued) Decisional conflict, knowledge, self-efficacy, self-management be haviours, health resource utilization, choice, Outcomes evidence of shared decision-making January 2016 Starting date Contact information Linda Li; University of British Columbia; Vancouver, Canada Notes Trial #: NCT02540044 NCT02611050 Treatment decisions for multi-vessel CAD Trial name or title RCT Methods 160 adults with stable multi-vessel CAD at relative equipois e for at least 2 potential treatment options Participants Interventions Option grid decision aid vs usual care Decisional conflict, CollaboRATE score, knowledge, patient exp erience, treatment received Outcomes Starting date December 2015 Contact information Elizabeth L Nichols; the Dartmouth Institute Notes Trial #: NCT02611050 Oostendorp 2011 Trial name or title ncer by providing information with a decision Assessing the information desire of patients with advanced ca aid, which is evaluated in a randomized trial: a study protocol Methods RCT Participants Patients with advanced colorectal, breast, or ovarian cancer and have started treatment with first-line palliative chemotherapy Interventions Patients are randomized to receive either usual care or usual ca re + decision aid Outcomes Not yet assessed Starting date Not yet assessed Contact information Linda JM Oostendorp, Radbound University Notes Netherlands Trial Register (NTR): NTR1113 204 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

208 Yu 2015 Trial name or title l setting decision aid for patients with diabetes Impact of an interprofessional shared decision-making and goa Cluster-randomized controlled trial Methods 112 patients with diabetes Participants Interventions Multicomponent patient decision aid toolkit vs patient educat ion pamphlet Decisional conflict, diabetes distress, health-related qualit Outcomes y of life, chronic illness care, intention to engage in SDM Starting date April 2015 Contact information [email protected] Notes Trial # NCT02379078 CA-125 : cancer antigen 125; CAD : coronary artery disease; CT : computerized tomography; NIH : National Institutes of Health; NSW : : shared decision making. New South Wales; : osteoarthritis; RCT : randomized controlled trial; SDM OA 205 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

209 D A T A A N D A N A L Y S E S Comparison 1. Knowledge No. of No. of Statistical method Effect size Outcome or subgroup title studies participants 13316 Mean Difference (IV, Random, 95% CI) 13.27 [11.32, 15.2 1 Knowledge - all studies 52 3] 52 Mean Difference (IV, Random, 95% CI) Subtotals only 2 Knowledge - subgroup by timing of intervention (in consultation versus in preparation for consultation) 8 922 Mean Difference (IV, Random, 95% CI) 10.57 [4.79, 16.36] 2.1 In consultation 12394 Mean Difference (IV, Random, 95% CI) 13.77 [11.61, 15.9 3] 2.2 In preparation for 44 consultation 12327 47 3 Knowledge - studies without Mean Difference (IV, Random, 95% CI) 13.43 [11.37, 15.4 9] high risk of bias Comparison 2. Accurate risk perceptions No. of No. of Effect size Statistical method Outcome or subgroup title participants studies 5096 Risk Ratio (M-H, Random, 95% CI) 2.10 [1.66, 2.66] 1 Accurate risk perceptions - all 17 studies 2 Accurate risk perceptions Subtotals only Risk Ratio (M-H, Random, 95% CI) 17 - subgroup by timing of intervention (in consultation versus in preparation for consultation) 2.1 In consultation 6 898 Risk Ratio (M-H, Random, 95% CI) 1.79 [1.28, 2.52] 2.2 In preparation for 11 4198 Risk Ratio (M-H, Random, 95% CI) 2.25 [1.65, 3.07] consultation 3 Accurate risk perceptions - 2.02 [1.57, 2.59] 15 4732 Risk Ratio (M-H, Random, 95% CI) studies without high risk of bias 206 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

210 Comparison 3. Informed values-choice congruence No. of No. of Effect size Statistical method Outcome or subgroup title participants studies 10 1 Informed values-choice 2.06 [1.46, 2.91] 4626 Risk Ratio (M-H, Random, 95% CI) congruence - all studies 8 Risk Ratio (M-H, Random, 95% CI) 2.13 [1.44, 3.14] 2 Informed values-choice 4154 congruence - actual choice only 3 Informed values-chose 4365 Risk Ratio (M-H, Random, 95% CI) 2.08 [1.40, 3.08] 8 congruence -using MMIC 2 261 Risk Ratio (M-H, Random, 95% CI) 2.02 [1.44, 2.83] 4 Informed values-chose congruence - heterogeneous measures 4626 2.06 [1.46, 2.91] 10 5 Informed values-choice Risk Ratio (M-H, Random, 95% CI) congruence - without studies of high risk of bias Comparison 4. Decisional conflict No. of No. of Statistical method Effect size Outcome or subgroup title participants studies 42 Mean Difference (IV, Random, 95% CI) Subtotals only 1 Decisional conflict - all studies Mean Difference (IV, Random, 95% CI) 38 8785 1.1 Total decisional conflict -7.22 [-9.12, -5.31] score 1.2 Uninformed subscale 5707 Mean Difference (IV, Random, 95% CI) -9.28 [-12.20, -6.36] 27 23 5068 1.3 Unclear values subscale -8.81 [-11.99, -5.63] Mean Difference (IV, Random, 95% CI) 1.4 Uncertainty subscale 28 6200 Mean Difference (IV, Random, 95% CI) -4.04 [-6.27, -1.81] 1.5 Unsupported subscale 24 5214 Mean Difference (IV, Random, 95% CI) -6.27 [-8.86, -3.68] 1.6 Ineffective choice subscale 24 Mean Difference (IV, Random, 95% CI) -6.31 [-8.93, -3.70] 5241 Mean Difference (IV, Random, 95% CI) Subtotals only 2 Decisional conflict - in 6 consultation 2.1 Uncertainty subscale 2 310 Mean Difference (IV, Random, 95% CI) -6.45 [-18.29, 5.38] 4 545 Mean Difference (IV, Random, 95% CI) -6.37 [-14.58, 1.85] 2.2 Uninformed subscale 2.3 Unclear values subscale 1 Mean Difference (IV, Random, 95% CI) -17.2 [-23.77, -10. 204 63] 2.4 Unsupported subscale 354 Mean Difference (IV, Random, 95% CI) -7.16 [-13.28, -1.03] 2 2 307 Mean Difference (IV, Random, 95% CI) -2.37 [-7.31, 2.58] 2.5 Ineffective choice subscale 2.6 Total decisional conflict -6.46 [-12.78, -0.14] 5 735 Mean Difference (IV, Random, 95% CI) score 3 Decisional conflict - in 36 Mean Difference (IV, Random, 95% CI) Subtotals only preparation for consultation 3.1 Uncertainty subscale 5890 Mean Difference (IV, Random, 95% CI) -3.83 [-6.12, -1.55] 26 23 5162 Mean Difference (IV, Random, 95% CI) -9.81 [-13.00, -6.61] 3.2 Uninformed subscale 3.3 Unclear values subscale 22 4864 Mean Difference (IV, Random, 95% CI) -8.40 [-11.59, -5.21] 3.4 Unsupported subscale 4860 Mean Difference (IV, Random, 95% CI) -6.18 [-8.96, -3.40] 22 3.5 Ineffective choice subscale 22 4934 Mean Difference (IV, Random, 95% CI) -6.75 [-9.59, -3.90] 207 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

211 3.6 Total decisional conflict Mean Difference (IV, Random, 95% CI) -7.32 [-9.35, -5.28] 33 8050 score 39 Subtotals only Mean Difference (IV, Random, 95% CI) 4 Decisional conflict - without studies having high risk of bias 4.1 Uncertainty subscale 26 5809 Mean Difference (IV, Random, 95% CI) -4.53 [-6.87, -2.18] 5316 Mean Difference (IV, Random, 95% CI) 4.2 Uninformed subscale 25 -9.96 [-13.13, -6.78] 21 Mean Difference (IV, Random, 95% CI) -9.55 [-13.08, -6.02] 4.3 Unclear values subscale 4677 4823 Mean Difference (IV, Random, 95% CI) -7.00 [-9.76, -4.24] 4.4 Unsupported subscale 22 22 4850 Mean Difference (IV, Random, 95% CI) -6.97 [-9.76, -4.18] 4.5 Ineffective choice subscale 35 Mean Difference (IV, Random, 95% CI) -7.81 [-9.84, -5.77] 4.6 Total decisional conflict 8240 score Comparison 5. Participation in decision making No. of No. of Statistical method Effect size Outcome or subgroup title participants studies 16 1 Participation in decision making Risk Ratio (M-H, Random, 95% CI) Subtotals only - all studies 1.1 Clinician-controlled 16 3180 Risk Ratio (M-H, Random, 95% CI) 0.68 [0.55, 0.83] decision making 1.28 [1.05, 1.55] 15 3009 Risk Ratio (M-H, Random, 95% CI) 1.2 Patient-controlled decision making 1.3 Shared decision making 15 Risk Ratio (M-H, Random, 95% CI) 0.95 [0.83, 1.10] 2973 Risk Ratio (M-H, Random, 95% CI) Subtotals only 2 Participation in decision making 3 - in consultation 0.89 [0.70, 1.12] Risk Ratio (M-H, Random, 95% CI) 3 650 2.1 Clinician-controlled decision making - in consultation 2.2 Patient-controlled decision 2 479 Risk Ratio (M-H, Random, 95% CI) 1.01 [0.80, 1.27] making - in consultation 2.3 Shared decision making - 1.14 [0.84, 1.55] 2 479 Risk Ratio (M-H, Random, 95% CI) in consultation 13 Risk Ratio (M-H, Random, 95% CI) Subtotals only 3 Participation in decision making - in preparation for consultation 2530 13 3.1 Clinician-controlled Risk Ratio (M-H, Random, 95% CI) 0.60 [0.48, 0.75] decision making 3.2 Patient-controlled 13 2530 Risk Ratio (M-H, Random, 95% CI) 1.37 [1.08, 1.73] decision making 3.3 Shared decision making 13 2494 Risk Ratio (M-H, Random, 95% CI) 0.94 [0.80, 1.09] 208 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

212 Comparison 6. Proportion undecided No. of No. of Statistical method Effect size Outcome or subgroup title studies participants 1 Proportion undecided - all 0.64 [0.52, 0.79] 5256 22 Risk Ratio (M-H, Random, 95% CI) studies Comparison 7. Satisfaction No. of No. of Statistical method Effect size Outcome or subgroup title participants studies 1 Satisfaction with the choice - all 11 Mean Difference (IV, Random, 95% CI) Totals not selected studies Mean Difference (IV, Random, 95% CI) 1 2 Satisfaction with the choice - in Totals not selected consultation 3 Satisfaction with the choice - in Mean Difference (IV, Random, 95% CI) Totals not selected 10 preparation for consultation 9 Mean Difference (IV, Random, 95% CI) Totals not selected 4 Satisfaction with the decision making process - all studies Mean Difference (IV, Random, 95% CI) Totals not selected 1 5 Satisfaction with the decision making process - in consultation Totals not selected 6 Satisfaction with the decision 8 Mean Difference (IV, Random, 95% CI) making process - in preparation for consultation Comparison 8. Choice No. of No. of Effect size Statistical method Outcome or subgroup title participants studies Risk Ratio (M-H, Random, 95% CI) Subtotals only 1 Choice: surgery over conservative 18 option 1.1 Per-protocol analysis 18 3286 Risk Ratio (M-H, Random, 95% CI) 0.87 [0.75, 1.01] 18 1.2 Intention-to-treat analysis Risk Ratio (M-H, Random, 95% CI) 0.86 [0.75, 1.00] 3844 1.3 Per-protocol analysis 17 3108 Risk Ratio (M-H, Random, 95% CI) 0.84 [0.73, 0.97] without prophylactic mastectomy 2 Choice for screening 25 Risk Ratio (M-H, Random, 95% CI) Subtotals only 2.1 PSA screening 10 Risk Ratio (M-H, Random, 95% CI) 0.88 [0.80, 0.98] 3996 2.2 Colorectal cancer 10 4529 Risk Ratio (M-H, Random, 95% CI) 1.12 [0.95, 1.31] screening 2.3 Breast cancer genetic 0.99 [0.71, 1.38] 3 738 Risk Ratio (M-H, Random, 95% CI) testing 209 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

213 2.4 Prenatal diagnostic testing 2 Risk Ratio (M-H, Random, 95% CI) 0.99 [0.91, 1.09] 1100 3 Choice: diabetes medication Risk Ratio (M-H, Random, 95% CI) 1.65 [1.06, 2.56] 4 447 (uptake new medication) Analysis 1.1. Comparison 1 Knowledge, Outcome 1 Knowledge - all studies. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 1 Knowledge Outcome: 1 Knowledge - all studies Mean Mean Difference Usual Care Difference Decision Aid Weight Study or subgroup N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 291 66 (35.48) 334 60 (29.24) 2.0 % 6.00 [ 0.86, 11.14 ] Allen 2010 7.00 [ 2.33, 11.67 ] 2.1 % 77 72 (12) 75 Arterburn 2011 65 (17) 104 75 (45) 123 54 (45) Barry 1997 21.00 [ 9.25, 32.75 ] 1.3 % 2.0 % 2.50 [ -3.31, 8.31 ] 50 74 (14.5) 56 71.5 (16) Bekker 2004 61 83 (16) 48 58 (16) Bernstein 1998 25.00 [ 18.95, 31.05 ] 1.9 % 2.2 % 6.00 [ 2.31, 9.69 ] 204 71 (20) 182 77 (17) Bjorklund 2012 61 75.7 (19) 61 49.9 (16) Chabrera 2015 25.80 [ 19.57, 32.03 ] 1.9 % 2.1 % 9.00 [ 4.69, 13.31 ] 155 81.4 (18.7) 151 72.4 (19.7) Frosch 2008a Gattellari 2003 106 50 (18.4) 108 45 (15.9) 2.1 % 5.00 [ 0.39, 9.61 ] 15.00 [ 10.40, 19.60 ] 2.1 % 131 57.2 (21.3) Gattellari 2005 136 42.2 (16.7) 95 (7) 65 (21) 29 Green 2001 14 1.3 % 30.00 [ 18.71, 41.29 ] 2.0 % 8.90 [ 3.60, 14.20 ] 129 79.5 (21.64) Hanson 2011 127 88.4 (21.64) 101 51.43 (18.2) 103 42.86 (18.3) Hess 2012 8.57 [ 3.56, 13.58 ] 2.0 % 1.4 % 17.63 [ 7.33, 27.93 ] 44 61.22 (20.38) Jibaja-Weiss 2011 39 43.59 (26.61) 32 92.6 (11) 35 85.2 (15.6) Johnson 2006 7.40 [ 0.98, 13.82 ] 1.9 % 2.0 % 4.62 [ -0.45, 9.69 ] Knops 2014 80 76.92 (16.92) 84 72.3 (16.15) Krist 2007 196 69 (33.21) 75 54 (33.21) 1.6 % 15.00 [ 6.16, 23.84 ] 1.6 % 33.00 [ 24.27, 41.73 ] 50 31 27 (16.7) Kupke 2013 60 (23.3) 357 62.7 (21.3) 353 57.3 (21.3) Kuppermann 2014 5.40 [ 2.27, 8.53 ] 2.2 % 2.0 % 2.00 [ -3.49, 7.49 ] Lam 2013 113 61 (21) 112 59 (21) Laupacis 2006 53 83 (19.5) 53 67.4 (17) 1.8 % 15.60 [ 8.64, 22.56 ] -20 -10 0 10 20 Favours Usual Care Favours Decision Aid ( Continued . . . ) 210 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

214 ( . . . ) Continued Mean Mean Decision Aid Weight Study or subgroup Difference Usual Care Difference N Mean(SD) IV,Random,95% CI Mean(SD) N IV,Random,95% CI 60 (26.86) 12.50 [ 5.05, 19.95 ] 100 Leighl 2011 100 72.5 (26.86) 1.8 % 215 61.6 (0.13) Lepore 2012 216 54.7 (0.13) 2.4 % 6.90 [ 6.88, 6.92 ] 19.90 [ 15.17, 24.63 ] 2.1 % 49 (21.7) 122 68.9 (19) Lerman 1997 164 107 46.7 (34.01) 93 45.1 (34.01) Lewis 2010 -1.60 [ -11.05, 7.85 ] 1.5 % 2.2 % 9.45 [ 5.68, 13.22 ] 137 75.91 (15.72) 136 66.46 (16.07) Man-Son-Hing 1999 134 41.29 (21) 273 64.14 (21.86) Mann E 2010 2.1 % 22.85 [ 18.45, 27.25 ] 1.9 % 10.80 [ 4.37, 17.23 ] 113 73.5 (27.6) 189 62.7 (27.6) Mathieu 2010 77 McCaffery 2010 72 (23.51) 81 (23.51) 71 9.00 [ 1.42, 16.58 ] 1.7 % 1.9 % 15.00 [ 8.39, 21.61 ] 60 (18) Montgomery 2003 75 (17) 50 58 69.7 (18) 202 57.5 (18.5) 196 Montgomery 2007 2.2 % 12.20 [ 8.61, 15.79 ] 20.00 [ 8.09, 31.91 ] 1.3 % 46 43.3 (29.61) Montori 2011 49 63.3 (29.61) 86 75 (32.04) 94 62 (32.04) Morgan 2000 13.00 [ 3.63, 22.37 ] 1.5 % 10.50 [ 1.44, 19.56 ] 1.6 % 53 (18.2) 48 63.5 (24.4) Mullan 2009 37 98 88 (19) 90 79 (18) Nassar 2007 9.00 [ 3.71, 14.29 ] 2.0 % 1.8 % 10.90 [ 3.73, 18.07 ] Protheroe 2007 54 59.7 (18.4) 54 48.8 (19.6) Sawka 2012 37 97 (6) 37 78 (13) 2.1 % 19.00 [ 14.39, 23.61 ] 17.50 [ 13.99, 21.01 ] 2.2 % 223 89.17 (15) 231 71.67 (22.5) Schroy 2011 74 60 (30) 76 Schwalm 2012 40 (26) 1.6 % 20.00 [ 11.02, 28.98 ] 2.2 % 8.57 [ 5.55, 11.59 ] 191 65.71 (14.29) 190 57.14 (15.71) Schwartz 2001 75.33 (15) 92 60.53 (17.07) 99 Shorten 2005 2.1 % 14.80 [ 10.23, 19.37 ] 2.2 % 20.00 [ 16.42, 23.58 ] 357 54.17 (27.83) Smith 2010 173 34.17 (14.25) Stacey 2014a 66 46.6 (21.4) 66 71.2 (23.7) 24.60 [ 16.90, 32.30 ] 1.7 % 2.3 % 22.50 [ 20.23, 24.77 ] 785 53.75 (28.75) 792 31.25 (15) Steckelberg 2011 80 77.3 (15.5) Taylor 2006 74 62.7 (11.8) 2.1 % 14.60 [ 10.27, 18.93 ] 2.0 % 0.56 [ -4.77, 5.89 ] Thomson 2007 53 62.91 (14.26) 56 62.35 (14.1) Van Peperstraten 2010 123 62 (28.3) 132 43 (20.5) 1.9 % 19.00 [ 12.90, 25.10 ] 1.7 % 25.00 [ 16.83, 33.17 ] 70 79 49 (23.33) Vandemheen 2009 74 (27.07) 78 48 (21.6) 80 31 (18.8) Volk 1999 17.00 [ 10.68, 23.32 ] 1.9 % 2.1 % 8.50 [ 4.37, 12.63 ] Whelan 2003 82 80.2 (14.4) 93 71.7 (13.3) Williams 2013 196 64.4 (18.5) 185 61.7 (17.8) 2.2 % 2.70 [ -0.95, 6.35 ] 2.0 % 25.00 [ 19.60, 30.40 ] Wong 2006 85 (26.7) 159 60 (21.7) 154 -20 -10 0 10 20 Favours Usual Care Favours Decision Aid ( Continued . . . ) 211 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

215 ( . . . Continued ) Mean Mean Usual Care Study or subgroup Difference Weight Decision Aid Difference Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) N N 6537 100.0 % 13.27 [ 11.32, 15.23 ] Total (95% CI) 6779 2 2 2 = 41.98; Chi =93% = 717.60, df = 51 (P<0.00001); I Heterogeneity: Tau Test for overall effect: Z = 13.31 (P < 0.00001) Test for subgroup differences: Not applicable -10 0 10 20 -20 Favours Decision Aid Favours Usual Care subgroup by timing of intervention (in Analysis 1.2. Comparison 1 Knowledge, Outcome 2 Knowledge - consultation versus in preparation for consultation). creening decisions Review: Decision aids for people facing health treatment or s Comparison: 1 Knowledge Outcome: 2 Knowledge - subgroup by timing of intervention (i n consultation versus in preparation for consultation) Mean Mean Difference Weight Decision Aid Difference Study or subgroup Usual Care Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI N 1 In consultation Bekker 2004 50 74 (14.5) 56 71.5 (16) 13.3 % 2.50 [ -3.31, 8.31 ] 13.8 % 8.57 [ 3.56, 13.58 ] 103 42.86 (18.3) Hess 2012 101 51.43 (18.2) Johnson 2006 32 35 85.2 (15.6) 92.6 (11) 7.40 [ 0.98, 13.82 ] 13.0 % 50 60 (23.3) 31 27 (16.7) Kupke 2013 11.4 % 33.00 [ 24.27, 41.73 ] 9.3 % 20.00 [ 8.09, 31.91 ] 49 63.3 (29.61) 46 43.3 (29.61) Montori 2011 48 63.5 (24.4) 53 (18.2) Mullan 2009 37 10.50 [ 1.44, 19.56 ] 11.2 % 13.6 % 0.56 [ -4.77, 5.89 ] 53 62.91 (14.26) 56 62.35 (14.1) Thomson 2007 Whelan 2003 82 80.2 (14.4) 93 71.7 (13.3) 14.3 % 8.50 [ 4.37, 12.63 ] 100.0 % 10.57 [ 4.79, 16.36 ] 465 Subtotal (95% CI) 457 2 2 2 = 46.70, df = 7 (P<0.00001); I Heterogeneity: Tau =85% = 56.40; Chi Test for overall effect: Z = 3.58 (P = 0.00034) 2 In preparation for consultation -20 -10 0 10 20 Favours Usual Care Favours Decision Aid ( Continued . . . ) 212 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

216 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 334 6.00 [ 0.86, 11.14 ] 2.4 % 66 (35.48) Allen 2010 291 60 (29.24) 77 65 (17) 75 Arterburn 2011 72 (12) 2.4 % 7.00 [ 2.33, 11.67 ] 1.5 % 21.00 [ 9.25, 32.75 ] 104 75 (45) 54 (45) Barry 1997 123 83 (16) 58 (16) 61 Bernstein 1998 48 2.3 % 25.00 [ 18.95, 31.05 ] 2.5 % 6.00 [ 2.31, 9.69 ] 182 204 71 (20) Bjorklund 2012 77 (17) 61 75.7 (19) 61 49.9 (16) Chabrera 2015 25.80 [ 19.57, 32.03 ] 2.2 % 2.5 % 9.00 [ 4.69, 13.31 ] Frosch 2008a 151 72.4 (19.7) 155 81.4 (18.7) 108 50 (18.4) 106 Gattellari 2003 45 (15.9) 2.4 % 5.00 [ 0.39, 9.61 ] 15.00 [ 10.40, 19.60 ] 2.4 % 136 42.2 (16.7) 131 57.2 (21.3) Gattellari 2005 29 95 (7) 14 65 (21) Green 2001 1.6 % 30.00 [ 18.71, 41.29 ] 2.4 % 8.90 [ 3.60, 14.20 ] 127 88.4 (21.64) Hanson 2011 129 79.5 (21.64) Jibaja-Weiss 2011 39 43.59 (26.61) 44 61.22 (20.38) 17.63 [ 7.33, 27.93 ] 1.7 % 2.4 % 4.62 [ -0.45, 9.69 ] 80 76.92 (16.92) 84 72.3 (16.15) Knops 2014 196 69 (33.21) Krist 2007 54 (33.21) 75 1.9 % 15.00 [ 6.16, 23.84 ] 2.6 % 5.40 [ 2.27, 8.53 ] Kuppermann 2014 357 62.7 (21.3) 353 57.3 (21.3) Lam 2013 113 61 (21) 112 59 (21) 2.3 % 2.00 [ -3.49, 7.49 ] 2.1 % 15.60 [ 8.64, 22.56 ] 53 Laupacis 2006 83 (19.5) 53 67.4 (17) 100 100 72.5 (26.86) Leighl 2011 60 (26.86) 2.1 % 12.50 [ 5.05, 19.95 ] 2.7 % 6.90 [ 6.88, 6.92 ] 215 61.6 (0.13) 216 54.7 (0.13) Lepore 2012 68.9 (19) 164 49 (21.7) 122 Lerman 1997 2.4 % 19.90 [ 15.17, 24.63 ] 1.8 % -1.60 [ -11.05, 7.85 ] 93 45.1 (34.01) Lewis 2010 107 46.7 (34.01) Man-Son-Hing 1999 136 66.46 (16.07) 137 75.91 (15.72) 9.45 [ 5.68, 13.22 ] 2.5 % 2.5 % 22.85 [ 18.45, 27.25 ] 273 64.14 (21.86) 134 41.29 (21) Mann E 2010 113 73.5 (27.6) 189 62.7 (27.6) Mathieu 2010 10.80 [ 4.37, 17.23 ] 2.2 % 2.1 % 9.00 [ 1.42, 16.58 ] McCaffery 2010 77 81 (23.51) 71 72 (23.51) Montgomery 2003 50 75 (17) 58 60 (18) 2.2 % 15.00 [ 8.39, 21.61 ] 2.6 % 12.20 [ 8.61, 15.79 ] 196 69.7 (18) Montgomery 2007 202 57.5 (18.5) 75 (32.04) 94 62 (32.04) Morgan 2000 86 1.8 % 13.00 [ 3.63, 22.37 ] 2.4 % 9.00 [ 3.71, 14.29 ] 90 79 (18) 98 88 (19) Nassar 2007 Protheroe 2007 54 59.7 (18.4) 54 48.8 (19.6) 2.1 % 10.90 [ 3.73, 18.07 ] 2.4 % 19.00 [ 14.39, 23.61 ] Sawka 2012 97 (6) 37 78 (13) 37 -20 -10 0 10 20 Favours Usual Care Favours Decision Aid ( Continued . . . ) 213 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

217 ( . . . ) Continued Mean Mean Decision Aid Study or subgroup Difference Usual Care Weight Difference Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) N N 89.17 (15) 231 71.67 (22.5) 2.6 % 17.50 [ 13.99, 21.01 ] Schroy 2011 223 60 (30) 74 40 (26) 76 Schwalm 2012 1.9 % 20.00 [ 11.02, 28.98 ] 2.6 % 8.57 [ 5.55, 11.59 ] 191 65.71 (14.29) Schwartz 2001 190 57.14 (15.71) Shorten 2005 75.33 (15) 92 60.53 (17.07) 99 14.80 [ 10.23, 19.37 ] 2.4 % 2.6 % 20.00 [ 16.42, 23.58 ] 357 54.17 (27.83) 173 34.17 (14.25) Smith 2010 66 71.2 (23.7) 66 46.6 (21.4) Stacey 2014a 2.0 % 24.60 [ 16.90, 32.30 ] 22.50 [ 20.23, 24.77 ] 2.7 % 785 53.75 (28.75) 792 31.25 (15) Steckelberg 2011 Taylor 2006 80 77.3 (15.5) 74 62.7 (11.8) 2.5 % 14.60 [ 10.27, 18.93 ] 2.3 % 19.00 [ 12.90, 25.10 ] Van Peperstraten 2010 43 (20.5) 123 132 62 (28.3) 70 79 49 (23.33) Vandemheen 2009 74 (27.07) 2.0 % 25.00 [ 16.83, 33.17 ] 2.2 % 17.00 [ 10.68, 23.32 ] Volk 1999 31 (18.8) 78 48 (21.6) 80 Williams 2013 196 64.4 (18.5) 185 61.7 (17.8) 2.5 % 2.70 [ -0.95, 6.35 ] 25.00 [ 19.60, 30.40 ] 2.3 % 85 (26.7) 154 Wong 2006 159 60 (21.7) Subtotal (95% CI) 6314 6080 100.0 % 13.77 [ 11.61, 15.93 ] 2 2 2 = 669.38, df = 43 (P<0.00001); I = 44.14; Chi =94% Heterogeneity: Tau Test for overall effect: Z = 12.50 (P < 0.00001) 2 2 = 1.03, df = 1 (P = 0.31), I Test for subgroup differences: Chi =3% -20 0 10 20 -10 Favours Usual Care Favours Decision Aid 214 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

218 Analysis 1.3. Comparison 1 Knowledge, Outcome 3 Knowledge - studies without high risk of bias. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 1 Knowledge Outcome: 3 Knowledge - studies without high risk of bias Mean Mean Decision Aid Study or subgroup Difference Usual Care Difference Weight Mean(SD) IV,Random,95% CI IV,Random,95% CI N N Mean(SD) 66 (35.48) 334 60 (29.24) 2.2 % 6.00 [ 0.86, 11.14 ] Allen 2010 291 72 (12) Arterburn 2011 65 (17) 75 77 7.00 [ 2.33, 11.67 ] 2.3 % 1.4 % 21.00 [ 9.25, 32.75 ] Barry 1997 75 (45) 123 104 54 (45) 74 (14.5) 56 71.5 (16) 50 Bekker 2004 2.50 [ -3.31, 8.31 ] 2.2 % 2.1 % 25.00 [ 18.95, 31.05 ] 61 58 (16) Bernstein 1998 48 83 (16) 182 77 (17) 71 (20) Bjorklund 2012 204 6.00 [ 2.31, 9.69 ] 2.4 % 2.1 % 25.80 [ 19.57, 32.03 ] Chabrera 2015 61 61 49.9 (16) 75.7 (19) 155 81.4 (18.7) 151 72.4 (19.7) Frosch 2008a 9.00 [ 4.69, 13.31 ] 2.3 % 2.3 % 5.00 [ 0.39, 9.61 ] 108 45 (15.9) 106 50 (18.4) Gattellari 2003 131 57.2 (21.3) 136 42.2 (16.7) Gattellari 2005 15.00 [ 10.40, 19.60 ] 2.3 % 1.5 % 30.00 [ 18.71, 41.29 ] Green 2001 29 14 65 (21) 95 (7) 8.90 [ 3.60, 14.20 ] 2.2 % 129 79.5 (21.64) 127 88.4 (21.64) Hanson 2011 101 51.43 (18.2) 103 42.86 (18.3) Hess 2012 2.3 % 8.57 [ 3.56, 13.58 ] 1.6 % 17.63 [ 7.33, 27.93 ] Jibaja-Weiss 2011 39 43.59 (26.61) 44 61.22 (20.38) 92.6 (11) 35 85.2 (15.6) Johnson 2006 32 7.40 [ 0.98, 13.82 ] 2.1 % 2.4 % 5.40 [ 2.27, 8.53 ] Kuppermann 2014 357 62.7 (21.3) 353 57.3 (21.3) 113 Lam 2013 112 59 (21) 61 (21) 2.2 % 2.00 [ -3.49, 7.49 ] 2.0 % 15.60 [ 8.64, 22.56 ] 53 53 67.4 (17) Laupacis 2006 83 (19.5) 100 72.5 (26.86) Leighl 2011 60 (26.86) 100 1.9 % 12.50 [ 5.05, 19.95 ] Lepore 2012 215 61.6 (0.13) 216 54.7 (0.13) 2.6 % 6.90 [ 6.88, 6.92 ] 2.3 % 19.90 [ 15.17, 24.63 ] Lerman 1997 49 (21.7) 122 164 68.9 (19) 273 64.14 (21.86) 41.29 (21) Mann E 2010 134 22.85 [ 18.45, 27.25 ] 2.3 % 10.80 [ 4.37, 17.23 ] 2.1 % 113 73.5 (27.6) 189 62.7 (27.6) Mathieu 2010 McCaffery 2010 77 81 (23.51) 71 72 (23.51) 1.9 % 9.00 [ 1.42, 16.58 ] 2.0 % 15.00 [ 8.39, 21.61 ] Montgomery 2003 75 (17) 58 60 (18) 50 -20 -10 0 10 20 Favours Usual Care Favours Decision Aid ( Continued . . . ) 215 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

219 ( . . . ) Continued Mean Mean Decision Aid Study or subgroup Difference Usual Care Weight Difference Mean(SD) IV,Random,95% CI N N Mean(SD) IV,Random,95% CI 69.7 (18) 2.4 % 12.20 [ 8.61, 15.79 ] 196 Montgomery 2007 202 57.5 (18.5) 46 43.3 (29.61) Montori 2011 49 63.3 (29.61) 1.4 % 20.00 [ 8.09, 31.91 ] 1.7 % 13.00 [ 3.63, 22.37 ] Morgan 2000 75 (32.04) 94 62 (32.04) 86 37 53 (18.2) Mullan 2009 48 63.5 (24.4) 10.50 [ 1.44, 19.56 ] 1.7 % 2.2 % 9.00 [ 3.71, 14.29 ] Nassar 2007 79 (18) 98 88 (19) 90 Protheroe 2007 54 48.8 (19.6) 54 59.7 (18.4) 10.90 [ 3.73, 18.07 ] 2.0 % 2.3 % 19.00 [ 14.39, 23.61 ] 97 (6) 37 78 (13) 37 Sawka 2012 89.17 (15) 231 71.67 (22.5) Schroy 2011 223 17.50 [ 13.99, 21.01 ] 2.4 % 1.7 % 20.00 [ 11.02, 28.98 ] 74 40 (26) 76 Schwalm 2012 60 (30) Schwartz 2001 190 57.14 (15.71) 191 65.71 (14.29) 8.57 [ 5.55, 11.59 ] 2.5 % 2.3 % 14.80 [ 10.23, 19.37 ] Shorten 2005 75.33 (15) 92 60.53 (17.07) 99 Smith 2010 357 54.17 (27.83) 173 34.17 (14.25) 2.4 % 20.00 [ 16.42, 23.58 ] 24.60 [ 16.90, 32.30 ] 1.9 % 66 71.2 (23.7) Stacey 2014a 66 46.6 (21.4) 792 785 53.75 (28.75) Steckelberg 2011 31.25 (15) 2.5 % 22.50 [ 20.23, 24.77 ] 2.3 % 14.60 [ 10.27, 18.93 ] 74 62.7 (11.8) Taylor 2006 80 77.3 (15.5) 53 62.91 (14.26) Thomson 2007 56 62.35 (14.1) 2.2 % 0.56 [ -4.77, 5.89 ] 19.00 [ 12.90, 25.10 ] 2.1 % 132 43 (20.5) 123 Van Peperstraten 2010 62 (28.3) Vandemheen 2009 74 (27.07) 79 49 (23.33) 70 1.8 % 25.00 [ 16.83, 33.17 ] 17.00 [ 10.68, 23.32 ] 2.1 % 48 (21.6) 80 Volk 1999 31 (18.8) 78 93 71.7 (13.3) 82 80.2 (14.4) Whelan 2003 8.50 [ 4.37, 12.63 ] 2.4 % 2.4 % 2.70 [ -0.95, 6.35 ] Williams 2013 196 64.4 (18.5) 185 61.7 (17.8) 154 85 (26.7) 159 60 (21.7) Wong 2006 2.2 % 25.00 [ 19.60, 30.40 ] 100.0 % 13.43 [ 11.37, 15.49 ] 6223 6104 Total (95% CI) 2 2 2 Heterogeneity: Tau = 42.59; Chi = 674.45, df = 46 (P<0.00001); I =93% Test for overall effect: Z = 12.78 (P < 0.00001) Test for subgroup differences: Not applicable -20 -10 0 10 20 Favours Usual Care Favours Decision Aid 216 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

220 Analysis 2.1. Comparison 2 Accurate risk perceptions, Outc ome 1 Accurate risk perceptions - all studies. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 2 Accurate risk perceptions Outcome: 1 Accurate risk perceptions - all studies Decision Aid Risk Ratio Weight Ris k Ratio Study or subgroup Control M- M- H,Random,95% H,Random,95% n/N n/N CI CI 11/108 5.2 % 57/106 Gattellari 2003 5.28 [ 2.93, 9.50 ] Hess 2012 24/101 1/103 1.2 % 24.48 [ 3.37, 177.53 ] 14/47 5/50 Laupacis 2006 2.98 [ 1.16, 7.63 ] 3.5 % 5.6 % 2.70 [ 1.59, 4.58 ] 23/32 12/45 LeBlanc 2015 Lerman 1997 90/122 108/164 7.5 % 1.12 [ 0.96, 1.31 ] 2.80 [ 2.05, 3.83 ] 6.8 % 35/148 Man-Son-Hing 1999 92/139 35/80 22/70 Mann D 2010 1.39 [ 0.91, 2.13 ] 6.2 % 3.4 % 13.22 [ 5.05, 34.62 ] Mathers 2012 4/75 67/95 McAlister 2005 66/175 25/155 6.3 % 2.34 [ 1.56, 3.51 ] 7.2 % 1.37 [ 1.09, 1.73 ] McBride 2002 109/265 82/274 23/49 Montori 2011 10/43 5.0 % 2.02 [ 1.09, 3.75 ] 6.7 % 1.58 [ 1.13, 2.20 ] Schwalm 2012 47/76 29/74 7.5 % 2.58 [ 2.18, 3.05 ] 141/792 361/785 Steckelberg 2011 46/70 23/79 Vandemheen 2009 6.4 % 2.26 [ 1.54, 3.31 ] 6.7 % 1.55 [ 1.12, 2.15 ] Whelan 2003 47/82 34/92 Whelan 2004 73/94 62/107 7.4 % 1.34 [ 1.10, 1.63 ] 7.4 % 1.31 [ 1.10, 1.56 ] 189/266 72/133 Wolf 2000 2584 2512 Total (95% CI) 2.10 [ 1.66, 2.66 ] 100.0 % Total events: 1363 (Decision Aid), 676 (Control) 2 2 2 = 0.19; Chi = 151.38, df = 16 (P<0.00001); I Heterogeneity: Tau =89% Test for overall effect: Z = 6.16 (P < 0.00001) Test for subgroup differences: Not applicable 0.1 0.2 0.5 1 2 5 10 Favours Control Favours Decision Aid 217 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

221 Analysis 2.2. Comparison 2 Accurate risk perceptions, Outc ome 2 Accurate risk perceptions - subgroup by ion for consultation). timing of intervention (in consultation versus in preparat creening decisions Review: Decision aids for people facing health treatment or s Comparison: 2 Accurate risk perceptions nsultation) intervention (in consultation versus in preparation for co Outcome: 2 Accurate risk perceptions - subgroup by timing of Risk Ratio Weight Study or subgroup k Ratio Decision Aid Control Ris M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 In consultation Hess 2012 1/103 24/101 2.6 % 24.48 [ 3.37, 177.53 ] 16.5 % 2.70 [ 1.59, 4.58 ] LeBlanc 2015 23/32 12/45 35/80 22/70 Mann D 2010 1.39 [ 0.91, 2.13 ] 19.2 % 2.02 [ 1.09, 3.75 ] 14.4 % 23/49 10/43 Montori 2011 34/92 Whelan 2003 47/82 1.55 [ 1.12, 2.15 ] 22.0 % 1.34 [ 1.10, 1.63 ] 25.3 % 62/107 Whelan 2004 73/94 Subtotal (95% CI) 438 460 1.79 [ 1.28, 2.52 ] 100.0 % Total events: 225 (Decision Aid), 141 (Control) 2 2 2 Heterogeneity: Tau = 0.11; Chi = 17.62, df = 5 (P = 0.003); I =72% Test for overall effect: Z = 3.37 (P = 0.00075) 2 In preparation for consultation 7.9 % 5.28 [ 2.93, 9.50 ] 57/106 11/108 Gattellari 2003 5.5 % 2.98 [ 1.16, 7.63 ] 5/50 Laupacis 2006 14/47 Lerman 1997 90/122 108/164 1.12 [ 0.96, 1.31 ] 10.8 % 10.0 % 2.80 [ 2.05, 3.83 ] 35/148 Man-Son-Hing 1999 92/139 67/95 Mathers 2012 4/75 5.4 % 13.22 [ 5.05, 34.62 ] McAlister 2005 66/175 25/155 2.34 [ 1.56, 3.51 ] 9.3 % 10.4 % 1.37 [ 1.09, 1.73 ] 109/265 McBride 2002 82/274 47/76 29/74 Schwalm 2012 1.58 [ 1.13, 2.20 ] 9.8 % 10.7 % 2.58 [ 2.18, 3.05 ] Steckelberg 2011 361/785 141/792 Vandemheen 2009 46/70 23/79 9.5 % 2.26 [ 1.54, 3.31 ] 10.7 % 1.31 [ 1.10, 1.56 ] 189/266 72/133 Wolf 2000 2146 2052 Subtotal (95% CI) 2.25 [ 1.65, 3.07 ] 100.0 % Total events: 1138 (Decision Aid), 535 (Control) 2 2 2 Heterogeneity: Tau = 130.92, df = 10 (P<0.00001); I = 0.23; Chi =92% Test for overall effect: Z = 5.11 (P < 0.00001) 2 2 = 0.94, df = 1 (P = 0.33), I =0.0% Test for subgroup differences: Chi 0.1 0.2 0.5 1 2 5 10 Favours Control Favours Decision Aid 218 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

222 Analysis 2.3. Comparison 2 Accurate risk perceptions, Outc ome 3 Accurate risk perceptions - studies without high risk of bias. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 2 Accurate risk perceptions risk of bias Outcome: 3 Accurate risk perceptions - studies without high Decision Aid Control Risk Ratio Weight Ris k Ratio Study or subgroup M- M- H,Random,95% H,Random,95% n/N n/N CI CI Gattellari 2003 11/108 6.0 % 5.28 [ 2.93, 9.50 ] 57/106 24/101 1/103 Hess 2012 24.48 [ 3.37, 177.53 ] 1.3 % 3.9 % 2.98 [ 1.16, 7.63 ] 14/47 5/50 Laupacis 2006 Lerman 1997 90/122 108/164 8.6 % 1.12 [ 0.96, 1.31 ] 7.1 % 1.39 [ 0.91, 2.13 ] 35/80 22/70 Mann D 2010 67/95 4/75 Mathers 2012 3.8 % 13.22 [ 5.05, 34.62 ] 7.2 % 2.34 [ 1.56, 3.51 ] McAlister 2005 66/175 25/155 1.37 [ 1.09, 1.73 ] 8.2 % 82/274 109/265 McBride 2002 23/49 10/43 Montori 2011 5.7 % 2.02 [ 1.09, 3.75 ] 1.58 [ 1.13, 2.20 ] 7.7 % Schwalm 2012 47/76 29/74 Steckelberg 2011 361/785 141/792 2.58 [ 2.18, 3.05 ] 8.5 % 7.3 % 2.26 [ 1.54, 3.31 ] 46/70 Vandemheen 2009 23/79 47/82 Whelan 2003 34/92 7.7 % 1.55 [ 1.12, 2.15 ] Whelan 2004 73/94 62/107 8.4 % 1.34 [ 1.10, 1.63 ] 8.5 % 1.31 [ 1.10, 1.56 ] Wolf 2000 189/266 72/133 Total (95% CI) 2319 2413 100.0 % 2.02 [ 1.57, 2.59 ] Total events: 1248 (Decision Aid), 629 (Control) 2 2 2 = 0.18; Chi Heterogeneity: Tau =90% = 136.15, df = 14 (P<0.00001); I Test for overall effect: Z = 5.51 (P < 0.00001) Test for subgroup differences: Not applicable 0.1 0.2 0.5 1 2 5 10 Favours Control Favours Decision Aid 219 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

223 Analysis 3.1. Comparison 3 Informed values-choice congrue nce, Outcome 1 Informed values-choice congruence - all studies. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 3 Informed values-choice congruence Outcome: 1 Informed values-choice congruence - all studies Study or subgroup Comparison Risk Ratio Weight Risk Ratio Decision Aid M- M- H,Random,95% H,Random,95% n/N n/N CI CI 128/179 123/197 11.9 % Bjorklund 2012 1.15 [ 0.99, 1.32 ] Fagerlin 2011 6/102 202/383 8.97 [ 4.10, 19.60 ] 7.5 % 11.9 % 1.51 [ 1.31, 1.73 ] 227/309 136/279 Mathieu 2007 65/91 Mathieu 2010 70/110 11.7 % 1.12 [ 0.93, 1.36 ] 11.9 % 1.17 [ 1.02, 1.35 ] Nagle 2008 127/167 111/171 Schwalm 2012 36/76 19/74 10.0 % 1.84 [ 1.17, 2.91 ] 2.78 [ 1.81, 4.25 ] 10.2 % 21/172 121/357 Smith 2010 14/56 Stacey 2014a 31/55 2.25 [ 1.35, 3.75 ] 9.6 % 11.6 % 3.45 [ 2.83, 4.20 ] Steckelberg 2011 345/785 101/792 14/134 2/137 Trevena 2008 3.8 % 7.16 [ 1.66, 30.89 ] 100.0 % 2.06 [ 1.46, 2.91 ] 2090 Total (95% CI) 2536 Total events: 1296 (Decision Aid), 603 (Comparison) 2 2 2 = 0.26; Chi = 186.48, df = 9 (P<0.00001); I =95% Heterogeneity: Tau Test for overall effect: Z = 4.11 (P = 0.000040) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Favours usual care Favours decision aid 220 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

224 Analysis 3.2. Comparison 3 Informed values-choice congrue nce, Outcome 2 Informed values-choice congruence - actual choice only. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 3 Informed values-choice congruence Outcome: 2 Informed values-choice congruence - actual choi ce only Comparison Risk Ratio Weight Risk Ratio Study or subgroup Decision Aid M- M- H,Random,95% H,Random,95% n/N n/N CI CI 128/179 123/197 13.9 % 1.15 [ 0.99, 1.32 ] Bjorklund 2012 202/383 Fagerlin 2011 6/102 9.0 % 8.97 [ 4.10, 19.60 ] 1.51 [ 1.31, 1.73 ] 14.0 % 136/279 227/309 Mathieu 2007 Nagle 2008 127/167 111/171 13.9 % 1.17 [ 1.02, 1.35 ] 11.9 % 1.84 [ 1.17, 2.91 ] 19/74 Schwalm 2012 36/76 Smith 2010 121/357 21/172 2.78 [ 1.81, 4.25 ] 12.1 % 11.4 % 2.25 [ 1.35, 3.75 ] 31/55 14/56 Stacey 2014a 345/785 Steckelberg 2011 101/792 13.7 % 3.45 [ 2.83, 4.20 ] 100.0 % 2.13 [ 1.44, 3.14 ] Total (95% CI) 2311 1843 Total events: 1217 (Decision Aid), 531 (Comparison) 2 2 2 Heterogeneity: Tau = 0.28; Chi = 167.61, df = 7 (P<0.00001); I =96% Test for overall effect: Z = 3.80 (P = 0.00014) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Favours usual care Favours decision aid 221 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

225 Analysis 3.3. Comparison 3 Informed values-choice congrue nce, Outcome 3 Informed values-chose congruence -using MMIC. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 3 Informed values-choice congruence Outcome: 3 Informed values-chose congruence -using MMIC Comparison Risk Ratio Weight Risk Ratio Study or subgroup Decision Aid M- M- H,Random,95% H,Random,95% n/N n/N CI CI 128/179 123/197 14.7 % 1.15 [ 0.99, 1.32 ] Bjorklund 2012 202/383 Fagerlin 2011 6/102 9.4 % 8.97 [ 4.10, 19.60 ] 1.51 [ 1.31, 1.73 ] 14.7 % 136/279 227/309 Mathieu 2007 Mathieu 2010 65/91 70/110 14.5 % 1.12 [ 0.93, 1.36 ] 14.7 % 1.17 [ 1.02, 1.35 ] 111/171 Nagle 2008 127/167 Smith 2010 121/357 21/172 2.78 [ 1.81, 4.25 ] 12.7 % 14.4 % 3.45 [ 2.83, 4.20 ] 345/785 101/792 Steckelberg 2011 14/134 Trevena 2008 2/137 4.9 % 7.16 [ 1.66, 30.89 ] 100.0 % 2.08 [ 1.40, 3.08 ] Total (95% CI) 2405 1960 Total events: 1229 (Decision Aid), 570 (Comparison) 2 2 2 Heterogeneity: Tau = 0.27; Chi = 184.27, df = 7 (P<0.00001); I =96% Test for overall effect: Z = 3.63 (P = 0.00028) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Favours decision aid Favours comparison 222 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

226 Analysis 3.4. Comparison 3 Informed values-choice congrue nce, Outcome 4 Informed values-chose congruence - heterogeneous measures. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 3 Informed values-choice congruence Outcome: 4 Informed values-chose congruence - heterogeneo us measures Comparison Risk Ratio Weight Risk Ratio Study or subgroup Decision Aid M- M- H,Random,95% H,Random,95% n/N n/N CI CI Schwalm 2012 36/76 19/74 55.7 % 1.84 [ 1.17, 2.91 ] Stacey 2014a 31/55 14/56 44.3 % 2.25 [ 1.35, 3.75 ] 100.0 % 2.02 [ 1.44, 2.83 ] 131 130 Total (95% CI) Total events: 67 (Decision Aid), 33 (Comparison) 2 2 2 Heterogeneity: Tau = 0.33, df = 1 (P = 0.56); I =0.0% = 0.0; Chi Test for overall effect: Z = 4.05 (P = 0.000051) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Favours Decision Aid Favours Comparison 223 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

227 Analysis 3.5. Comparison 3 Informed values-choice congrue nce, Outcome 5 Informed values-choice congruence - without studies of high risk of bias. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 3 Informed values-choice congruence Outcome: 5 Informed values-choice congruence - without stu dies of high risk of bias Comparison Risk Ratio Weight Risk Ratio Study or subgroup Decision Aid M- M- H,Random,95% H,Random,95% n/N n/N CI CI 128/179 123/197 11.9 % 1.15 [ 0.99, 1.32 ] Bjorklund 2012 8.97 [ 4.10, 19.60 ] 7.5 % 202/383 6/102 Fagerlin 2011 136/279 227/309 Mathieu 2007 1.51 [ 1.31, 1.73 ] 11.9 % 11.7 % 1.12 [ 0.93, 1.36 ] Mathieu 2010 65/91 70/110 127/167 111/171 Nagle 2008 11.9 % 1.17 [ 1.02, 1.35 ] 10.0 % 1.84 [ 1.17, 2.91 ] 19/74 Schwalm 2012 36/76 Smith 2010 121/357 21/172 2.78 [ 1.81, 4.25 ] 10.2 % 9.6 % 2.25 [ 1.35, 3.75 ] 31/55 14/56 Stacey 2014a 345/785 101/792 Steckelberg 2011 11.6 % 3.45 [ 2.83, 4.20 ] 3.8 % 7.16 [ 1.66, 30.89 ] Trevena 2008 14/134 2/137 Total (95% CI) 2536 2090 100.0 % 2.06 [ 1.46, 2.91 ] Total events: 1296 (Decision Aid), 603 (Comparison) 2 2 2 = 0.26; Chi = 186.48, df = 9 (P<0.00001); I =95% Heterogeneity: Tau Test for overall effect: Z = 4.11 (P = 0.000040) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Favours usual care Favours decision aid 224 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

228 Analysis 4.1. Comparison 4 Decisional conflict, Outcome 1 De cisional conflict - all studies. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 4 Decisional conflict Outcome: 1 Decisional conflict - all studies Mean Mean Decision Aid Usual Care Difference Study or subgroup Difference Weight Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) N N 1 Total decisional conflict score 14 (34.29) 334 20 (37.83) Allen 2010 291 -6.00 [ -11.66, -0.34 ] 2.6 % 15.8 (13.9) 51 14.1 (16.1) Brazell 2014 53 1.70 [ -4.09, 7.49 ] 2.6 % -20.50 [ -24.71, -16.29 ] 2.9 % 51.7 (13.3) Chabrera 2015 61 61 31.2 (10.2) 69 23.4 (14.95) 69 29.2 (16.61) De Achaval 2012 -5.80 [ -11.07, -0.53 ] 2.7 % 2.2 % -5.00 [ -12.64, 2.64 ] 41 Dolan 2002 37 25.75 (20.25) 20.75 (13) 89 38.08 (24.15) Evans 2010 103 49.58 (24.15) 2.3 % -11.50 [ -18.35, -4.65 ] -33.70 [ -50.79, -16.61 ] 0.9 % Fagerlin 2011 22 (42.2) 690 160 55.7 (108.4) 115 24.25 (18.55) 118 16.25 (18.55) Hanson 2011 2.8 % -8.00 [ -12.76, -3.24 ] -20.00 [ -25.46, -14.54 ] 2.6 % 103 43.3 (18.97) 101 23.3 (20.76) Hess 2012 44 16.53 (19.9) Jibaja-Weiss 2011 39 22.16 (25.29) 1.8 % -5.63 [ -15.51, 4.25 ] 2.7 % -2.00 [ -7.38, 3.38 ] 81 Knops 2014 73 22 (17) 24 (17) Kuppermann 2014 357 12.9 (14.1) 353 13.8 (15.6) 3.2 % -0.90 [ -3.09, 1.29 ] 2.9 % -4.10 [ -8.26, 0.06 ] Lam 2013 19.9 (16.3) 113 112 15.8 (15.5) 53 17.5 (13.75) 54 25.25 (14.25) Laupacis 2006 2.7 % -7.75 [ -13.06, -2.44 ] -4.00 [ -10.32, 2.32 ] 2.5 % Legare 2008a 23 (14.25) 43 41 27 (15.25) 215 34.15 (24.03) 216 39.85 (24.04) Lepore 2012 -5.70 [ -10.24, -1.16 ] 2.8 % 3.1 % -2.25 [ -5.12, 0.62 ] Man-Son-Hing 1999 139 16.25 (11.25) 18.5 (13.5) 148 80 25.5 (11.14) Mann D 2010 70 28.5 (11.14) -3.00 [ -6.57, 0.57 ] 3.0 % 2.9 % -7.80 [ -11.93, -3.67 ] 95 Mathers 2012 80 25.2 (14.9) 17.4 (12.6) Mathieu 2007 315 20.06 (14.51) 295 21.89 (14.51) 3.2 % -1.83 [ -4.13, 0.47 ] 3.2 % -2.50 [ -4.93, -0.07 ] 205 202 17.5 (12.5) McAlister 2005 15 (12.5) Montgomery 2003 27.1 (10) 58 44.2 (19.3) 50 2.6 % -17.10 [ -22.79, -11.41 ] -4.20 [ -7.12, -1.28 ] 3.1 % 23.6 (15.1) 27.8 (14.6) 198 Montgomery 2007 201 Montori 2011 49 14.4 (24.92) 46 16.2 (24.92) 1.7 % -1.80 [ -11.83, 8.23 ] -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 225 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

229 ( . . . Continued ) Mean Mean Usual Care Study or subgroup Difference Weight Decision Aid Difference Mean(SD) IV,Random,95% CI N IV,Random,95% CI N Mean(SD) 94 27.5 (37.5) 1.6 % 0.0 [ -10.97, 10.97 ] Morgan 2000 86 27.5 (37.5) 37 14.95 (12.68) 48 14.1 (17.89) Mullan 2009 2.4 % -0.85 [ -7.35, 5.65 ] 2.9 % -7.50 [ -11.89, -3.11 ] 40 (12.5) 57 32.5 (10) Murray 2001a 48 94 96 45 (15) Murray 2001b 37.5 (12.5) -7.50 [ -11.42, -3.58 ] 3.0 % 3.1 % 1.50 [ -1.27, 4.27 ] Nagle 2008 171 16.25 (13.75) 167 17.75 (12.25) 4.6 (9) Nassar 2007 13.5 (19.2) 98 98 -8.90 [ -13.10, -4.70 ] 2.9 % 2.6 % -17.10 [ -22.58, -11.62 ] 69 40.5 (18.3) Protheroe 2007 69 23.4 (14.3) 37 Sawka 2012 52.1 (21.9) 25.2 (13.4) 37 -26.90 [ -35.17, -18.63 ] 2.1 % 2.8 % -4.70 [ -9.18, -0.22 ] Schwalm 2012 14.8 (10.5) 74 19.5 (16.7) 76 23.5 (12.5) Shorten 2005 99 88 29.5 (18.25) -6.00 [ -10.54, -1.46 ] 2.8 % 2.3 % -23.00 [ -30.29, -15.71 ] Shourie 2013 43 14.25 (17.25) 67 37.25 (21.5) 70 11.6 (13.6) 79 20.4 (16.9) Vandemheen 2009 -8.80 [ -13.70, -3.90 ] 2.8 % 2.6 % -4.25 [ -9.88, 1.38 ] 55 20.5 (14.75) 56 24.75 (15.5) Vodermaier 2009 94 10 (12) 107 15.5 (12.9) Whelan 2004 -5.50 [ -8.94, -2.06 ] 3.0 % 100.0 % -7.22 [ -9.12, -5.31 ] Subtotal (95% CI) 4635 4150 2 2 2 Heterogeneity: Tau = 255.34, df = 37 (P<0.00001); I = 28.03; Chi =86% Test for overall effect: Z = 7.42 (P < 0.00001) 2 Uninformed subscale 50 Bekker 2004 32.5 (15) 56 31.67 (14.17) 0.83 [ -4.74, 6.40 ] 3.9 % 1.00 [ -4.39, 6.39 ] 3.9 % 11.1 (15.2) Brazell 2014 12.1 (12.7) 51 53 39.7 (10.6) 61 61.1 (19.7) Chabrera 2015 61 3.8 % -21.40 [ -27.01, -15.79 ] -11.40 [ -16.81, -5.99 ] 3.9 % 69 27.3 (16.61) De Achaval 2012 69 15.9 (15.78) 41 15.75 (13) 37 24.5 (21.25) Dolan 2002 -8.75 [ -16.67, -0.83 ] 3.4 % -48.70 [ -66.15, -31.25 ] 1.7 % 160 57.4 (110.7) 690 Fagerlin 2011 8.7 (43.2) 101 22.8 (22.8) 103 40.6 (21.53) Hess 2012 3.8 % -17.80 [ -23.89, -11.71 ] 2.7 % -8.42 [ -19.58, 2.74 ] Jibaja-Weiss 2011 44 15 (22.26) 39 23.42 (28.72) Laupacis 2006 54 16.25 (13.75) 54 27.25 (15) 3.9 % -11.00 [ -16.43, -5.57 ] 2.8 % -4.50 [ -14.97, 5.97 ] 43 29.75 (22.75) 34.25 (26) Legare 2008a 41 139 15.75 (13.25) 148 21 (14.75) Man-Son-Hing 1999 -5.25 [ -8.49, -2.01 ] 4.2 % 3.8 % -6.70 [ -12.35, -1.05 ] Mann D 2010 80 27.1 (17.6) 70 33.8 (17.6) Mathers 2012 95 18.1 (13.3) 80 26 (16.6) 4.0 % -7.90 [ -12.41, -3.39 ] 4.3 % -2.48 [ -4.96, 0.00 ] Mathieu 2007 315 20.78 (15.59) 295 23.26 (15.59) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 226 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

230 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI Mean(SD) N IV,Random,95% CI 202 -5.00 [ -7.68, -2.32 ] 4.3 % 15 (12.5) McAlister 2005 205 20 (15) 50 22.17 (9.47) Montgomery 2003 58 49.14 (25.4) 3.6 % -26.97 [ -34.01, -19.93 ] -0.70 [ -5.43, 4.03 ] 4.0 % 35.8 (22.7) 199 35.1 (25.6) Montgomery 2007 203 20 (21.5) 27.5 (21.5) Morgan 2000 86 94 3.7 % -7.50 [ -13.79, -1.21 ] -1.63 [ -9.14, 5.88 ] 3.5 % 37 15.28 (15.49) Mullan 2009 48 13.65 (19.84) Murray 2001a 52 27.56 (10.51) 45 38.88 (20.02) 3.7 % -11.32 [ -17.83, -4.81 ] 3.8 % -8.96 [ -14.73, -3.19 ] 93 38.89 (22.53) 93 29.93 (17.26) Murray 2001b 167 15.25 (14.5) Nagle 2008 171 12.75 (14.75) 4.3 % 2.50 [ -0.62, 5.62 ] -6.60 [ -12.17, -1.03 ] 3.9 % 22.3 (20.5) Schwalm 2012 15.7 (13.5) 76 74 69 46.25 (26) 44 11.25 (15.25) Shourie 2013 3.4 % -35.00 [ -42.61, -27.39 ] -12.70 [ -17.77, -7.63 ] 3.9 % 79 4.5 (9.6) 70 Vandemheen 2009 17.2 (20.6) 55 22 (15.75) 56 30 (22.5) Vodermaier 2009 -8.00 [ -15.21, -0.79 ] 3.5 % 4.2 % -4.00 [ -7.50, -0.50 ] 146 25.75 (15) 136 21.75 (15) Wong 2006 3116 2591 Subtotal (95% CI) 100.0 % -9.28 [ -12.20, -6.36 ] 2 2 2 Heterogeneity: Tau = 231.76, df = 26 (P<0.00001); I =89% = 49.45; Chi Test for overall effect: Z = 6.23 (P < 0.00001) 3 Unclear values subscale 4.3 % -1.90 [ -8.82, 5.02 ] 51 17.2 (20.1) Brazell 2014 15.3 (15.5) 53 61 28.1 (11.2) 53.2 (14.5) Chabrera 2015 61 -25.10 [ -29.70, -20.50 ] 4.8 % 4.6 % -8.20 [ -13.92, -2.48 ] De Achaval 2012 69 26.1 (19.11) 69 17.9 (14.95) Dolan 2002 29.25 (24) 41 19.75 (15.75) 37 -9.50 [ -18.61, -0.39 ] 3.7 % 1.7 % -35.10 [ -55.35, -14.85 ] 160 47.7 (128.4) Fagerlin 2011 690 12.6 (50.3) 101 24.2 (25.64) Hess 2012 103 41.4 (22.05) -17.20 [ -23.77, -10.63 ] 4.4 % 2.4 % -15.35 [ -30.66, -0.04 ] 44 14.38 (27.08) 39 29.73 (41.6) Jibaja-Weiss 2011 54 18.75 (16.5) 55 Laupacis 2006 30 (17) 4.4 % -11.25 [ -17.54, -4.96 ] 4.1 % -3.50 [ -11.36, 4.36 ] Legare 2008a 43 19.75 (16.5) 41 23.25 (20) Man-Son-Hing 1999 139 16.25 (12.5) 148 19 (14.75) 5.1 % -2.75 [ -5.91, 0.41 ] 4.8 % -10.00 [ -14.87, -5.13 ] 95 80 26.7 (18.2) Mathers 2012 16.7 (13.9) 315 19.51 (16.3) 295 22.59 (80) Mathieu 2007 -3.08 [ -12.38, 6.22 ] 3.7 % 5.2 % -2.50 [ -5.18, 0.18 ] McAlister 2005 205 15 (12.5) 202 17.5 (15) Montgomery 2003 50 28.5 (12.5) 58 51.29 (25.73) 4.1 % -22.79 [ -30.26, -15.32 ] 5.2 % -6.50 [ -9.34, -3.66 ] Montgomery 2007 201 17.6 (13.2) 203 24.1 (15.8) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 227 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

231 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 94 0.0 [ -0.95, 0.95 ] 5.4 % 30 (3.25) Morgan 2000 86 30 (3.25) 53 35.38 (12.33) Murray 2001a 45 40.56 (16.44) 4.6 % -5.18 [ -11.02, 0.66 ] -5.35 [ -10.16, -0.54 ] 4.8 % 84 42.85 (16.57) Murray 2001b 82 37.5 (15) 19 (15.25) 167 171 15.5 (15.75) Nagle 2008 5.1 % 3.50 [ 0.20, 6.80 ] 4.4 % -8.00 [ -14.50, -1.50 ] Schwalm 2012 76 18 (15.3) 74 26 (24.2) 44 11.25 (13) 69 37.5 (24.25) Shourie 2013 -26.25 [ -33.14, -19.36 ] 4.3 % 4.5 % -6.90 [ -13.12, -0.68 ] Vandemheen 2009 16.8 (21) 70 79 9.9 (17.7) 55 20.75 (15.5) 56 24.75 (15.5) Vodermaier 2009 -4.00 [ -9.77, 1.77 ] 4.6 % 100.0 % -8.81 [ -11.99, -5.63 ] Subtotal (95% CI) 2274 2794 2 2 2 = 273.77, df = 22 (P<0.00001); I =92% Heterogeneity: Tau = 48.93; Chi Test for overall effect: Z = 5.43 (P < 0.00001) 4 Uncertainty subscale 50 45 (20.83) 56 45 (25.83) Bekker 2004 0.0 [ -8.89, 8.89 ] 2.9 % 3.0 % 2.90 [ -5.71, 11.51 ] 53 51 18.8 (23.3) Brazell 2014 21.7 (21.4) 61 32 (11.4) Chabrera 2015 43.8 (8.82) 61 4.7 % -11.80 [ -15.42, -8.18 ] 3.3 % -2.50 [ -10.12, 5.12 ] De Achaval 2012 69 33.4 (23.26) 69 35.9 (22.43) Dolan 2002 41 27 (19.25) 37 26 (24.25) 2.6 % 1.00 [ -8.79, 10.79 ] 0.7 % -35.80 [ -60.98, -10.62 ] 690 37.4 (62.3) Fagerlin 2011 160 73.2 (159.7) 42.5 (20) Gattellari 2003 106 108 42.5 (33.33) 3.4 % 0.0 [ -7.35, 7.35 ] 1.66 [ -2.49, 5.81 ] 4.5 % 29.17 (15) 136 Gattellari 2005 131 30.83 (19.25) Hanson 2011 118 22 (20) 115 28.75 (20) 4.2 % -6.75 [ -11.89, -1.61 ] 3.7 % -12.10 [ -18.54, -5.66 ] Hess 2012 103 36.8 (23.59) 101 24.7 (23.33) 44 15.38 (32.25) Jibaja-Weiss 2011 39 12.83 (22.53) 2.1 % 2.55 [ -9.32, 14.42 ] 3.4 % -2.50 [ -9.97, 4.97 ] 23 (21) 55 Laupacis 2006 54 20.5 (18.75) Legare 2008a 43 26.5 (23) 41 33.25 (25.25) 2.5 % -6.75 [ -17.09, 3.59 ] 1.25 [ -3.39, 5.89 ] 4.4 % 21 (21) 148 139 Man-Son-Hing 1999 19.75 (19) 95 80 29.4 (20.8) 20.1 (16.6) Mathers 2012 4.0 % -9.30 [ -14.95, -3.65 ] -0.42 [ -3.51, 2.67 ] 4.9 % 295 22.65 (19.46) Mathieu 2007 315 22.23 (19.46) 205 20 (20) 202 17.5 (17.5) McAlister 2005 2.50 [ -1.15, 6.15 ] 4.7 % 3.0 % -12.49 [ -21.10, -3.88 ] Montgomery 2003 50 35.5 (20.47) 58 47.99 (25.14) Montgomery 2007 201 22.1 (18.4) 203 27.3 (18.8) 4.7 % -5.20 [ -8.83, -1.57 ] 4.7 % 2.50 [ -1.30, 6.30 ] Morgan 2000 86 35 (13) 94 32.5 (13) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 228 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

232 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 35 (20) 42.5 (20) 3.3 % -7.50 [ -15.18, 0.18 ] 57 Murray 2001a 48 52.5 (25) 96 60 (27.5) Murray 2001b 94 3.4 % -7.50 [ -14.97, -0.03 ] 4.5 % -0.25 [ -4.65, 4.15 ] Nagle 2008 24 (19.75) 171 24.25 (21.5) 167 18 (18.8) 19.6 (19.9) 76 Schwalm 2012 74 -1.60 [ -7.80, 4.60 ] 3.8 % 2.9 % -22.00 [ -30.85, -13.15 ] Shourie 2013 68 38.25 (29.5) 44 16.25 (18.25) 70 26.4 (25.9) 79 Vandemheen 2009 36.4 (27.8) 3.0 % -10.00 [ -18.63, -1.37 ] 3.5 % -3.00 [ -9.92, 3.92 ] 55 Vodermaier 2009 30 (10) 27 (24.25) 56 146 136 38.25 (22.5) Wong 2006 40 (20.83) 4.2 % -1.75 [ -6.82, 3.32 ] 100.0 % -4.04 [ -6.27, -1.81 ] Subtotal (95% CI) 3351 2849 2 2 2 = 107.12, df = 27 (P<0.00001); I =75% Heterogeneity: Tau = 24.09; Chi Test for overall effect: Z = 3.55 (P = 0.00039) 5 Unsupported subscale Brazell 2014 11.5 (14.4) 51 9.5 (13.9) 53 4.3 % 2.00 [ -3.44, 7.44 ] 4.5 % -21.20 [ -26.02, -16.38 ] Chabrera 2015 51.7 (15.3) 61 30.5 (11.6) 61 De Achaval 2012 69 20.5 (14.95) 69 25 (15.78) 4.4 % -4.50 [ -9.63, 0.63 ] -2.25 [ -9.91, 5.41 ] 3.6 % 21 (13.5) 37 Dolan 2002 23.25 (20) 41 18.1 (46.9) Fagerlin 2011 160 43.3 (119.4) 690 1.4 % -25.20 [ -44.03, -6.37 ] 4.1 % -10.70 [ -16.89, -4.51 ] 101 18.5 (22.56) Hess 2012 103 29.2 (22.56) 39 22.07 (28.88) Jibaja-Weiss 2011 44 19.17 (26.3) 2.5 % -2.90 [ -14.84, 9.04 ] -6.75 [ -12.98, -0.52 ] 4.1 % 24 (17.25) 53 17.25 (15.75) 55 Laupacis 2006 Legare 2008a 43 24.25 (19.5) 41 23.5 (17.25) 0.75 [ -7.11, 8.61 ] 3.6 % 4.9 % -0.25 [ -3.37, 2.87 ] 139 Man-Son-Hing 1999 16.5 (14) 16.25 (13) 148 80 25.2 (13.72) Mann D 2010 70 29.6 (13.72) 4.6 % -4.40 [ -8.80, 0.00 ] 4.7 % -3.40 [ -7.66, 0.86 ] 80 20.8 (15.3) Mathers 2012 95 17.4 (13.1) 315 20.9 (15.58) 295 22.98 (15.58) Mathieu 2007 5.1 % -2.08 [ -4.55, 0.39 ] 5.0 % 0.0 [ -2.91, 2.91 ] 15 (15) 202 15 (15) McAlister 2005 205 Montgomery 2003 50 23.67 (10.96) 58 40.52 (19.83) -16.85 [ -22.79, -10.91 ] 4.2 % 4.9 % -6.30 [ -9.75, -2.85 ] 200 22.2 (16.5) 201 28.5 (18.7) Montgomery 2007 86 Morgan 2000 94 32.5 (24.75) 30 (24.75) 3.8 % -2.50 [ -9.74, 4.74 ] 4.1 % -7.86 [ -13.92, -1.80 ] Murray 2001a 53 32.7 (12.75) 45 40.56 (17.1) Murray 2001b 85 36.47 (14.43) 82 48.68 (15.46) 4.6 % -12.21 [ -16.75, -7.67 ] 4.9 % 0.75 [ -2.40, 3.90 ] Nagle 2008 167 15.25 (13.75) 171 14.5 (15.75) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 229 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

233 ( . . . Continued ) Mean Mean Weight Usual Care Difference Study or subgroup Decision Aid Difference Mean(SD) IV,Random,95% CI N IV,Random,95% CI N Mean(SD) 74 14.9 (16.9) Schwalm 2012 -2.70 [ -7.85, 2.45 ] 76 12.2 (15.2) 4.4 % 69 38 (21.75) 43 13.25 (17.25) Shourie 2013 3.7 % -24.75 [ -32.02, -17.48 ] 4.5 % -7.60 [ -12.45, -2.75 ] 6.9 (12.3) 70 14.5 (17.7) Vandemheen 2009 79 56 55 16.25 (16.25) Vodermaier 2009 21 (15.75) 4.2 % -4.75 [ -10.70, 1.20 ] 100.0 % -6.27 [ -8.86, -3.68 ] Subtotal (95% CI) 2874 2340 2 2 2 = 32.67; Chi = 155.62, df = 23 (P<0.00001); I =85% Heterogeneity: Tau Test for overall effect: Z = 4.75 (P < 0.00001) 6 Ineffective choice subscale 50 22.5 (13.75) 56 21.88 (14.38) Bekker 2004 0.62 [ -4.74, 5.98 ] 4.3 % 3.7 % 4.00 [ -3.19, 11.19 ] 17.8 (19.1) Brazell 2014 51 53 13.8 (18.3) 61 27.1 (11.7) 61 Chabrera 2015 49.5 (14.3) 4.5 % -22.40 [ -27.04, -17.76 ] 4.0 % -3.50 [ -9.74, 2.74 ] De Achaval 2012 69 27.7 (18.27) 69 31.2 (19.11) 41 20.5 (14.5) 37 25.75 (21) Dolan 2002 3.5 % -5.25 [ -13.34, 2.84 ] 1.2 % -25.50 [ -46.61, -4.39 ] 160 55.5 (133.9) Fagerlin 2011 690 30 (52.3) 118 115 19.25 (15.55) Hanson 2011 14 (15.55) 4.6 % -5.25 [ -9.24, -1.26 ] -6.25 [ -12.00, -0.50 ] 4.2 % 55 15 (14.5) 53 21.25 (16) Laupacis 2006 43 16.5 (14.75) 41 22.25 (19) Legare 2008a -5.75 [ -13.05, 1.55 ] 3.7 % 4.8 % -2.00 [ -5.21, 1.21 ] 13.5 (13) 148 15.5 (14.75) Man-Son-Hing 1999 139 Mathers 2012 95 80 23.3 (15.2) 16.1 (14.4) -7.20 [ -11.61, -2.79 ] 4.5 % 5.0 % -0.78 [ -3.16, 1.60 ] 295 19.19 (14.96) Mathieu 2007 315 18.41 (14.96) 15 (12.5) 17.5 (15) 205 McAlister 2005 202 4.9 % -2.50 [ -5.18, 0.18 ] 4.3 % -9.13 [ -14.52, -3.74 ] 26 (11.11) 58 35.13 (17.2) Montgomery 2003 50 86 20 (32) 94 Morgan 2000 22.5 (32) 3.1 % -2.50 [ -11.86, 6.86 ] 4.4 % -5.00 [ -9.97, -0.03 ] 57 48 30 (15) Murray 2001a 25 (10) 94 30 (15) 96 Murray 2001b 37.5 (17.5) 4.5 % -7.50 [ -12.13, -2.87 ] 4.9 % 1.25 [ -1.74, 4.24 ] Nagle 2008 167 16.25 (13.75) 171 15 (14.25) Schwalm 2012 76 11.3 (11.4) 74 15.9 (15.9) 4.5 % -4.60 [ -9.04, -0.16 ] 4.1 % -19.50 [ -25.38, -13.62 ] 44 68 30.5 (19.5) Shourie 2013 11 (12.25) 70 10.4 (16.4) 79 Vandemheen 2009 17.9 (20.4) 4.1 % -7.50 [ -13.42, -1.58 ] 3.6 % -6.75 [ -14.33, 0.83 ] Vodermaier 2009 55 28.25 (20.75) 56 35 (20) Whelan 2004 94 12.5 (12) 107 17 (13) 4.8 % -4.50 [ -7.96, -1.04 ] 4.7 % -17.29 [ -21.00, -13.58 ] Wong 2006 136 19.38 (13.13) 159 36.67 (19.17) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 230 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

234 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N IV,Random,95% CI IV,Random,95% CI Mean(SD) N Mean(SD) 2380 100.0 % -6.31 [ -8.93, -3.70 ] Subtotal (95% CI) 2861 2 2 2 = 34.12; Chi = 175.19, df = 23 (P<0.00001); I =87% Heterogeneity: Tau Test for overall effect: Z = 4.73 (P < 0.00001) -10 0 10 20 -20 Favours Usual Care Favours Decision Aid cisional conflict - in consultation. Analysis 4.2. Comparison 4 Decisional conflict, Outcome 2 De Review: Decision aids for people facing health treatment or s creening decisions Comparison: 4 Decisional conflict Outcome: 2 Decisional conflict - in consultation Mean Mean Study or subgroup Decision Aid Usual Care Difference Weight Difference N N Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) 1 Uncertainty subscale Bekker 2004 50 45 (20.83) 56 45 (25.83) 46.7 % 0.0 [ -8.89, 8.89 ] -12.10 [ -18.54, -5.66 ] 53.3 % 103 36.8 (23.59) 101 24.7 (23.33) Hess 2012 151 Subtotal (95% CI) 159 100.0 % -6.45 [ -18.29, 5.38 ] 2 2 2 = 57.51; Chi =79% Heterogeneity: Tau = 4.67, df = 1 (P = 0.03); I Test for overall effect: Z = 1.07 (P = 0.28) 2 Uninformed subscale 25.7 % 0.83 [ -4.74, 6.40 ] 56 31.67 (14.17) 32.5 (15) Bekker 2004 50 101 22.8 (22.8) Hess 2012 103 40.6 (21.53) 25.2 % -17.80 [ -23.89, -11.71 ] 25.7 % -6.70 [ -12.35, -1.05 ] 80 27.1 (17.6) Mann D 2010 70 33.8 (17.6) Mullan 2009 48 13.65 (19.84) 37 15.28 (15.49) 23.5 % -1.63 [ -9.14, 5.88 ] 100.0 % -6.37 [ -14.58, 1.85 ] Subtotal (95% CI) 279 266 2 2 2 = 21.50, df = 3 (P = 0.00008); I Heterogeneity: Tau =86% = 60.19; Chi Test for overall effect: Z = 1.52 (P = 0.13) 3 Unclear values subscale Hess 2012 101 24.2 (25.64) 103 41.4 (22.05) 100.0 % -17.20 [ -23.77, -10.63 ] -20 0 10 20 -10 Favours Decision Aid Favours Usual Care ( Continued . . . ) 231 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

235 ( . . . Continued ) Mean Mean Usual Care Study or subgroup Difference Weight Decision Aid Difference Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) N N Subtotal (95% CI) 100.0 % -17.20 [ -23.77, -10.63 ] 101 103 Heterogeneity: not applicable Test for overall effect: Z = 5.13 (P < 0.00001) 4 Unsupported subscale 101 18.5 (22.56) 103 29.2 (22.56) Hess 2012 -10.70 [ -16.89, -4.51 ] 43.8 % 56.2 % -4.40 [ -8.80, 0.00 ] Mann D 2010 80 25.2 (13.72) 70 29.6 (13.72) 181 Subtotal (95% CI) 173 100.0 % -7.16 [ -13.28, -1.03 ] 2 2 2 = 2.64, df = 1 (P = 0.10); I =62% = 12.33; Chi Heterogeneity: Tau Test for overall effect: Z = 2.29 (P = 0.022) 5 Ineffective choice subscale 41.7 % 0.62 [ -4.74, 5.98 ] Bekker 2004 50 22.5 (13.75) 56 21.88 (14.38) 94 12.5 (12) 107 17 (13) Whelan 2004 58.3 % -4.50 [ -7.96, -1.04 ] 100.0 % -2.37 [ -7.31, 2.58 ] 163 Subtotal (95% CI) 144 2 2 2 Heterogeneity: Tau =60% = 7.81; Chi = 2.48, df = 1 (P = 0.12); I Test for overall effect: Z = 0.94 (P = 0.35) 6 Total decisional conflict score 20.5 % -20.00 [ -25.46, -14.54 ] 101 23.3 (20.76) 103 43.3 (18.97) Hess 2012 80 25.5 (11.14) 70 28.5 (11.14) Mann D 2010 -3.00 [ -6.57, 0.57 ] 22.5 % 15.1 % -1.80 [ -11.83, 8.23 ] 49 14.4 (24.92) 46 16.2 (24.92) Montori 2011 48 14.1 (17.89) 37 14.95 (12.68) Mullan 2009 -0.85 [ -7.35, 5.65 ] 19.3 % 22.6 % -5.50 [ -8.94, -2.06 ] 10 (12) 107 15.5 (12.9) Whelan 2004 94 Subtotal (95% CI) 372 363 100.0 % -6.46 [ -12.78, -0.14 ] 2 2 2 Heterogeneity: Tau = 42.94; Chi = 31.11, df = 4 (P<0.00001); I =87% Test for overall effect: Z = 2.00 (P = 0.045) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care 232 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

236 Analysis 4.3. Comparison 4 Decisional conflict, Outcome 3 De cisional conflict - in preparation for consultation. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 4 Decisional conflict Outcome: 3 Decisional conflict - in preparation for consulta tion Mean Mean Study or subgroup Usual Care Difference Difference Decision Aid Weight Mean(SD) IV,Random,95% CI IV,Random,95% CI N N Mean(SD) 1 Uncertainty subscale 21.7 (21.4) Brazell 2014 18.8 (23.3) 53 51 2.90 [ -5.71, 11.51 ] 3.2 % 32 (11.4) 43.8 (8.82) 61 Chabrera 2015 61 5.1 % -11.80 [ -15.42, -8.18 ] 3.5 % -2.50 [ -10.12, 5.12 ] 69 33.4 (23.26) De Achaval 2012 69 35.9 (22.43) Dolan 2002 41 27 (19.25) 37 26 (24.25) 1.00 [ -8.79, 10.79 ] 2.8 % 0.7 % -35.80 [ -60.98, -10.62 ] 690 Fagerlin 2011 37.4 (62.3) 160 73.2 (159.7) 42.5 (20) Gattellari 2003 106 108 42.5 (33.33) 3.6 % 0.0 [ -7.35, 7.35 ] 1.66 [ -2.49, 5.81 ] 4.9 % 29.17 (15) 136 Gattellari 2005 131 30.83 (19.25) Hanson 2011 118 22 (20) 115 28.75 (20) -6.75 [ -11.89, -1.61 ] 4.5 % 2.3 % 2.55 [ -9.32, 14.42 ] 44 15.38 (32.25) 39 12.83 (22.53) Jibaja-Weiss 2011 Laupacis 2006 54 20.5 (18.75) 55 23 (21) -2.50 [ -9.97, 4.97 ] 3.6 % 2.7 % -6.75 [ -17.09, 3.59 ] 43 26.5 (23) 41 33.25 (25.25) Legare 2008a 4.7 % 1.25 [ -3.39, 5.89 ] 21 (21) 148 19.75 (19) Man-Son-Hing 1999 139 Mathers 2012 95 80 29.4 (20.8) 20.1 (16.6) -9.30 [ -14.95, -3.65 ] 4.3 % -0.42 [ -3.51, 2.67 ] 5.3 % 295 22.65 (19.46) 315 22.23 (19.46) Mathieu 2007 20 (20) McAlister 2005 17.5 (17.5) 205 202 2.50 [ -1.15, 6.15 ] 5.1 % -12.49 [ -21.10, -3.88 ] 3.2 % 58 47.99 (25.14) Montgomery 2003 50 35.5 (20.47) 201 22.1 (18.4) 203 27.3 (18.8) Montgomery 2007 -5.20 [ -8.83, -1.57 ] 5.1 % 86 35 (13) 94 32.5 (13) Morgan 2000 2.50 [ -1.30, 6.30 ] 5.0 % 3.5 % -7.50 [ -15.18, 0.18 ] 57 35 (20) 48 42.5 (20) Murray 2001a Murray 2001b 94 52.5 (25) 96 60 (27.5) 3.6 % -7.50 [ -14.97, -0.03 ] 4.8 % -0.25 [ -4.65, 4.15 ] 167 171 24.25 (21.5) Nagle 2008 24 (19.75) 76 18 (18.8) 74 Schwalm 2012 19.6 (19.9) 4.1 % -1.60 [ -7.80, 4.60 ] 3.1 % -22.00 [ -30.85, -13.15 ] Shourie 2013 44 16.25 (18.25) 68 38.25 (29.5) Vandemheen 2009 70 26.4 (25.9) 79 36.4 (27.8) 3.2 % -10.00 [ -18.63, -1.37 ] -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 233 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

237 ( . . . ) Continued Mean Mean Decision Aid Weight Study or subgroup Difference Usual Care Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 56 3.8 % 27 (24.25) -3.00 [ -9.92, 3.92 ] Vodermaier 2009 30 (10) 55 40 (20.83) 146 Wong 2006 136 38.25 (22.5) 4.5 % -1.75 [ -6.82, 3.32 ] 100.0 % -3.83 [ -6.12, -1.55 ] 3200 Subtotal (95% CI) 2690 2 2 2 = 23.28; Chi = 98.99, df = 25 (P<0.00001); I =75% Heterogeneity: Tau Test for overall effect: Z = 3.29 (P = 0.00099) 2 Uninformed subscale 12.1 (12.7) 51 11.1 (15.2) Brazell 2014 53 4.6 % 1.00 [ -4.39, 6.39 ] 4.5 % -21.40 [ -27.01, -15.79 ] 39.7 (10.6) 61 61 Chabrera 2015 61.1 (19.7) 69 15.9 (15.78) De Achaval 2012 69 27.3 (16.61) 4.6 % -11.40 [ -16.81, -5.99 ] -8.75 [ -16.67, -0.83 ] 4.0 % Dolan 2002 15.75 (13) 37 24.5 (21.25) 41 8.7 (43.2) 160 57.4 (110.7) Fagerlin 2011 690 2.0 % -48.70 [ -66.15, -31.25 ] -8.42 [ -19.58, 2.74 ] 3.2 % 39 23.42 (28.72) 44 Jibaja-Weiss 2011 15 (22.26) 54 16.25 (13.75) 54 27.25 (15) Laupacis 2006 -11.00 [ -16.43, -5.57 ] 4.6 % -4.50 [ -14.97, 5.97 ] 3.4 % 34.25 (26) 43 29.75 (22.75) Legare 2008a 41 139 15.75 (13.25) 148 21 (14.75) Man-Son-Hing 1999 5.0 % -5.25 [ -8.49, -2.01 ] 4.8 % -7.90 [ -12.41, -3.39 ] Mathers 2012 95 18.1 (13.3) 80 26 (16.6) Mathieu 2007 315 20.78 (15.59) 295 23.26 (15.59) 5.1 % -2.48 [ -4.96, 0.00 ] 5.1 % -5.00 [ -7.68, -2.32 ] 15 (12.5) McAlister 2005 202 205 20 (15) 58 49.14 (25.4) 50 22.17 (9.47) Montgomery 2003 -26.97 [ -34.01, -19.93 ] 4.2 % 4.7 % -0.70 [ -5.43, 4.03 ] 35.8 (22.7) 199 Montgomery 2007 203 35.1 (25.6) 86 20 (21.5) 94 27.5 (21.5) Morgan 2000 -7.50 [ -13.79, -1.21 ] 4.4 % 4.3 % -11.32 [ -17.83, -4.81 ] 52 27.56 (10.51) Murray 2001a 45 38.88 (20.02) Murray 2001b 93 29.93 (17.26) 93 38.89 (22.53) 4.5 % -8.96 [ -14.73, -3.19 ] 2.50 [ -0.62, 5.62 ] 5.0 % 171 12.75 (14.75) 167 15.25 (14.5) Nagle 2008 76 15.7 (13.5) 74 22.3 (20.5) Schwalm 2012 -6.60 [ -12.17, -1.03 ] 4.5 % 4.0 % -35.00 [ -42.61, -27.39 ] 69 46.25 (26) Shourie 2013 44 11.25 (15.25) Vandemheen 2009 70 17.2 (20.6) 4.5 (9.6) 79 -12.70 [ -17.77, -7.63 ] 4.6 % 4.1 % -8.00 [ -15.21, -0.79 ] Vodermaier 2009 55 56 30 (22.5) 22 (15.75) 136 21.75 (15) 146 25.75 (15) Wong 2006 -4.00 [ -7.50, -0.50 ] 4.9 % 100.0 % -9.81 [ -13.00, -6.61 ] Subtotal (95% CI) 2837 2325 2 2 2 Heterogeneity: Tau = 210.26, df = 22 (P<0.00001); I = 50.54; Chi =90% Test for overall effect: Z = 6.02 (P < 0.00001) 3 Unclear values subscale -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 234 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

238 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 51 -1.90 [ -8.82, 5.02 ] 4.5 % 15.3 (15.5) Brazell 2014 53 17.2 (20.1) 61 53.2 (14.5) 61 Chabrera 2015 28.1 (11.2) 5.1 % -25.10 [ -29.70, -20.50 ] -8.20 [ -13.92, -2.48 ] 4.8 % 69 17.9 (14.95) De Achaval 2012 69 26.1 (19.11) 37 41 19.75 (15.75) 29.25 (24) Dolan 2002 3.9 % -9.50 [ -18.61, -0.39 ] 1.7 % -35.10 [ -55.35, -14.85 ] 690 12.6 (50.3) Fagerlin 2011 160 47.7 (128.4) 44 14.38 (27.08) 39 29.73 (41.6) Jibaja-Weiss 2011 -15.35 [ -30.66, -0.04 ] 2.5 % 4.6 % -11.25 [ -17.54, -4.96 ] Laupacis 2006 54 18.75 (16.5) 30 (17) 55 43 19.75 (16.5) 41 Legare 2008a 23.25 (20) -3.50 [ -11.36, 4.36 ] 4.2 % -2.75 [ -5.91, 0.41 ] 5.4 % Man-Son-Hing 1999 148 19 (14.75) 139 16.25 (12.5) 16.7 (13.9) 80 26.7 (18.2) Mathers 2012 95 5.0 % -10.00 [ -14.87, -5.13 ] 3.8 % -3.08 [ -12.38, 6.22 ] 22.59 (80) 315 19.51 (16.3) Mathieu 2007 295 205 202 17.5 (15) 15 (12.5) McAlister 2005 5.5 % -2.50 [ -5.18, 0.18 ] 4.3 % -22.79 [ -30.26, -15.32 ] 58 51.29 (25.73) 28.5 (12.5) Montgomery 2003 50 201 Montgomery 2007 203 24.1 (15.8) 17.6 (13.2) 5.4 % -6.50 [ -9.34, -3.66 ] 0.0 [ -0.95, 0.95 ] 5.6 % 30 (3.25) 94 86 Morgan 2000 30 (3.25) 53 35.38 (12.33) Murray 2001a 45 40.56 (16.44) 4.8 % -5.18 [ -11.02, 0.66 ] -5.35 [ -10.16, -0.54 ] 5.0 % Murray 2001b 37.5 (15) 84 42.85 (16.57) 82 19 (15.25) 171 15.5 (15.75) Nagle 2008 167 5.3 % 3.50 [ 0.20, 6.80 ] -8.00 [ -14.50, -1.50 ] 4.6 % 74 18 (15.3) 76 Schwalm 2012 26 (24.2) 44 11.25 (13) 69 37.5 (24.25) Shourie 2013 -26.25 [ -33.14, -19.36 ] 4.5 % -6.90 [ -13.12, -0.68 ] 4.7 % 79 9.9 (17.7) 70 Vandemheen 2009 16.8 (21) 55 20.75 (15.5) 56 24.75 (15.5) Vodermaier 2009 4.8 % -4.00 [ -9.77, 1.77 ] 100.0 % -8.40 [ -11.59, -5.21 ] Subtotal (95% CI) 2693 2171 2 2 2 Heterogeneity: Tau = 255.50, df = 21 (P<0.00001); I =92% = 46.63; Chi Test for overall effect: Z = 5.17 (P < 0.00001) 4 Unsupported subscale 53 51 9.5 (13.9) Brazell 2014 11.5 (14.4) 4.7 % 2.00 [ -3.44, 7.44 ] 4.9 % -21.20 [ -26.02, -16.38 ] 61 Chabrera 2015 61 30.5 (11.6) 51.7 (15.3) De Achaval 2012 69 20.5 (14.95) 69 25 (15.78) 4.8 % -4.50 [ -9.63, 0.63 ] -2.25 [ -9.91, 5.41 ] 4.0 % 21 (13.5) 37 41 Dolan 2002 23.25 (20) 690 160 43.3 (119.4) Fagerlin 2011 18.1 (46.9) -25.20 [ -44.03, -6.37 ] 1.6 % 2.8 % -2.90 [ -14.84, 9.04 ] 44 19.17 (26.3) 39 22.07 (28.88) Jibaja-Weiss 2011 -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 235 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

239 ( . . . ) Continued Mean Mean Decision Aid Study or subgroup Difference Usual Care Weight Difference Mean(SD) IV,Random,95% CI N IV,Random,95% CI N Mean(SD) 24 (17.25) 4.5 % -6.75 [ -12.98, -0.52 ] Laupacis 2006 53 17.25 (15.75) 55 41 23.5 (17.25) 43 24.25 (19.5) Legare 2008a 3.9 % 0.75 [ -7.11, 8.61 ] 5.4 % -0.25 [ -3.37, 2.87 ] 16.5 (14) Man-Son-Hing 1999 139 16.25 (13) 148 17.4 (13.1) 95 Mathers 2012 80 20.8 (15.3) -3.40 [ -7.66, 0.86 ] 5.1 % 5.5 % -2.08 [ -4.55, 0.39 ] 315 20.9 (15.58) Mathieu 2007 295 22.98 (15.58) 15 (15) 202 15 (15) McAlister 2005 205 0.0 [ -2.91, 2.91 ] 5.4 % 4.6 % -16.85 [ -22.79, -10.91 ] 50 23.67 (10.96) Montgomery 2003 58 40.52 (19.83) Montgomery 2007 22.2 (16.5) 201 28.5 (18.7) 200 -6.30 [ -9.75, -2.85 ] 5.3 % 4.1 % -2.50 [ -9.74, 4.74 ] Morgan 2000 30 (24.75) 94 32.5 (24.75) 86 45 40.56 (17.1) 53 32.7 (12.75) Murray 2001a -7.86 [ -13.92, -1.80 ] 4.5 % 5.0 % -12.21 [ -16.75, -7.67 ] Murray 2001b 85 36.47 (14.43) 82 48.68 (15.46) 167 15.25 (13.75) 171 14.5 (15.75) Nagle 2008 5.4 % 0.75 [ -2.40, 3.90 ] 4.8 % -2.70 [ -7.85, 2.45 ] 76 74 14.9 (16.9) Schwalm 2012 12.2 (15.2) 43 13.25 (17.25) 69 38 (21.75) Shourie 2013 -24.75 [ -32.02, -17.48 ] 4.1 % 4.9 % -7.60 [ -12.45, -2.75 ] Vandemheen 2009 70 6.9 (12.3) 79 14.5 (17.7) Vodermaier 2009 55 16.25 (16.25) 56 21 (15.75) 4.6 % -4.75 [ -10.70, 1.20 ] 100.0 % -6.18 [ -8.96, -3.40 ] 2693 2167 Subtotal (95% CI) 2 2 2 = 34.85; Chi = 151.76, df = 21 (P<0.00001); I =86% Heterogeneity: Tau Test for overall effect: Z = 4.36 (P = 0.000013) 5 Ineffective choice subscale 17.8 (19.1) 51 13.8 (18.3) Brazell 2014 53 4.1 % 4.00 [ -3.19, 11.19 ] 4.9 % -22.40 [ -27.04, -17.76 ] 27.1 (11.7) 49.5 (14.3) 61 Chabrera 2015 61 69 27.7 (18.27) 69 31.2 (19.11) De Achaval 2012 4.4 % -3.50 [ -9.74, 2.74 ] -5.25 [ -13.34, 2.84 ] 3.9 % 37 20.5 (14.5) 41 Dolan 2002 25.75 (21) 690 30 (52.3) 160 55.5 (133.9) Fagerlin 2011 -25.50 [ -46.61, -4.39 ] 1.4 % 5.1 % -5.25 [ -9.24, -1.26 ] Hanson 2011 118 14 (15.55) 115 19.25 (15.55) Laupacis 2006 53 15 (14.5) 55 21.25 (16) 4.6 % -6.25 [ -12.00, -0.50 ] 4.1 % -5.75 [ -13.05, 1.55 ] 43 16.5 (14.75) 22.25 (19) Legare 2008a 41 Man-Son-Hing 1999 139 13.5 (13) 148 15.5 (14.75) -2.00 [ -5.21, 1.21 ] 5.3 % 5.0 % -7.20 [ -11.61, -2.79 ] 95 16.1 (14.4) 80 23.3 (15.2) Mathers 2012 Mathieu 2007 315 18.41 (14.96) 295 19.19 (14.96) 5.4 % -0.78 [ -3.16, 1.60 ] 5.4 % -2.50 [ -5.18, 0.18 ] McAlister 2005 15 (12.5) 202 17.5 (15) 205 -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 236 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

240 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI Mean(SD) N IV,Random,95% CI 58 35.13 (17.2) -9.13 [ -14.52, -3.74 ] 26 (11.11) 4.7 % Montgomery 2003 50 Morgan 2000 22.5 (32) 86 20 (32) 94 -2.50 [ -11.86, 6.86 ] 3.5 % -5.00 [ -9.97, -0.03 ] 4.8 % 30 (15) 57 25 (10) Murray 2001a 48 30 (15) 37.5 (17.5) Murray 2001b 94 96 4.9 % -7.50 [ -12.13, -2.87 ] 1.25 [ -1.74, 4.24 ] 5.3 % 15 (14.25) 171 Nagle 2008 167 16.25 (13.75) Schwalm 2012 76 11.3 (11.4) 74 15.9 (15.9) -4.60 [ -9.04, -0.16 ] 5.0 % 4.5 % -19.50 [ -25.38, -13.62 ] 44 30.5 (19.5) Shourie 2013 68 11 (12.25) 70 10.4 (16.4) 17.9 (20.4) Vandemheen 2009 79 -7.50 [ -13.42, -1.58 ] 4.5 % 4.0 % -6.75 [ -14.33, 0.83 ] Vodermaier 2009 56 35 (20) 55 28.25 (20.75) 159 36.67 (19.17) Wong 2006 136 19.38 (13.13) 5.1 % -17.29 [ -21.00, -13.58 ] 100.0 % -6.75 [ -9.59, -3.90 ] 2217 Subtotal (95% CI) 2717 2 2 2 = 170.48, df = 21 (P<0.00001); I Heterogeneity: Tau =88% = 37.39; Chi Test for overall effect: Z = 4.65 (P < 0.00001) 6 Total decisional conflict score 291 14 (34.29) 334 20 (37.83) Allen 2010 -6.00 [ -11.66, -0.34 ] 3.0 % 3.0 % 1.70 [ -4.09, 7.49 ] 53 15.8 (13.9) 51 14.1 (16.1) Brazell 2014 Chabrera 2015 61 31.2 (10.2) 61 51.7 (13.3) 3.3 % -20.50 [ -24.71, -16.29 ] 3.1 % -5.80 [ -11.07, -0.53 ] De Achaval 2012 69 29.2 (16.61) 69 23.4 (14.95) 20.75 (13) 41 Dolan 2002 37 25.75 (20.25) 2.5 % -5.00 [ -12.64, 2.64 ] -11.50 [ -18.35, -4.65 ] 2.7 % 103 49.58 (24.15) 89 38.08 (24.15) Evans 2010 690 22 (42.2) 160 55.7 (108.4) Fagerlin 2011 1.0 % -33.70 [ -50.79, -16.61 ] 3.2 % -8.00 [ -12.76, -3.24 ] 118 16.25 (18.55) Hanson 2011 115 24.25 (18.55) Jibaja-Weiss 2011 39 22.16 (25.29) 44 16.53 (19.9) -5.63 [ -15.51, 4.25 ] 2.0 % -2.00 [ -7.38, 3.38 ] 3.1 % 81 22 (17) 73 Knops 2014 24 (17) 357 Kuppermann 2014 353 13.8 (15.6) 12.9 (14.1) 3.7 % -0.90 [ -3.09, 1.29 ] -4.10 [ -8.26, 0.06 ] 3.3 % 15.8 (15.5) 112 113 Lam 2013 19.9 (16.3) 53 17.5 (13.75) 54 25.25 (14.25) Laupacis 2006 3.1 % -7.75 [ -13.06, -2.44 ] -4.00 [ -10.32, 2.32 ] 2.8 % 41 23 (14.25) 43 Legare 2008a 27 (15.25) 215 34.15 (24.03) 216 39.85 (24.04) Lepore 2012 3.3 % -5.70 [ -10.24, -1.16 ] 3.6 % -2.25 [ -5.12, 0.62 ] Man-Son-Hing 1999 139 16.25 (11.25) 148 18.5 (13.5) Mathers 2012 95 17.4 (12.6) 80 25.2 (14.9) 3.3 % -7.80 [ -11.93, -3.67 ] 3.7 % -1.83 [ -4.13, 0.47 ] Mathieu 2007 315 20.06 (14.51) 295 21.89 (14.51) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 237 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

241 ( . . . ) Continued Mean Mean Decision Aid Weight Study or subgroup Difference Usual Care Difference N IV,Random,95% CI IV,Random,95% CI Mean(SD) Mean(SD) N 15 (12.5) 202 17.5 (12.5) 3.7 % McAlister 2005 205 -2.50 [ -4.93, -0.07 ] 50 27.1 (10) 58 44.2 (19.3) Montgomery 2003 3.0 % -17.10 [ -22.79, -11.41 ] 3.6 % -4.20 [ -7.12, -1.28 ] 201 27.8 (14.6) 198 Montgomery 2007 23.6 (15.1) 86 27.5 (37.5) 27.5 (37.5) Morgan 2000 94 0.0 [ -10.97, 10.97 ] 1.8 % 3.3 % -7.50 [ -11.89, -3.11 ] Murray 2001a 57 48 40 (12.5) 32.5 (10) 94 Murray 2001b 96 45 (15) 37.5 (12.5) 3.4 % -7.50 [ -11.42, -3.58 ] 3.6 % 1.50 [ -1.27, 4.27 ] 167 17.75 (12.25) 171 16.25 (13.75) Nagle 2008 Nassar 2007 98 4.6 (9) 98 13.5 (19.2) 3.3 % -8.90 [ -13.10, -4.70 ] 3.0 % -17.10 [ -22.58, -11.62 ] Protheroe 2007 40.5 (18.3) 69 69 23.4 (14.3) 37 37 52.1 (21.9) 25.2 (13.4) Sawka 2012 2.4 % -26.90 [ -35.17, -18.63 ] 3.3 % -4.70 [ -9.18, -0.22 ] 14.8 (10.5) 74 76 Schwalm 2012 19.5 (16.7) 99 88 29.5 (18.25) Shorten 2005 23.5 (12.5) -6.00 [ -10.54, -1.46 ] 3.3 % 2.6 % -23.00 [ -30.29, -15.71 ] 43 14.25 (17.25) 67 37.25 (21.5) Shourie 2013 Vandemheen 2009 70 11.6 (13.6) 79 20.4 (16.9) 3.2 % -8.80 [ -13.70, -3.90 ] 3.0 % -4.25 [ -9.88, 1.38 ] 55 20.5 (14.75) 56 24.75 (15.5) Vodermaier 2009 4263 3787 Subtotal (95% CI) 100.0 % -7.32 [ -9.35, -5.28 ] 2 2 2 Heterogeneity: Tau = 223.37, df = 32 (P<0.00001); I = 27.77; Chi =86% Test for overall effect: Z = 7.05 (P < 0.00001) -10 0 10 20 -20 Favours Decision Aid Favours Usual Care 238 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

242 Analysis 4.4. Comparison 4 Decisional conflict, Outcome 4 De cisional conflict - without studies having high risk of bias. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 4 Decisional conflict risk of bias Outcome: 4 Decisional conflict - without studies having high Mean Mean Weight Difference Decision Aid Difference Study or subgroup Usual Care Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) N N 1 Uncertainty subscale Bekker 2004 56 45 (25.83) 50 45 (20.83) 0.0 [ -8.89, 8.89 ] 3.1 % 32 (11.4) 43.8 (8.82) Chabrera 2015 61 61 5.1 % -11.80 [ -15.42, -8.18 ] 3.6 % -2.50 [ -10.12, 5.12 ] 69 35.9 (22.43) De Achaval 2012 69 33.4 (23.26) 41 27 (19.25) 37 26 (24.25) Dolan 2002 1.00 [ -8.79, 10.79 ] 2.9 % 0.8 % -35.80 [ -60.98, -10.62 ] 690 37.4 (62.3) Fagerlin 2011 160 73.2 (159.7) 42.5 (20) 106 Gattellari 2003 108 42.5 (33.33) 3.7 % 0.0 [ -7.35, 7.35 ] 4.9 % 1.66 [ -2.49, 5.81 ] 131 30.83 (19.25) 29.17 (15) Gattellari 2005 136 118 22 (20) Hanson 2011 28.75 (20) 115 4.5 % -6.75 [ -11.89, -1.61 ] 4.0 % -12.10 [ -18.54, -5.66 ] 101 24.7 (23.33) Hess 2012 103 36.8 (23.59) 44 15.38 (32.25) 39 12.83 (22.53) Jibaja-Weiss 2011 2.3 % 2.55 [ -9.32, 14.42 ] 3.6 % -2.50 [ -9.97, 4.97 ] Laupacis 2006 54 20.5 (18.75) 55 23 (21) -6.75 [ -17.09, 3.59 ] 2.7 % 26.5 (23) 41 33.25 (25.25) Legare 2008a 43 95 20.1 (16.6) 29.4 (20.8) Mathers 2012 80 -9.30 [ -14.95, -3.65 ] 4.3 % 5.2 % -0.42 [ -3.51, 2.67 ] 295 22.65 (19.46) Mathieu 2007 315 22.23 (19.46) 20 (20) 202 17.5 (17.5) McAlister 2005 205 2.50 [ -1.15, 6.15 ] 5.1 % 3.2 % -12.49 [ -21.10, -3.88 ] Montgomery 2003 50 35.5 (20.47) 58 47.99 (25.14) 201 203 27.3 (18.8) Montgomery 2007 22.1 (18.4) -5.20 [ -8.83, -1.57 ] 5.1 % Morgan 2000 35 (13) 94 32.5 (13) 86 5.0 % 2.50 [ -1.30, 6.30 ] -7.50 [ -15.18, 0.18 ] 3.6 % 35 (20) 48 Murray 2001a 42.5 (20) 57 52.5 (25) 96 94 Murray 2001b 60 (27.5) 3.6 % -7.50 [ -14.97, -0.03 ] 4.8 % -0.25 [ -4.65, 4.15 ] 171 24.25 (21.5) Nagle 2008 167 24 (19.75) 76 18 (18.8) 74 19.6 (19.9) Schwalm 2012 4.1 % -1.60 [ -7.80, 4.60 ] 3.2 % -22.00 [ -30.85, -13.15 ] 68 38.25 (29.5) Shourie 2013 44 16.25 (18.25) Vandemheen 2009 26.4 (25.9) 79 36.4 (27.8) 70 3.2 % -10.00 [ -18.63, -1.37 ] -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 239 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

243 ( . . . ) Continued Mean Mean Decision Aid Weight Study or subgroup Difference Usual Care Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 56 3.8 % 27 (24.25) -3.00 [ -9.92, 3.92 ] Vodermaier 2009 30 (10) 55 40 (20.83) 146 Wong 2006 136 38.25 (22.5) 4.5 % -1.75 [ -6.82, 3.32 ] 100.0 % -4.53 [ -6.87, -2.18 ] 3159 Subtotal (95% CI) 2650 2 2 2 = 24.77; Chi = 101.56, df = 25 (P<0.00001); I =75% Heterogeneity: Tau Test for overall effect: Z = 3.79 (P = 0.00015) 2 Uninformed subscale 50 32.5 (15) 56 31.67 (14.17) Bekker 2004 4.2 % 0.83 [ -4.74, 6.40 ] 4.2 % -21.40 [ -27.01, -15.79 ] 39.7 (10.6) 61 61 Chabrera 2015 61.1 (19.7) 69 15.9 (15.78) De Achaval 2012 69 27.3 (16.61) 4.2 % -11.40 [ -16.81, -5.99 ] -8.75 [ -16.67, -0.83 ] 3.7 % Dolan 2002 15.75 (13) 37 24.5 (21.25) 41 8.7 (43.2) 160 57.4 (110.7) Fagerlin 2011 690 2.0 % -48.70 [ -66.15, -31.25 ] -17.80 [ -23.89, -11.71 ] 4.1 % 103 40.6 (21.53) 101 Hess 2012 22.8 (22.8) 44 15 (22.26) 39 23.42 (28.72) Jibaja-Weiss 2011 -8.42 [ -19.58, 2.74 ] 3.0 % -11.00 [ -16.43, -5.57 ] 4.2 % 27.25 (15) 54 16.25 (13.75) Laupacis 2006 54 43 29.75 (22.75) 41 34.25 (26) Legare 2008a 3.2 % -4.50 [ -14.97, 5.97 ] 4.2 % -6.70 [ -12.35, -1.05 ] Mann D 2010 80 27.1 (17.6) 70 33.8 (17.6) Mathers 2012 95 18.1 (13.3) 80 26 (16.6) 4.4 % -7.90 [ -12.41, -3.39 ] 4.7 % -2.48 [ -4.96, 0.00 ] Mathieu 2007 295 23.26 (15.59) 315 20.78 (15.59) 202 15 (12.5) 205 McAlister 2005 20 (15) 4.6 % -5.00 [ -7.68, -2.32 ] -26.97 [ -34.01, -19.93 ] 3.9 % 58 49.14 (25.4) 50 22.17 (9.47) Montgomery 2003 199 35.1 (25.6) 203 35.8 (22.7) Montgomery 2007 4.3 % -0.70 [ -5.43, 4.03 ] 4.0 % -7.50 [ -13.79, -1.21 ] 94 27.5 (21.5) 86 Morgan 2000 20 (21.5) Mullan 2009 37 15.28 (15.49) 48 13.65 (19.84) -1.63 [ -9.14, 5.88 ] 3.8 % 4.0 % -11.32 [ -17.83, -4.81 ] 52 27.56 (10.51) 45 38.88 (20.02) Murray 2001a 93 29.93 (17.26) Murray 2001b 93 38.89 (22.53) 4.1 % -8.96 [ -14.73, -3.19 ] 4.6 % 2.50 [ -0.62, 5.62 ] Nagle 2008 167 15.25 (14.5) 171 12.75 (14.75) Schwalm 2012 76 15.7 (13.5) 74 22.3 (20.5) 4.2 % -6.60 [ -12.17, -1.03 ] 3.8 % -35.00 [ -42.61, -27.39 ] 44 11.25 (15.25) 46.25 (26) Shourie 2013 69 70 4.5 (9.6) 79 17.2 (20.6) Vandemheen 2009 -12.70 [ -17.77, -7.63 ] 4.3 % 3.9 % -8.00 [ -15.21, -0.79 ] Vodermaier 2009 55 22 (15.75) 56 30 (22.5) Wong 2006 136 21.75 (15) 146 25.75 (15) 4.5 % -4.00 [ -7.50, -0.50 ] 100.0 % -9.96 [ -13.13, -6.78 ] 2924 2392 Subtotal (95% CI) -10 0 10 20 -20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 240 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

244 ( . . . Continued ) Mean Mean Usual Care Study or subgroup Difference Weight Decision Aid Difference Mean(SD) IV,Random,95% CI N IV,Random,95% CI N Mean(SD) 2 2 2 =89% = 54.55; Chi Heterogeneity: Tau = 223.82, df = 24 (P<0.00001); I Test for overall effect: Z = 6.15 (P < 0.00001) 3 Unclear values subscale 28.1 (11.2) 61 Chabrera 2015 61 53.2 (14.5) -25.10 [ -29.70, -20.50 ] 5.3 % -8.20 [ -13.92, -2.48 ] 5.0 % 69 26.1 (19.11) De Achaval 2012 69 17.9 (14.95) Dolan 2002 41 19.75 (15.75) 37 29.25 (24) 4.2 % -9.50 [ -18.61, -0.39 ] 2.0 % -35.10 [ -55.35, -14.85 ] 12.6 (50.3) 160 47.7 (128.4) Fagerlin 2011 690 Hess 2012 101 24.2 (25.64) 103 41.4 (22.05) -17.20 [ -23.77, -10.63 ] 4.8 % -15.35 [ -30.66, -0.04 ] 2.8 % 44 14.38 (27.08) 39 29.73 (41.6) Jibaja-Weiss 2011 55 Laupacis 2006 30 (17) 54 18.75 (16.5) 4.9 % -11.25 [ -17.54, -4.96 ] -3.50 [ -11.36, 4.36 ] 4.5 % 23.25 (20) 43 19.75 (16.5) 41 Legare 2008a 16.7 (13.9) 80 Mathers 2012 95 26.7 (18.2) -10.00 [ -14.87, -5.13 ] 5.2 % 4.1 % -3.08 [ -12.38, 6.22 ] 22.59 (80) Mathieu 2007 315 19.51 (16.3) 295 205 202 17.5 (15) McAlister 2005 15 (12.5) 5.6 % -2.50 [ -5.18, 0.18 ] -22.79 [ -30.26, -15.32 ] 4.6 % 58 51.29 (25.73) 50 28.5 (12.5) Montgomery 2003 201 17.6 (13.2) 203 24.1 (15.8) Montgomery 2007 -6.50 [ -9.34, -3.66 ] 5.6 % 5.8 % 0.0 [ -0.95, 0.95 ] 30 (3.25) 94 30 (3.25) Morgan 2000 86 Murray 2001a 53 35.38 (12.33) 45 40.56 (16.44) -5.18 [ -11.02, 0.66 ] 5.0 % 5.3 % -5.35 [ -10.16, -0.54 ] Murray 2001b 37.5 (15) 84 42.85 (16.57) 82 19 (15.25) 167 Nagle 2008 171 15.5 (15.75) 5.5 % 3.50 [ 0.20, 6.80 ] 4.9 % -8.00 [ -14.50, -1.50 ] 76 74 26 (24.2) Schwalm 2012 18 (15.3) Shourie 2013 11.25 (13) 69 37.5 (24.25) 44 4.8 % -26.25 [ -33.14, -19.36 ] 4.9 % -6.90 [ -13.12, -0.68 ] 70 79 16.8 (21) Vandemheen 2009 9.9 (17.7) 55 20.75 (15.5) Vodermaier 2009 56 24.75 (15.5) 5.0 % -4.00 [ -9.77, 1.77 ] 100.0 % -9.55 [ -13.08, -6.02 ] Subtotal (95% CI) 2602 2075 2 2 2 = 55.88; Chi = 273.66, df = 20 (P<0.00001); I Heterogeneity: Tau =93% Test for overall effect: Z = 5.30 (P < 0.00001) 4 Unsupported subscale Chabrera 2015 61 51.7 (15.3) 61 30.5 (11.6) -21.20 [ -26.02, -16.38 ] 5.0 % 4.9 % -4.50 [ -9.63, 0.63 ] 25 (15.78) De Achaval 2012 69 20.5 (14.95) 69 41 21 (13.5) 37 23.25 (20) Dolan 2002 4.0 % -2.25 [ -9.91, 5.41 ] 1.6 % -25.20 [ -44.03, -6.37 ] 18.1 (46.9) 160 43.3 (119.4) Fagerlin 2011 690 Hess 2012 103 29.2 (22.56) 101 18.5 (22.56) 4.5 % -10.70 [ -16.89, -4.51 ] -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 241 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

245 ( . . . ) Continued Mean Mean Decision Aid Study or subgroup Difference Usual Care Weight Difference Mean(SD) IV,Random,95% CI N N IV,Random,95% CI Mean(SD) 2.8 % -2.90 [ -14.84, 9.04 ] Jibaja-Weiss 2011 44 19.17 (26.3) 39 22.07 (28.88) 24 (17.25) 55 Laupacis 2006 53 17.25 (15.75) 4.5 % -6.75 [ -12.98, -0.52 ] 4.0 % 0.75 [ -7.11, 8.61 ] 43 24.25 (19.5) 41 23.5 (17.25) Legare 2008a 70 29.6 (13.72) Mann D 2010 80 25.2 (13.72) -4.40 [ -8.80, 0.00 ] 5.1 % -3.40 [ -7.66, 0.86 ] 5.1 % 20.8 (15.3) 17.4 (13.1) 80 95 Mathers 2012 Mathieu 2007 315 20.9 (15.58) 295 22.98 (15.58) 5.6 % -2.08 [ -4.55, 0.39 ] 5.5 % 0.0 [ -2.91, 2.91 ] 202 McAlister 2005 205 15 (15) 15 (15) Montgomery 2003 58 40.52 (19.83) 50 23.67 (10.96) -16.85 [ -22.79, -10.91 ] 4.6 % 5.3 % -6.30 [ -9.75, -2.85 ] 28.5 (18.7) Montgomery 2007 22.2 (16.5) 201 200 30 (24.75) 94 32.5 (24.75) Morgan 2000 86 -2.50 [ -9.74, 4.74 ] 4.2 % 4.5 % -7.86 [ -13.92, -1.80 ] Murray 2001a 45 40.56 (17.1) 53 32.7 (12.75) 85 36.47 (14.43) Murray 2001b 82 48.68 (15.46) 5.0 % -12.21 [ -16.75, -7.67 ] 5.4 % 0.75 [ -2.40, 3.90 ] 167 15.25 (13.75) Nagle 2008 171 14.5 (15.75) 76 12.2 (15.2) 74 14.9 (16.9) Schwalm 2012 -2.70 [ -7.85, 2.45 ] 4.8 % 4.1 % -24.75 [ -32.02, -17.48 ] Shourie 2013 43 13.25 (17.25) 69 38 (21.75) Vandemheen 2009 70 6.9 (12.3) 79 14.5 (17.7) 4.9 % -7.60 [ -12.45, -2.75 ] 4.6 % -4.75 [ -10.70, 1.20 ] 55 16.25 (16.25) 56 Vodermaier 2009 21 (15.75) 2141 2682 Subtotal (95% CI) 100.0 % -7.00 [ -9.76, -4.24 ] 2 2 2 = 33.90; Chi =85% Heterogeneity: Tau = 141.02, df = 21 (P<0.00001); I Test for overall effect: Z = 4.98 (P < 0.00001) 5 Ineffective choice subscale 4.7 % 0.62 [ -4.74, 5.98 ] Bekker 2004 50 22.5 (13.75) 56 21.88 (14.38) Chabrera 2015 27.1 (11.7) 61 49.5 (14.3) 61 4.9 % -22.40 [ -27.04, -17.76 ] 4.4 % -3.50 [ -9.74, 2.74 ] 69 27.7 (18.27) De Achaval 2012 69 31.2 (19.11) 41 20.5 (14.5) 37 25.75 (21) Dolan 2002 -5.25 [ -13.34, 2.84 ] 3.8 % 1.3 % -25.50 [ -46.61, -4.39 ] Fagerlin 2011 690 30 (52.3) 160 55.5 (133.9) Hanson 2011 118 14 (15.55) 115 19.25 (15.55) 5.1 % -5.25 [ -9.24, -1.26 ] 4.6 % -6.25 [ -12.00, -0.50 ] 53 15 (14.5) 21.25 (16) 55 Laupacis 2006 41 22.25 (19) 43 16.5 (14.75) Legare 2008a 4.1 % -5.75 [ -13.05, 1.55 ] 5.0 % -7.20 [ -11.61, -2.79 ] 80 23.3 (15.2) 95 16.1 (14.4) Mathers 2012 Mathieu 2007 315 18.41 (14.96) 295 19.19 (14.96) 5.4 % -0.78 [ -3.16, 1.60 ] 5.4 % -2.50 [ -5.18, 0.18 ] McAlister 2005 15 (12.5) 202 17.5 (15) 205 -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 242 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

246 ( . . . ) Continued Mean Mean Decision Aid Weight Usual Care Difference Study or subgroup Difference N Mean(SD) IV,Random,95% CI N Mean(SD) IV,Random,95% CI 58 35.13 (17.2) -9.13 [ -14.52, -3.74 ] 26 (11.11) Montgomery 2003 50 4.7 % 94 22.5 (32) 86 Morgan 2000 20 (32) 3.5 % -2.50 [ -11.86, 6.86 ] 4.8 % -5.00 [ -9.97, -0.03 ] 57 25 (10) 30 (15) Murray 2001a 48 30 (15) 37.5 (17.5) 94 Murray 2001b 96 4.9 % -7.50 [ -12.13, -2.87 ] 5.3 % 1.25 [ -1.74, 4.24 ] 171 15 (14.25) Nagle 2008 167 16.25 (13.75) 76 11.3 (11.4) Schwalm 2012 15.9 (15.9) 74 4.9 % -4.60 [ -9.04, -0.16 ] 4.5 % -19.50 [ -25.38, -13.62 ] 44 11 (12.25) 30.5 (19.5) Shourie 2013 68 10.4 (16.4) 17.9 (20.4) 70 Vandemheen 2009 79 4.5 % -7.50 [ -13.42, -1.58 ] 4.0 % -6.75 [ -14.33, 0.83 ] 56 35 (20) Vodermaier 2009 55 28.25 (20.75) 94 12.5 (12) Whelan 2004 17 (13) 107 5.2 % -4.50 [ -7.96, -1.04 ] 5.1 % -17.29 [ -21.00, -13.58 ] 136 19.38 (13.13) Wong 2006 159 36.67 (19.17) Subtotal (95% CI) 2669 2181 100.0 % -6.97 [ -9.76, -4.18 ] 2 2 2 Heterogeneity: Tau = 35.80; Chi = 164.67, df = 21 (P<0.00001); I =87% Test for overall effect: Z = 4.90 (P < 0.00001) 6 Total decisional conflict score -6.00 [ -11.66, -0.34 ] 2.8 % 14 (34.29) 334 Allen 2010 20 (37.83) 291 31.2 (10.2) 51.7 (13.3) Chabrera 2015 61 61 3.2 % -20.50 [ -24.71, -16.29 ] -5.80 [ -11.07, -0.53 ] 2.9 % 69 29.2 (16.61) 69 23.4 (14.95) De Achaval 2012 20.75 (13) 37 25.75 (20.25) 41 Dolan 2002 -5.00 [ -12.64, 2.64 ] 2.4 % 2.6 % -11.50 [ -18.35, -4.65 ] 89 38.08 (24.15) 103 49.58 (24.15) Evans 2010 690 22 (42.2) 160 55.7 (108.4) Fagerlin 2011 -33.70 [ -50.79, -16.61 ] 1.0 % -8.00 [ -12.76, -3.24 ] 3.0 % 115 24.25 (18.55) 118 16.25 (18.55) Hanson 2011 103 43.3 (18.97) Hess 2012 101 23.3 (20.76) -20.00 [ -25.46, -14.54 ] 2.9 % 2.0 % -5.63 [ -15.51, 4.25 ] Jibaja-Weiss 2011 44 16.53 (19.9) 39 22.16 (25.29) 357 12.9 (14.1) 353 13.8 (15.6) Kuppermann 2014 3.5 % -0.90 [ -3.09, 1.29 ] 3.2 % -4.10 [ -8.26, 0.06 ] 15.8 (15.5) 112 19.9 (16.3) Lam 2013 113 Laupacis 2006 53 17.5 (13.75) 54 25.25 (14.25) -7.75 [ -13.06, -2.44 ] 2.9 % 2.7 % -4.00 [ -10.32, 2.32 ] 43 23 (14.25) 41 27 (15.25) Legare 2008a Lepore 2012 216 39.85 (24.04) 215 34.15 (24.03) 3.1 % -5.70 [ -10.24, -1.16 ] 3.3 % -3.00 [ -6.57, 0.57 ] Mann D 2010 80 25.5 (11.14) 70 28.5 (11.14) Mathers 2012 95 17.4 (12.6) 80 25.2 (14.9) 3.2 % -7.80 [ -11.93, -3.67 ] 3.5 % -1.83 [ -4.13, 0.47 ] Mathieu 2007 315 20.06 (14.51) 295 21.89 (14.51) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care ( Continued . . . ) 243 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

247 ( . . . ) Continued Mean Mean Decision Aid Weight Study or subgroup Difference Usual Care Difference N IV,Random,95% CI IV,Random,95% CI Mean(SD) Mean(SD) N 15 (12.5) 202 17.5 (12.5) 3.5 % McAlister 2005 205 -2.50 [ -4.93, -0.07 ] 50 27.1 (10) 58 44.2 (19.3) Montgomery 2003 2.8 % -17.10 [ -22.79, -11.41 ] -4.20 [ -7.12, -1.28 ] 3.4 % 201 23.6 (15.1) 198 Montgomery 2007 27.8 (14.6) 49 14.4 (24.92) 46 16.2 (24.92) Montori 2011 -1.80 [ -11.83, 8.23 ] 1.9 % 0.0 [ -10.97, 10.97 ] 1.8 % 27.5 (37.5) 94 86 27.5 (37.5) Morgan 2000 Mullan 2009 48 14.1 (17.89) 37 14.95 (12.68) -0.85 [ -7.35, 5.65 ] 2.7 % 3.1 % -7.50 [ -11.89, -3.11 ] 57 40 (12.5) 48 Murray 2001a 32.5 (10) 94 96 Murray 2001b 45 (15) 37.5 (12.5) -7.50 [ -11.42, -3.58 ] 3.2 % 3.4 % 1.50 [ -1.27, 4.27 ] 167 17.75 (12.25) 171 16.25 (13.75) Nagle 2008 Nassar 2007 98 4.6 (9) 98 13.5 (19.2) 3.2 % -8.90 [ -13.10, -4.70 ] 2.9 % -17.10 [ -22.58, -11.62 ] 69 40.5 (18.3) 69 23.4 (14.3) Protheroe 2007 25.2 (13.4) 37 52.1 (21.9) Sawka 2012 37 -26.90 [ -35.17, -18.63 ] 2.3 % 3.1 % -4.70 [ -9.18, -0.22 ] Schwalm 2012 19.5 (16.7) 76 14.8 (10.5) 74 Shorten 2005 23.5 (12.5) 88 29.5 (18.25) 99 -6.00 [ -10.54, -1.46 ] 3.1 % 2.5 % -23.00 [ -30.29, -15.71 ] 43 14.25 (17.25) 67 37.25 (21.5) Shourie 2013 Vandemheen 2009 70 11.6 (13.6) 79 20.4 (16.9) 3.0 % -8.80 [ -13.70, -3.90 ] 2.8 % -4.25 [ -9.88, 1.38 ] 55 20.5 (14.75) Vodermaier 2009 56 24.75 (15.5) 94 10 (12) 107 Whelan 2004 15.5 (12.9) 3.3 % -5.50 [ -8.94, -2.06 ] 100.0 % -7.81 [ -9.84, -5.77 ] Subtotal (95% CI) 4370 3870 2 2 2 = 29.64; Chi Heterogeneity: Tau =86% = 243.60, df = 34 (P<0.00001); I Test for overall effect: Z = 7.51 (P < 0.00001) -20 -10 0 10 20 Favours Decision Aid Favours Usual Care 244 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

248 Analysis 5.1. Comparison 5 Participation in decision makin g, Outcome 1 Participation in decision making - all studies. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 5 Participation in decision making Outcome: 1 Participation in decision making - all studies Decision Aid Risk Ratio Weight R isk Ratio Usual Care Study or subgroup M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 Clinician-controlled decision making 31/103 Auvinen 2004 73/100 0.41 [ 0.30, 0.57 ] 13.9 % Davison 1997 10/30 3/30 2.6 % 0.30 [ 0.09, 0.98 ] 3.5 % 1.17 [ 0.43, 3.19 ] 6/43 Dolan 2002 7/43 6/134 10/139 Kasper 2008 0.62 [ 0.23, 1.66 ] 3.6 % 0.55 [ 0.29, 1.03 ] 7.1 % 14/75 20/196 Krist 2007 Legare 2011 26/81 24/70 10.3 % 0.94 [ 0.59, 1.47 ] 15.0 % 0.90 [ 0.68, 1.20 ] 58/163 65/165 Legare 2012 16/137 23/146 Man-Son-Hing 1999 0.74 [ 0.41, 1.34 ] 7.6 % 5.1 % 0.42 [ 0.19, 0.92 ] 8/92 16/77 Mathers 2012 25/86 39/94 Morgan 2000 0.70 [ 0.47, 1.05 ] 11.4 % 2.4 % 1.05 [ 0.30, 3.70 ] 4/48 Murray 2001a 5/57 Murray 2001b 5/94 6/95 0.84 [ 0.27, 2.67 ] 2.8 % 3.1 % 0.44 [ 0.15, 1.32 ] 9/37 Sawka 2012 4/37 0/173 Smith 2010 3/357 0.5 % 3.40 [ 0.18, 65.50 ] 7.4 % 0.89 [ 0.48, 1.64 ] 16/54 14/53 Vodermaier 2009 6/80 Whelan 2003 12/91 3.9 % 0.57 [ 0.22, 1.45 ] 100.0 % 0.68 [ 0.55, 0.83 ] Subtotal (95% CI) 1437 1743 Total events: 237 (Decision Aid), 327 (Usual Care) 2 2 2 = 0.05; Chi = 23.59, df = 15 (P = 0.07); I =36% Heterogeneity: Tau Test for overall effect: Z = 3.74 (P = 0.00019) 2 Patient-controlled decision making 44/103 9/100 Auvinen 2004 4.75 [ 2.45, 9.20 ] 4.8 % 3.5 % 3.40 [ 1.44, 8.03 ] Davison 1997 5/30 17/30 Dolan 2002 9/43 15/43 4.5 % 0.60 [ 0.29, 1.22 ] 10.4 % 1.10 [ 0.97, 1.25 ] Kasper 2008 103/139 109/134 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid ( Continued . . . ) 245 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

249 ( . . . Continued ) Usual Care Risk Ratio R isk Ratio Decision Aid Study or subgroup Weight M- M- H,Random,95% H,Random,95% n/N n/N CI CI Krist 2007 35/75 9.0 % 1.16 [ 0.88, 1.52 ] 106/196 1.12 [ 0.79, 1.60 ] 8.1 % 30/70 39/81 Legare 2011 57/165 52/163 Legare 2012 0.92 [ 0.68, 1.26 ] 8.6 % 1.13 [ 0.93, 1.38 ] 9.8 % 80/146 85/137 Man-Son-Hing 1999 59/92 33/77 Mathers 2012 8.7 % 1.50 [ 1.11, 2.02 ] 5.0 % 1.33 [ 0.70, 2.53 ] Morgan 2000 17/86 14/94 18/57 Murray 2001a 2/48 1.6 % 7.58 [ 1.85, 31.03 ] 9.1 % 0.93 [ 0.72, 1.22 ] Murray 2001b 53/95 49/94 9/37 Sawka 2012 17/37 1.89 [ 0.97, 3.68 ] 4.8 % 10.9 % 0.98 [ 0.94, 1.02 ] Smith 2010 335/357 166/173 4/53 2/54 Vodermaier 2009 2.04 [ 0.39, 10.66 ] 1.2 % 100.0 % 1.28 [ 1.05, 1.55 ] 1346 1663 Subtotal (95% CI) Total events: 960 (Decision Aid), 613 (Usual Care) 2 2 2 = 109.06, df = 14 (P<0.00001); I Heterogeneity: Tau =87% = 0.09; Chi Test for overall effect: Z = 2.47 (P = 0.013) 3 Shared decision making 4.6 % 1.43 [ 0.82, 2.48 ] 17/100 Auvinen 2004 25/103 Davison 1997 10/30 15/30 0.67 [ 0.36, 1.24 ] 3.9 % 7.6 % 1.23 [ 0.85, 1.78 ] 22/43 Dolan 2002 27/43 26/103 19/134 Kasper 2008 0.56 [ 0.33, 0.96 ] 4.9 % 8.0 % 1.01 [ 0.71, 1.43 ] Krist 2007 71/196 27/75 16/81 16/70 Legare 2011 4.0 % 0.86 [ 0.47, 1.60 ] 8.3 % 1.25 [ 0.89, 1.75 ] 53/163 Legare 2012 43/165 Man-Son-Hing 1999 36/137 43/146 0.89 [ 0.61, 1.30 ] 7.5 % 6.2 % 0.75 [ 0.48, 1.17 ] 25/92 28/77 Mathers 2012 Morgan 2000 42/86 38/94 8.6 % 1.21 [ 0.87, 1.68 ] 0.68 [ 0.54, 0.87 ] 11.0 % Murray 2001a 42/48 34/57 36/95 40/94 Murray 2001b 1.12 [ 0.79, 1.59 ] 8.1 % 5.3 % 0.79 [ 0.48, 1.30 ] Sawka 2012 15/37 19/37 17/357 5/173 Smith 2010 1.8 % 1.65 [ 0.62, 4.39 ] 10.1 % 0.99 [ 0.76, 1.30 ] 36/54 Vodermaier 2009 35/53 Subtotal (95% CI) 1310 1663 100.0 % 0.95 [ 0.83, 1.10 ] Total events: 465 (Decision Aid), 413 (Usual Care) 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid ( Continued . . . ) 246 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

250 ( . . . ) Continued Decision Aid Usual Care Weight R isk Ratio Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 2 2 2 Heterogeneity: Tau =45% = 0.03; Chi = 25.31, df = 14 (P = 0.03); I Test for overall effect: Z = 0.64 (P = 0.52) 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid g, Outcome 2 Participation in decision making - Analysis 5.2. Comparison 5 Participation in decision makin in consultation. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 5 Participation in decision making Outcome: 2 Participation in decision making - in consultati on Usual Care Risk Ratio Weight Study or subgroup isk Ratio Decision Aid R M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 Clinician-controlled decision making - in consultation 26/81 24/70 Legare 2011 25.9 % 0.94 [ 0.59, 1.47 ] Legare 2012 58/163 65/165 67.9 % 0.90 [ 0.68, 1.20 ] 0.57 [ 0.22, 1.45 ] 6.1 % 6/80 12/91 Whelan 2003 324 Subtotal (95% CI) 326 100.0 % 0.89 [ 0.70, 1.12 ] Total events: 90 (Decision Aid), 101 (Usual Care) 2 2 2 = 0.0; Chi = 0.95, df = 2 (P = 0.62); I =0.0% Heterogeneity: Tau Test for overall effect: Z = 1.03 (P = 0.31) 2 Patient-controlled decision making - in consultation Legare 2011 39/81 30/70 43.2 % 1.12 [ 0.79, 1.60 ] Legare 2012 52/163 57/165 56.8 % 0.92 [ 0.68, 1.26 ] 100.0 % 1.01 [ 0.80, 1.27 ] Subtotal (95% CI) 244 235 Total events: 91 (Decision Aid), 87 (Usual Care) 2 2 2 Heterogeneity: Tau = 0.68, df = 1 (P = 0.41); I =0.0% = 0.0; Chi Test for overall effect: Z = 0.04 (P = 0.97) 3 Shared decision making - in consultation 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid ( Continued . . . ) 247 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

251 ( . . . ) Continued Decision Aid Usual Care Weight R isk Ratio Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 16/70 24.7 % Legare 2011 16/81 0.86 [ 0.47, 1.60 ] 75.3 % 1.25 [ 0.89, 1.75 ] 53/163 43/165 Legare 2012 Subtotal (95% CI) 244 235 1.14 [ 0.84, 1.55 ] 100.0 % Total events: 69 (Decision Aid), 59 (Usual Care) 2 2 2 = 1.05, df = 1 (P = 0.30); I = 0.00; Chi =5% Heterogeneity: Tau Test for overall effect: Z = 0.82 (P = 0.41) 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid Analysis 5.3. Comparison 5 Participation in decision makin g, Outcome 3 Participation in decision making - in preparation for consultation. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 5 Participation in decision making n for consultation Outcome: 3 Participation in decision making - in preparatio Decision Aid Study or subgroup Risk Ratio Weight R isk Ratio Usual Care M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 Clinician-controlled decision making 31/103 73/100 Auvinen 2004 23.6 % 0.41 [ 0.30, 0.57 ] 3.1 % 0.30 [ 0.09, 0.98 ] 10/30 3/30 Davison 1997 7/43 6/43 Dolan 2002 1.17 [ 0.43, 3.19 ] 4.3 % 4.4 % 0.62 [ 0.23, 1.66 ] 6/134 10/139 Kasper 2008 20/196 14/75 Krist 2007 0.55 [ 0.29, 1.03 ] 9.5 % 10.4 % 0.74 [ 0.41, 1.34 ] 23/146 Man-Son-Hing 1999 16/137 Mathers 2012 8/92 16/77 0.42 [ 0.19, 0.92 ] 6.5 % 17.8 % 0.70 [ 0.47, 1.05 ] Morgan 2000 25/86 39/94 Murray 2001a 5/57 4/48 2.8 % 1.05 [ 0.30, 3.70 ] 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid ( Continued . . . ) 248 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

252 ( Continued . . . ) Usual Care Study or subgroup R isk Ratio Risk Ratio Decision Aid Weight M- M- H,Random,95% H,Random,95% n/N n/N CI CI 3.3 % 0.84 [ 0.27, 2.67 ] 5/94 Murray 2001b 6/95 3.7 % 0.44 [ 0.15, 1.32 ] Sawka 2012 4/37 9/37 3/357 Smith 2010 0/173 3.40 [ 0.18, 65.50 ] 0.5 % 0.89 [ 0.48, 1.64 ] 10.0 % 16/54 Vodermaier 2009 14/53 1111 Subtotal (95% CI) 1419 100.0 % 0.60 [ 0.48, 0.75 ] Total events: 147 (Decision Aid), 226 (Usual Care) 2 2 2 = 14.51, df = 12 (P = 0.27); I Heterogeneity: Tau = 0.03; Chi =17% Test for overall effect: Z = 4.59 (P < 0.00001) 2 Patient-controlled decision making Auvinen 2004 9/100 6.3 % 4.75 [ 2.45, 9.20 ] 44/103 3.40 [ 1.44, 8.03 ] 4.7 % 17/30 Davison 1997 5/30 Dolan 2002 9/43 15/43 5.8 % 0.60 [ 0.29, 1.22 ] 11.8 % 1.10 [ 0.97, 1.25 ] Kasper 2008 109/134 103/139 106/196 Krist 2007 35/75 10.5 % 1.16 [ 0.88, 1.52 ] 1.13 [ 0.93, 1.38 ] 11.3 % 80/146 85/137 Man-Son-Hing 1999 Mathers 2012 59/92 33/77 10.2 % 1.50 [ 1.11, 2.02 ] 6.4 % 1.33 [ 0.70, 2.53 ] 14/94 Morgan 2000 17/86 Murray 2001a 18/57 2/48 7.58 [ 1.85, 31.03 ] 2.3 % 10.6 % 0.93 [ 0.72, 1.22 ] 53/95 Murray 2001b 49/94 9/37 17/37 Sawka 2012 1.89 [ 0.97, 3.68 ] 6.2 % 12.2 % 0.98 [ 0.94, 1.02 ] Smith 2010 335/357 166/173 4/53 2/54 Vodermaier 2009 1.8 % 2.04 [ 0.39, 10.66 ] 100.0 % 1.37 [ 1.08, 1.73 ] 1419 1111 Subtotal (95% CI) Total events: 869 (Decision Aid), 526 (Usual Care) 2 2 2 = 0.12; Chi Heterogeneity: Tau =90% = 123.58, df = 12 (P<0.00001); I Test for overall effect: Z = 2.59 (P = 0.0096) 3 Shared decision making Auvinen 2004 25/103 17/100 5.3 % 1.43 [ 0.82, 2.48 ] 0.67 [ 0.36, 1.24 ] 4.5 % Davison 1997 15/30 10/30 22/43 27/43 Dolan 2002 1.23 [ 0.85, 1.78 ] 8.7 % 5.6 % 0.56 [ 0.33, 0.96 ] Kasper 2008 19/134 26/103 71/196 27/75 Krist 2007 9.1 % 1.01 [ 0.71, 1.43 ] 8.6 % 0.89 [ 0.61, 1.30 ] 43/146 Man-Son-Hing 1999 36/137 Mathers 2012 28/77 25/92 7.0 % 0.75 [ 0.48, 1.17 ] 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid ( Continued . . . ) 249 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

253 ( . . . Continued ) Usual Care Weight R isk Ratio Decision Aid Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 42/86 38/94 9.8 % 1.21 [ 0.87, 1.68 ] Morgan 2000 12.5 % 0.68 [ 0.54, 0.87 ] 34/57 42/48 Murray 2001a 40/94 36/95 Murray 2001b 1.12 [ 0.79, 1.59 ] 9.2 % 6.1 % 0.79 [ 0.48, 1.30 ] 19/37 Sawka 2012 15/37 Smith 2010 5/173 17/357 2.1 % 1.65 [ 0.62, 4.39 ] 11.5 % 0.99 [ 0.76, 1.30 ] Vodermaier 2009 35/53 36/54 1419 1075 Subtotal (95% CI) 0.94 [ 0.80, 1.09 ] 100.0 % Total events: 396 (Decision Aid), 354 (Usual Care) 2 2 2 Heterogeneity: Tau = 22.37, df = 12 (P = 0.03); I = 0.03; Chi =46% Test for overall effect: Z = 0.86 (P = 0.39) 0.1 0.2 0.5 1 2 5 10 Favours Usual Care Favours Decision Aid 250 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

254 Analysis 6.1. Comparison 6 Proportion undecided, Outcome 1 Proportion undecided - all studies. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 6 Proportion undecided Outcome: 1 Proportion undecided - all studies Decision Aid Weight R isk Ratio Usual Care Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 13/90 Nassar 2007 0.07 [ 0.01, 0.53 ] 1/98 0.9 % 0/44 Jibaja-Weiss 2011 4/39 0.10 [ 0.01, 1.78 ] 0.5 % 0.12 [ 0.02, 0.92 ] 0.9 % Man-Son-Hing 1999 1/139 9/148 Miller 2011 22/132 72/132 6.5 % 0.31 [ 0.20, 0.46 ] 3.9 % 0.39 [ 0.18, 0.86 ] Protheroe 2007 18/56 7/56 8/184 20/179 Vuorma 2003 3.8 % 0.39 [ 0.18, 0.86 ] 0.43 [ 0.19, 1.01 ] 3.5 % Chambers 2012 6/48 17/59 Mathieu 2010 21/117 82/209 0.46 [ 0.30, 0.70 ] 6.5 % 5.4 % 0.48 [ 0.28, 0.84 ] Mathieu 2007 17/349 36/356 4/37 8/37 Sawka 2012 0.50 [ 0.16, 1.52 ] 2.5 % 5.0 % 0.53 [ 0.29, 0.97 ] Murray 2001b 13/94 25/96 Shorten 2005 14/99 20/93 4.9 % 0.66 [ 0.35, 1.22 ] 0.67 [ 0.48, 0.94 ] 7.2 % 33/100 Schwartz 2009a 56/114 67/102 171/383 Fagerlin 2011 8.3 % 0.68 [ 0.57, 0.81 ] 3.3 % 0.75 [ 0.30, 1.85 ] Mathers 2012 8/95 9/80 16/44 18/41 Legare 2008a 0.83 [ 0.49, 1.40 ] 5.7 % 8.3 % 0.89 [ 0.75, 1.07 ] 45/60 Bozic 2013 52/62 13/70 16/78 Vandemheen 2009 0.91 [ 0.47, 1.75 ] 4.7 % 4.2 % 1.04 [ 0.50, 2.15 ] Berry 2013 14/120 12/107 Allen 2010 34/291 36/334 6.3 % 1.08 [ 0.70, 1.69 ] 3.4 % 1.28 [ 0.54, 3.07 ] 10/75 8/77 Arterburn 2011 Stacey 2014a 20/66 9/66 2.22 [ 1.09, 4.51 ] 4.3 % 100.0 % 0.64 [ 0.52, 0.79 ] 2701 2555 Total (95% CI) Total events: 478 (Decision Aid), 607 (Usual Care) 2 2 2 Heterogeneity: Tau = 67.06, df = 21 (P<0.00001); I = 0.13; Chi =69% Test for overall effect: Z = 4.16 (P = 0.000031) Test for subgroup differences: Not applicable 0.1 0.2 0.5 1 2 5 10 Favours decision aid Favours usual care 251 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

255 Analysis 7.1. Comparison 7 Satisfaction, Outcome 1 Satisfa ction with the choice - all studies. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 7 Satisfaction Outcome: 1 Satisfaction with the choice - all studies Mean Mean Decision aid Usual care Study or subgroup Difference Difference N Mean(SD) IV,Random,95% CI IV,Random,95% CI N Mean(SD) 75.89 (17.2) 117 73.9 (18) 1.99 [ -2.65, 6.63 ] 104 Barry 1997 -4.60 [ -12.42, 3.22 ] 73.1 (20.9) 48 77.7 (20.5) Bernstein 1998 61 Chabrera 2015 95.7 (6.89) 61 79.3 (10.3) 61 16.40 [ 13.29, 19.51 ] 1.30 [ -2.57, 5.17 ] Hanson 2011 127 83.5 (16.19) 84.8 (15.19) 126 43 93.5 (12) 38 92.5 (15) Jibaja-Weiss 2011 1.00 [ -4.97, 6.97 ] 12.00 [ 3.18, 20.82 ] Laupacis 2006 56 61 (25.4) 54 73 (21.7) 212 209 80 (15) Montgomery 2007 85 (15) 5.00 [ 2.13, 7.87 ] 2.50 [ -5.10, 10.10 ] 80 (26) Morgan 2000 77.5 (26) 86 94 Nassar 2007 86 87.9 (12.5) 84 84.2 (15) 3.70 [ -0.46, 7.86 ] 2.50 [ -7.20, 12.20 ] Ozanne 2007 15 80 (12.25) 15 82.5 (14.75) 357 80.25 (11) 173 80.25 (10.75) Smith 2010 0.0 [ -1.97, 1.97 ] -100 -50 50 100 0 Favours control Favours decision aid 252 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

256 Analysis 7.2. Comparison 7 Satisfaction, Outcome 2 Satisfa ction with the choice - in consultation. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 7 Satisfaction Outcome: 2 Satisfaction with the choice - in consultation Mean Mean Decision aid Difference Difference Usual care Study or subgroup N Mean(SD) IV,Random,95% CI IV,Random,95% CI N Mean(SD) 82.5 (14.75) 15 80 (12.25) 2.50 [ -7.20, 12.20 ] 15 Ozanne 2007 -100 -50 0 50 100 Favours decision aid Favours control ction with the choice - in preparation for Analysis 7.3. Comparison 7 Satisfaction, Outcome 3 Satisfa consultation. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 7 Satisfaction Outcome: 3 Satisfaction with the choice - in preparation for c onsultation Mean Mean Difference Difference Study or subgroup Decision aid Usual care Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI N Barry 1997 75.89 (17.2) 117 73.9 (18) 1.99 [ -2.65, 6.63 ] 104 Bernstein 1998 61 73.1 (20.9) 48 77.7 (20.5) -4.60 [ -12.42, 3.22 ] 16.40 [ 13.29, 19.51 ] 95.7 (6.89) 61 79.3 (10.3) Chabrera 2015 61 126 127 16.5 (16.19) Hanson 2011 15.25 (15.19) -1.25 [ -5.12, 2.62 ] 1.00 [ -4.97, 6.97 ] 93.5 (12) 38 Jibaja-Weiss 2011 43 92.5 (15) Laupacis 2006 54 73 (21.7) 56 61 (25.4) 12.00 [ 3.18, 20.82 ] 5.00 [ 2.13, 7.87 ] 212 85 (15) 80 (15) Montgomery 2007 209 86 77.5 (26) 94 Morgan 2000 80 (26) 2.50 [ -5.10, 10.10 ] 3.70 [ -0.46, 7.86 ] Nassar 2007 87.9 (12.5) 84 84.2 (15) 86 Smith 2010 357 80.25 (11) 173 80.25 (10.75) 0.0 [ -1.97, 1.97 ] 0 50 100 -100 -50 Favours control Favours decision aid 253 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

257 Analysis 7.4. Comparison 7 Satisfaction, Outcome 4 Satisfa ction with the decision making process - all studies. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 7 Satisfaction l studies Outcome: 4 Satisfaction with the decision making process - al Mean Mean Difference Difference Study or subgroup Decision Aid Usual Care Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI N 104 Barry 1997 117 71.07 (18.4) 5.31 [ 0.71, 9.91 ] 76.38 (16.5) 61 76.5 (17.6) 48 Bernstein 1998 73.1 (20.6) -3.40 [ -10.58, 3.78 ] 3.30 [ -1.09, 7.69 ] Bozic 2013 94.4 (10) 62 91.1 (14.4) 60 Jibaja-Weiss 2011 43 94 (17) 38 92.5 (17) 1.50 [ -5.92, 8.92 ] 1.00 [ -4.53, 6.53 ] 74 (16) 80 73 (19) 74 Knops 2014 5.10 [ -2.31, 12.51 ] 91.4 (12.5) 31 86.3 (18.6) Kupke 2013 50 146 83.75 (14.79) 138 84.75 (13.04) Man-Son-Hing 1999 -1.00 [ -4.24, 2.24 ] 2.00 [ -3.81, 7.81 ] 86 72 (19.88) 94 70 (19.88) Morgan 2000 Schroy 2011 84.17 (10.33) 217 77.83 (13.17) 214 6.34 [ 4.11, 8.57 ] -10 -5 0 5 10 Favours simple DA Favours detailed DA 254 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

258 Analysis 7.5. Comparison 7 Satisfaction, Outcome 5 Satisfa ction with the decision making process - in consultation. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 7 Satisfaction consultation Outcome: 5 Satisfaction with the decision making process - in Mean Mean Decision Aid Usual Care Difference Difference Study or subgroup N N IV,Random,95% CI IV,Random,95% CI Mean(SD) Mean(SD) 50 91.4 (12.5) 31 86.3 (18.6) 5.10 [ -2.31, 12.51 ] Kupke 2013 -5 -10 5 10 0 Favours simple DA Favours detailed DA ction with the decision making process - in Analysis 7.6. Comparison 7 Satisfaction, Outcome 6 Satisfa preparation for consultation. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 7 Satisfaction Outcome: 6 Satisfaction with the decision making process - in preparation for consultation Mean Mean Decision Aid Usual Care Study or subgroup Difference Difference N N Mean(SD) IV,Random,95% CI IV,Random,95% CI Mean(SD) Barry 1997 104 76.38 (16.5) 117 71.07 (18.4) 5.31 [ 0.71, 9.91 ] -3.40 [ -10.58, 3.78 ] Bernstein 1998 76.5 (17.6) 61 73.1 (20.6) 48 60 62 91.1 (14.4) Bozic 2013 94.4 (10) 3.30 [ -1.09, 7.69 ] 1.50 [ -5.92, 8.92 ] 94 (17) 38 92.5 (17) Jibaja-Weiss 2011 43 74 74 (16) 80 73 (19) Knops 2014 1.00 [ -4.53, 6.53 ] -1.00 [ -4.24, 2.24 ] 146 83.75 (14.79) 138 84.75 (13.04) Man-Son-Hing 1999 Morgan 2000 86 94 70 (19.88) 72 (19.88) 2.00 [ -3.81, 7.81 ] 6.34 [ 4.11, 8.57 ] Schroy 2011 84.17 (10.33) 217 77.83 (13.17) 214 -10 -5 0 5 10 Favours simple DA Favours detailed DA 255 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

259 Analysis 8.1. Comparison 8 Choice, Outcome 1 Choice: surger y over conservative option. Review: Decision aids for people facing health treatment or s creening decisions Comparison: 8 Choice Outcome: 1 Choice: surgery over conservative option Risk Ratio Weight R Study or subgroup isk Ratio Decision Aid Usual Care M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 Per-protocol analysis 30/72 Arterburn 2011 43/73 6.8 % 0.71 [ 0.51, 0.99 ] 0.64 [ 0.54, 0.76 ] 8.9 % 91/100 Auvinen 2004 60/103 Barry 1997 8/103 16/116 0.56 [ 0.25, 1.26 ] 2.6 % 0.70 [ 0.48, 1.03 ] 6.1 % 25/61 28/48 Bernstein 1998 49/107 42/120 Berry 2013 7.0 % 0.76 [ 0.56, 1.05 ] 7.8 % 0.90 [ 0.70, 1.16 ] Bozic 2013 38/61 43/62 18/44 20/39 Jibaja-Weiss 2011 0.80 [ 0.50, 1.27 ] 5.1 % 8.1 % 0.78 [ 0.62, 0.99 ] 82/253 101/244 Kennedy 2002 39/91 36/87 Knops 2014 1.04 [ 0.73, 1.46 ] 6.6 % 7.1 % 1.18 [ 0.87, 1.60 ] 39/81 Lam 2013 38/67 Morgan 2000 45/86 63/95 0.79 [ 0.62, 1.01 ] 7.9 % 0.5 % 5.33 [ 0.67, 42.73 ] 1/48 6/54 Murray 2001a Protheroe 2007 3/56 7/56 1.2 % 2.33 [ 0.64, 8.57 ] 3.8 % 2.14 [ 1.16, 3.95 ] 18/64 15/114 Schwartz 2009a 55/69 Stacey 2014a 48/68 8.6 % 1.13 [ 0.93, 1.37 ] 0.8 % 0.42 [ 0.09, 2.04 ] 2/39 Vodermaier 2009 5/41 Vuorma 2003 98/184 88/179 8.5 % 1.08 [ 0.89, 1.32 ] 2.4 % 0.26 [ 0.11, 0.61 ] Whelan 2004 26/107 6/94 Subtotal (95% CI) 1621 1665 0.87 [ 0.75, 1.01 ] 100.0 % Total events: 617 (Decision Aid), 715 (Usual Care) 2 2 2 Heterogeneity: Tau = 55.99, df = 17 (P<0.00001); I = 0.06; Chi =70% 0.1 0.2 0.5 1 2 5 10 Favours decision aid Favours usual care ( Continued . . . ) 256 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

260 ( . . . ) Continued Decision Aid Study or subgroup Weight R isk Ratio Usual Care Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Test for overall effect: Z = 1.79 (P = 0.074) 2 Intention-to-treat analysis 0.72 [ 0.51, 1.01 ] 6.8 % 43/77 30/75 Arterburn 2011 91/106 60/104 Auvinen 2004 0.67 [ 0.56, 0.81 ] 9.4 % 2.4 % 0.59 [ 0.26, 1.33 ] 16/123 Barry 1997 8/104 28/53 Bernstein 1998 25/65 5.9 % 0.73 [ 0.49, 1.09 ] 0.73 [ 0.51, 1.07 ] 6.3 % 49/228 Berry 2013 42/266 38/61 43/62 Bozic 2013 0.90 [ 0.70, 1.16 ] 8.2 % 4.7 % 0.86 [ 0.52, 1.43 ] 20/49 Jibaja-Weiss 2011 18/51 101/298 82/300 Kennedy 2002 0.81 [ 0.63, 1.03 ] 8.4 % 6.7 % 1.04 [ 0.73, 1.46 ] Knops 2014 39/91 36/87 38/67 39/81 Lam 2013 7.3 % 1.18 [ 0.87, 1.60 ] 7.7 % 0.71 [ 0.54, 0.95 ] 45/120 Morgan 2000 63/120 Murray 2001a 6/57 1/55 5.79 [ 0.72, 46.54 ] 0.5 % 1.1 % 2.33 [ 0.63, 8.67 ] 7/72 3/72 Protheroe 2007 Schwartz 2009a 18/100 15/114 3.5 % 1.37 [ 0.73, 2.57 ] 9.0 % 1.15 [ 0.93, 1.41 ] Stacey 2014a 55/71 48/71 2/39 5/41 Vodermaier 2009 0.8 % 0.42 [ 0.09, 2.04 ] 1.08 [ 0.89, 1.32 ] 9.1 % 88/179 Vuorma 2003 98/184 26/107 Whelan 2004 6/94 2.3 % 0.26 [ 0.11, 0.61 ] 100.0 % 0.86 [ 0.75, 1.00 ] 1923 1921 Subtotal (95% CI) Total events: 617 (Decision Aid), 715 (Usual Care) 2 2 2 Heterogeneity: Tau = 0.05; Chi =63% = 46.00, df = 17 (P = 0.00017); I Test for overall effect: Z = 2.00 (P = 0.046) 3 Per-protocol analysis without prophylactic mastectomy Arterburn 2011 30/72 43/73 7.0 % 0.71 [ 0.51, 0.99 ] 0.64 [ 0.54, 0.76 ] 9.5 % 91/100 60/103 Auvinen 2004 8/103 16/116 Barry 1997 2.5 % 0.56 [ 0.25, 1.26 ] 0.70 [ 0.48, 1.03 ] 6.2 % 28/48 Bernstein 1998 25/61 42/120 49/107 Berry 2013 7.2 % 0.76 [ 0.56, 1.05 ] 8.2 % 0.90 [ 0.70, 1.16 ] Bozic 2013 38/61 43/62 Jibaja-Weiss 2011 18/44 20/39 5.1 % 0.80 [ 0.50, 1.27 ] 8.6 % 0.78 [ 0.62, 0.99 ] Kennedy 2002 82/253 101/244 0.1 0.2 0.5 1 2 5 10 Favours decision aid Favours usual care ( Continued . . . ) 257 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

261 ( . . . ) Continued Decision Aid Usual Care Weight R isk Ratio Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 39/91 36/87 6.8 % 1.04 [ 0.73, 1.46 ] Knops 2014 7.4 % 1.18 [ 0.87, 1.60 ] 38/67 Lam 2013 39/81 45/86 Morgan 2000 63/95 8.4 % 0.79 [ 0.62, 1.01 ] 0.5 % 5.33 [ 0.67, 42.73 ] 6/54 Murray 2001a 1/48 Protheroe 2007 7/56 3/56 1.1 % 2.33 [ 0.64, 8.57 ] 1.13 [ 0.93, 1.37 ] 9.2 % Stacey 2014a 48/68 55/69 5/41 2/39 Vodermaier 2009 0.42 [ 0.09, 2.04 ] 0.8 % 9.1 % 1.08 [ 0.89, 1.32 ] Vuorma 2003 98/184 88/179 6/94 26/107 Whelan 2004 2.3 % 0.26 [ 0.11, 0.61 ] 100.0 % 0.84 [ 0.73, 0.97 ] 1551 Subtotal (95% CI) 1557 Total events: 599 (Decision Aid), 700 (Usual Care) 2 2 2 = 0.05; Chi = 47.70, df = 16 (P = 0.00005); I =66% Heterogeneity: Tau Test for overall effect: Z = 2.32 (P = 0.020) 0.1 0.2 0.5 1 2 5 10 Favours decision aid Favours usual care 258 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

262 Analysis 8.2. Comparison 8 Choice, Outcome 2 Choice for scre ening. creening decisions Review: Decision aids for people facing health treatment or s Comparison: 8 Choice Outcome: 2 Choice for screening Experimental Weight Ri sk Ratio Control Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI 1 PSA screening Allen 2010 225/291 264/334 0.98 [ 0.90, 1.06 ] 19.0 % 11/123 4/127 Evans 2010 0.35 [ 0.12, 1.08 ] 0.8 % 4.0 % 1.10 [ 0.69, 1.77 ] 27/106 Gattellari 2003 25/108 Gattellari 2005 37/131 42/136 5.8 % 0.91 [ 0.63, 1.33 ] 17.3 % 0.97 [ 0.87, 1.09 ] Krist 2007 64/75 163/196 97/215 99/216 Lepore 2012 11.7 % 0.98 [ 0.80, 1.21 ] 0.90 [ 0.70, 1.16 ] 9.5 % Partin 2004 83/308 87/290 Volk 1999 48/78 64/80 0.77 [ 0.63, 0.95 ] 11.7 % 11.8 % 0.88 [ 0.72, 1.08 ] Watson 2006 119/465 149/512 40/103 Wolf 1996 68/102 0.58 [ 0.44, 0.77 ] 8.5 % 100.0 % 0.88 [ 0.80, 0.98 ] Subtotal (95% CI) 1976 2020 Total events: 843 (Experimental), 873 (Control) 2 2 2 = 0.01; Chi Heterogeneity: Tau = 21.43, df = 9 (P = 0.01); I =58% Test for overall effect: Z = 2.30 (P = 0.021) 2 Colorectal cancer screening Dolan 2002 7/43 2/45 1.0 % 0.27 [ 0.06, 1.24 ] 70/226 Lewis 2010 71/207 10.6 % 1.11 [ 0.84, 1.45 ] 1.39 [ 0.80, 2.42 ] 5.3 % 18/132 Miller 2011 25/132 Pignone 2000 46/124 28/124 7.8 % 1.64 [ 1.10, 2.45 ] 1.70 [ 1.24, 2.32 ] 9.6 % 33/87 56/87 Ruffin 2007 Schroy 2011 116/269 96/276 12.0 % 1.24 [ 1.00, 1.53 ] 14.0 % 0.79 [ 0.70, 0.89 ] 130/173 Smith 2010 211/357 Steckelberg 2011 141/785 134/792 1.06 [ 0.86, 1.32 ] 11.9 % 0.96 [ 0.89, 1.05 ] 14.7 % 124/137 117/134 Trevena 2008 173/266 79/133 Wolf 2000 13.1 % 1.09 [ 0.93, 1.29 ] 1.12 [ 0.95, 1.31 ] 100.0 % 2123 Subtotal (95% CI) 2406 0.2 0.5 1 2 5 Reduces preference Increase preference ( Continued . . . ) 259 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

263 ( . . . ) Continued Experimental Control Weight Ri sk Ratio Study or subgroup Risk Ratio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Total events: 958 (Experimental), 719 (Control) 2 2 2 = 48.11, df = 9 (P<0.00001); I = 0.04; Chi Heterogeneity: Tau =81% Test for overall effect: Z = 1.37 (P = 0.17) 3 Breast cancer genetic testing 21.9 % 1.18 [ 0.67, 2.06 ] Green 2001 16/42 13/29 46.0 % 1.14 [ 0.93, 1.40 ] 87/164 Lerman 1997 74/122 35/191 49/190 Schwartz 2001 0.71 [ 0.48, 1.04 ] 32.1 % Subtotal (95% CI) 342 396 100.0 % 0.99 [ 0.71, 1.38 ] Total events: 122 (Experimental), 152 (Control) 2 2 2 = 5.15, df = 2 (P = 0.08); I = 0.05; Chi =61% Heterogeneity: Tau Test for overall effect: Z = 0.07 (P = 0.94) 4 Prenatal diagnostic testing 92/184 111/206 Bjorklund 2012 0.93 [ 0.77, 1.12 ] 21.8 % 78.2 % 1.01 [ 0.92, 1.12 ] Kuppermann 2014 238/353 244/357 Subtotal (95% CI) 541 559 100.0 % 0.99 [ 0.91, 1.09 ] Total events: 336 (Experimental), 349 (Control) 2 2 2 = 0.0; Chi = 0.66, df = 1 (P = 0.42); I =0.0% Heterogeneity: Tau Test for overall effect: Z = 0.12 (P = 0.90) 0.2 0.5 1 2 5 Reduces preference Increase preference 260 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

264 Analysis 8.3. Comparison 8 Choice, Outcome 3 Choice: diabet es medication (uptake new medication). creening decisions Review: Decision aids for people facing health treatment or s Comparison: 8 Choice Outcome: 3 Choice: diabetes medication (uptake new medicat ion) Risk Ratio Weight R isk Ratio Usual care Decision Aid Study or subgroup M- M- H,Random,95% H,Random,95% n/N n/N CI CI 9/80 3/70 12.1 % Mann D 2010 2.63 [ 0.74, 9.32 ] 34.6 % 1.60 [ 0.76, 3.39 ] Mathers 2012 9/78 17/92 16/48 Mullan 2009 8/37 36.2 % 1.54 [ 0.74, 3.21 ] 7/23 4/19 Weymiller 2007 17.1 % 1.45 [ 0.50, 4.20 ] 100.0 % 1.65 [ 1.06, 2.56 ] Total (95% CI) 243 204 Total events: 49 (Decision Aid), 24 (Usual care) 2 2 2 = 0.0; Chi = 0.62, df = 3 (P = 0.89); I Heterogeneity: Tau =0.0% Test for overall effect: Z = 2.22 (P = 0.026) Test for subgroup differences: Not applicable 0.01 0.1 1 10 100 Reduce preference Increase preference A D D I T I O N A L T A B L E S Table 1. Decision aids evaluated in the trials Topic Availability Source Contact Information Study Prostate cancer screening No Allen, Allen 2010 Requested access Center for Community- Based Research, Dana- Farber Cancer Institute, Boston, MA, USA, 2010 Arterburn 2011 Yes Informed Medical Deci- Bariatric surgery sions Foundation, MA, informedmedicaldecisions.org/ USA, 2010 imdf ̇decision ̇aid/ making-decisions-about- weight-loss-surgery/ Auvinen 2004 Prostate cancer Included in publication Yes Auvinen, Helsinki, Fin- land, 1993 treatment 261 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

265 Table 1. Decision aids evaluated in the trials ( Continued) Benign prostate disease Informed Medical Deci- Barry 1997 Yes treatment sions Foundation, MA, informedmedicaldecisions.org/ USA, 2001 imdf ̇decision ̇aid/ treatment-options- for-benign-prostatic- hyperplasia/ Yes Bekker, Leeds, UK, 2003 Bekker 2004 Prenatal screening Included in publication Ischaemic heart disease Yes Informed Medical Deci- Bernstein 1998 treatment sions Foundation, MA, informedmedicaldecisions.org/ USA, 2002 imdf ̇decision ̇aid/ treatment-choices-for- carotid-artery-disease/ Prostate cancer Berry 2013 No Berry, Phyllis F. Cantor [email protected] donna Center, MA, USA, 2011 harvard.edu treatment Antenatal Down syn- Yes Bjorklund 2012 vimeo.com/34600615/ Södersjukhuset, Depart- drome screening ment of Obstetrics and Gynecology, Stockholm, Sweden Bozic 2013 Osteoarthritis of the www.healthdialog.com No Informed Medical De- cisions Foundation and knee or hip Health Dialog; USA Brazell 2014 Pelvic Organ Prolapse Yes Healthwise, USA decisionaid.ohri.ca Prostate cancer C Chabrera. School of No Chabrera 2015 [email protected] Health Sciences, Depart- cat treatment ment of Nursing. Mataro, Spain Chambers 2012 Healthcare personnel’s Yes A McCarthy. Ottawa In- decisionaid.ohri.ca/ fluenza Decision decaids.html#oida influenza immunization Aid Planning Group, CA, 2008 Clancy 1988 Hepatitis B Vaccine No Clancy, Richmond VA, - USA, 1983 Davison 1997 Prostate cancer - No Davison, Manitoba CA, 1992-1996 treatment De Achaval 2012 Total Yes Informed Medical Deci- knee arthroplasty treat- sions Foundation, MA, informedmedicaldecisions.org/ ment USA imdf ̇decision ̇aid/ 262 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

266 Table 1. Decision aids evaluated in the trials ( Continued) treatment-choices-for- knee-osteoarthritis/ No Dolan, Rochester NY, Dolan 2002 - Colon cancer screening USA, 1999 Prostate cancer screening Elwyn, Cardiff, UK www.prosdex.com Evans 2010 Yes Yes Fagerlin, Ann Arbor, MI, Fagerlin 2011 - Breast cancer prevention USA Osteoarthritis knee treat- No Fraenkel, New Haven Author said DA never Fraenkel 2007 fully developed, all info in ment CT, USA paper Fraenkel 2012 No Vet- Atrial fibrillation Obtained from author erans Affairs Connecticut [email protected] Healthcare System, USA Frosch 2008a Prostate cancer screening No Frosch, Los Angeles, USA Screenshots from author Prostate cancer screening Yes Gattellari 2003 included in publication Gatellari, Sydney, AU, 2003 Gattellari 2005 Prostate cancer screening Yes Gatellari, Sydney, AU, Included in publication 2003 Green 2001 Breast cancer genetic 1-800-757-4868 Yes Green, Hershey PA, USA, 2000 testing [email protected] Schizophrenia treatment Hamann, Munich, GER Emailed by author (in Hamann 2006 Yes German) Feeding options in ad- decisionaid.ohri.ca/ Hanson 2011 Yes Mitchell, Tetroe, O’Connor; 2001 vanced decaids.html#feedingtube dementia (updated 2008) Heller 2008 Breast reconstruction Yes University of Texas MD Disc mailed Anderson Cancer Center, Houston TX, USA, 2003 Stress testing for chest Included in publication Yes Hess, Rochester, MN, Hess 2012 USA, 2012 pain Jibaja-Weiss 2011 Breast cancer treatment Yes Jibaja-Weiss, Baylor Col- www.bcm.edu/ lege of Medicine, 2010 patchworkoflife Included in publication Johnson 2006 Yes Johnson, Chicago, USA, Endodontic treatment 2004 263 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

267 Table 1. Decision aids evaluated in the trials ( Continued) Multiple sclerosis No - Kasper 2008 Jürgen Kasper Kennedy/Coulter, - Abnormal uterine bleed- Kennedy 2002 No ing treatment London UK, 1996 Knops 2014 Asymptomatic Abdom- Yes Amsterdam, The Nether- www.keuzehulp.info/ lands amc/AAA/landing-page inal Aortic Aneurysm treatment Prostate cancer screening Yes Krist, Fairfax VA, USA Krist 2007 www.familymedicine.vcu.edu/ research/misc/psa/ index.html Kupke 2013 Yes University of Cologne, [email protected] Dental - posterior tooth decay Cologne, Germany Prenatal screening No Kuppermann, San Fran- Kuppermann 2014 Interactive web-based de- cision aid cisco CA, USA Lam 2013 Breast cancer treatment Yes Kwong Wah Hospital, Obtained from author. Hong Kong, China [email protected] Langston 2010 Contraceptive method www.who.int/ Yes World Health Organiza- tion, 2005 choice reproductivehealth/ publications/ family ̇planning/ 9241593229index/en/ index.html Laupacis 2006 Pre-operative autologous No Laupacis, Ottawa, CA, Decisionaid.ohri.ca/ decaids-archive.html blood donation 2001 LeBlanc 2015 Treatment for osteoporo- Yes Mayo Clinic - sis Legare 2008a No Legare, Quebec City, CA, - Natural health products 2006 Legare 2011 Use of antibiotics for Legare, Quebec City, CA, Yes acute 2007 www.decision.chaire.fmed.ulaval.ca/ respiratory infections index.php?id=192&L=2 Legare 2012 Antibiotics for acute res- Yes Legare, Quebec City, CA piratory infections www.decision.chaire.fmed.ulaval.ca/ index.php? Leighl 2011 Advanced colorectal can- [email protected] Yes Princess Margaret Hospi- tal, Toronto, 2011 cer chemotherapy ca 264 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

268 Table 1. Decision aids evaluated in the trials ( Continued) Prostate cancer screening Yes Lepore 2012 Obtained from author Sally Weinrich University [email protected] of Louisville, USA Lerman 1997 No Lerman/Schwartz, Wash- - Breast cancer genetic testing ington DC, USA, 1997 Lewis 2010 Yes Lewis, Colorectal cancer University of North Car- screening decisionsupport.unc.edu/ olina, Chapel Hill, NC, CHOICE6/ USA, 2010 Depression treatment Yes Loh, Freiburg, GER Emailed to us by author - Loh 2007 in German Man-Son-Hing 1999 Atrial fibrillation treat- No McAlister/Laupacis, Ot- decisionaid.ohri.ca/ tawa CA, 2000 ment decaids-archive.html Mann D 2010 Diabetes treatment - Yes Montori, Rochester MN, mayoresearch.mayo.edu/ statins mayo/research/ker ̇unit/ USA form.cfm Mann E 2010 Diabetes Yes Marteau, King’s College Additional file 2 of publi- screening London, London, Eng- cation land, 2010 Marteau 2010 Diabetes Provided by author, same Yes Marteau, King’s College Mann E 2010 London, London, Eng- DA as screening land, 2010 Mammography Mathieu, Sydney, AU DA emailed by author Mathieu 2007 Yes Diabetes treatment Yes The Univer- Mathers 2012 Obtained from author [email protected] sity of Sheffield, Sheffield, UK, 2008 Mammography Yes Mathieu 2010 www.psych.usyd.edu.au/ Mathieu, University of Sydney, AUS, 2010 cemped/ com ̇decision ̇aids.shtml McAlister 2005 Atrial fibrillation treat- decisionaid.ohri.ca/ No McAlister/Laupacis, Ot- tawa CAN, 2000 decaids-archive.html ment McBride 2002 Hormone replacement Yes, update in progress Sigler/Bastien, Durham [email protected] therapy NC, USA, 1998 Screening after mildly McCaffery 2010 Yes Screening & test evalua- [email protected] tion program, School of edu.au abnormal pap smear public health, University of Sydney 2007 265 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

269 Table 1. Decision aids evaluated in the trials ( Continued) BRCA1/BRCA2 gene Miller, Fox Chase PA, - Miller 2005 No USA testing University of North Car- Colorectal Miller 2011 Yes cancer screening olina, Chapel Hill, NC, intmedweb.wakehealth.edu/ USA, 2007 choice/choice.html (no longer available) Hypertension treatment No Montgomery, UK, 2000 Montgomery 2003 - Montgomery 2007 Birthing options after Yes Montgomery, Bristol, UK, last update 2004 caesarean www.computing.dundee.ac.uk/ acstaff/cjones/diamond/ Information.html Osteoporosis treatment Yes Montori 2011 Montori, Mayo Founda- tion for Medical Educa- shareddecisions.mayoclinic.org/ tion and Research, 2007 decision-aids-for- diabetes/other-decision- aids/ Morgan 2000 Ischaemic heart disease Informed Medical Deci- Yes treatment sions Foundation, MA, informedmedicaldecisions.org/ USA, 2002 imdf ̇decision ̇aid/ treatment-choices-for- carotid-artery-disease/ PTSD treatment Yes Michael E DeBakey Vet- Obtained from author Mott 2014 erans Affairs Medical [email protected] Center, Houston, USA Mullan 2009 Diabetes treatment Yes Montori or Mayo Foun- Included in publication dation(?) Rochester MN, USA, Murray 2001a Benign prostate disease Informed Medical Deci- Yes treatment sions Foundation, MA, informedmedicaldecisions.org/ USA, 2001 imdf ̇decision ̇aid/ treatment-options- for-benign-prostatic- hyperplasia/ Murray 2001b Hormone replacement Informed Medical Deci- No, update in progress therapy sions Foundation, MA, informedmedicaldecisions.org/ USA imdf ̇decision ̇aid/ treatment-choices-for- managing-menopause/ 266 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

270 Table 1. Decision aids evaluated in the trials ( Continued) Prenatal screening Nagle 2008 www.mcri.edu.au/ Yes Nagle, Victoria, AU Downloads/ PrenatalTestingDecisionAid.pdf sydney.edu.au/medicine/ Nassar 2007 Nassar, West Perth WA, Birth breech presenta- Yes AU public-health/shdg/ tion resources/ decision ̇aids.php Osteoporosis treatment No Cranney, Ottawa CA, decisionaid.ohri.ca/ Oakley 2006 2002 decaids-archive.html Breast cancer prevention - Ozanne, Boston MA, Ozanne 2007 No USA Yes Informed Medical Deci- Partin 2004 Prostate cancer screening informedmedicaldecisions.org/ sions Foundation, MA, imdf ̇decision ̇aid/ USA, 2001 deciding-if-the-psa-test- is-right-for-you/ Pignone 2000 Colon cancer screening Yes Pignone, Chapel Hill www.med.unc.edu/ medicine/edusrc/ NC, USA, 1999 colon.htm Protheroe 2007 Protheroe, Manchester, Menorrhagia treatment Computer- No UK ized decision aid, Clinical Guidance Tree - no longer in existence, author sent chapter in thesis Prostate cancer screening No Centers for Disease Con- Rubel 2010 No longer available trol and Prevention (CDC), USA, 2010 Ruffin 2007 Colorectal cancer screen- Yes Regents of the Univer- colorectalweb.org ing sity of Michigan (copy- right info), Ann Arbor MI, USA, 2006 Sawka 2012 - University Health Net- Adjuvant radioactive io- No work, Toronto, Canada, dine treatment for pa- tients with early-stage 2009 papillary thyroid cancer Schroy 2011 Colorectal [email protected] Yes Schroy III, Boston, USA cancer screening 267 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

271 Table 1. Decision aids evaluated in the trials ( Continued) Coronary angiogram ac- Schwalm, Hamilton, www.phri.ca/workfiles/ Schwalm 2012 Yes studies/presentations/ ON, Canada, 2009 cess site PtDA%20Vascular%20Access%2023- − May 2012.pdf - No Schwartz/Lerman, Wash- Schwartz 2001 Breast cancer genetic ington DC, USA, 1997 testing Schwartz 2009a BRCA mutation prophy- No Schwartz, Washington - lactic surgery DC, USA Cardiovascular preven- Sheridan 2006 Yes Sheridan, Chapel Hill, www.med- NC, USA tion decisions.com/cvtool/ Sheridan 2011 Coronary heart Yes Sheridan, University of www.med- decisions.com/h2hv3/ disease prevention North Carolina at Chapel Hill, Division of General Internal Medicine, North Carolina, USA, 2011 Shorten 2005 Birthing options after Yes (updated 2006) Shorten, Wollongong, [email protected] or AU, 2000 www.capersbookstore.com.au/ previous caesarean product.asp?id=301 Shourie 2013 Yes University of Leeds, UK Measles mumps and www.leedsmmr.co.uk rubella vaccination & NSIRS Australia Smith 2010 Bowel Yes Smith, Sydney, AU 2008 sydney.edu.au/medicine/ public-health/shdg/ cancer screening resources/ decision ̇aids.php www.healthdialog.com Osteoarthritis of the hip No Informed Medical De- Stacey 2014a and knee cisions Foundation and Health Dialog; USA Steckelberg 2011 Colorectal cancer screen- Yes Steckelberg, Hamburg, - Germany ing Taylor 2006 Yes George- Obtained from author Prostate cancer screening town University Medical [email protected] Center, Washington DC, edu USA, 2000 Thomson 2007 Atrial fibrillation treat- Disc sent by mail Yes Thomson, Newcastle Upon Thyne, UK ment 268 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

272 Table 1. Decision aids evaluated in the trials ( Continued) Colorectal cancer screen Yes sydney.edu.au/medicine/ Trevena 2008 Trevena, Sydney, AU public-health/shdg/ resources/ decision ̇aids.php Yes Radboud University Ni- www.umcn.nl/ivfda-en Embryos transplant Van Peperstraten 2010 jmegen Medical Centre; 2006 Vandemheen 2009 Cystic Fibrosis referral Yes Aaron, Ottawa ON, CA, decisionaid.ohri.ca/ decaids.html#cfda transplant 2009 (last update 2011) Vodermaier 2009 Breast cancer surgery Vodermaier, Vancouver Yes Received by email (in German) BC, CA Prostate cancer screening Yes Informed Medical Deci- Volk 1999 sions Foundation, MA, informedmedicaldecisions.org/ USA, 1999 imdf ̇decision ̇aid/ deciding-if-the-psa-test- is-right-for-you/ Vuorma 2003 Menorrhagia treatment No Vuorma, Helsinki Fin- - land, 1996 Watson 2006 Prostate cancer screening Yes Oxford, UK Included in publication Weymiller 2007 Diabetes mellitus type 2 Yes Montori, Rochester MN, mayoresearch.mayo.edu/ USA treatment mayo/research/ker ̇unit/ form.cfm Williams 2013 Prostate cancer screening Yes Georgetown University, Obtained from author Washington, DC, USA [email protected] edu Whelan 2003 Breast cancer chemo- Yes Whelan, Hamilton CA, Included in publication therapy 1995 Whelan 2004 Breast cancer surgery Yes Whelan, Hamilton CA, Included in publication 1997 Yes Wolf, Charlottesville VA, Script in publication Prostate cancer screening Wolf 1996 USA, 1996 Colon cancer screening Yes Wolf, Charlottesville VA, Script in publication Wolf 2000 USA, 2000 Wong 2006 Pregnancy termination No Bekker, Leeds, UK, 2002 - 269 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

273 Table 2. Knowledge Study Timing N decision Scale used Decision aid - Notes N Comparison- comparison mean aid mean Bozic 2013 P = 0.01 33.3% 60 Decision qual- After 1st con- 58.3% 60 sultation with ity in- surgeon strument, 19 re items knowledge (> 50%) 12 true or false P < 0.001 2.17 Immediately 103 4.9 89 Evans 2010 post questions; scores ranging − from 12 to 12 93.1% P < 0.001 Insuffi- 102 31.8% Immediately Fagerlin 2011 383 post 50% ≤ cient ( correct) Sufficient Immediately 383 61.9% 102 6.9% - post Fraenkel 2012 Open-ended OR 3.5 (95% 31% 62 61% 66 Postinterven- tion questions CI: 1.6 to 7.7, about medica- P = 0.001) tionoptionsto reduce stroke - knows medi- cations 53 1. OR Open-ended 37% Postinterven- 46 49% tion 9 (95%CI: 0.9 questions about side ef- to 4.0; P = 0. fects of medi- 07) - cations knows side ef- fects Hamann P = 0.01 7- 10.9 (5.4 SD) 58 15 (4.4 SD) 49 On discharge item multiple 2006 (~ 1 month) choice knowl- edge test (unable to standardize re- sults) *mean Pre- Heller 2008 66 14%* 67 8%* 12-item mul- tiple choice operatively increase from 270 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

274 Table 2. Knowledge ( Continued) baseline P = 0.02 P = 0.11 13-item ques- LeBlanc 2015 45 7 (4.5 to 9.0) Immediately 32 5.5 (2.5 to 8. post (in consulta- tionnaire (me- 0) dian, IQR) to- tion) tal score P = 0.01 32 Immediately 4 (2.0 to 8.0) 9- 6 (3.5 to 6.5) 45 post items knowl- edge based on decision aid 10-item yes/ Legare 2008a No difference 0.51 ± 1.47 P 41 0.86 ± 1.77 43 Change scores between from baseline no/ P = 0.002 = 0.031 to 2 weeks groups (P = 0. unsure general knowledge 162) test about nat- ural health products (not specific to out- comes of op- tions) Mann D 2010 14-item Immediately - - - - No differ- survey (in consulta- ence in level of post tion) knowledge be- tween groups Mathers 2012 P < 0.001 28.8% 80 51.6% 6 months Correctly an- 95 swers question postinterven- tion about best op- tion to lower blood sugar Correctly an- P = 0.90 29% 80 31.0% 6 months 95 swers question postinterven- tion about best op- tion to lower complications Mathieu 2007 9-item-4con- Signif- - 357 - 351 - cept questions icantly higher mean increase and 5 numeric questions for the inter- vention group (2.62 ) com- pared to con- trol group (0. 271 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

275 Table 2. Knowledge ( Continued) 68) from base- line, P < 0.001 Interven- - Miller 2005 - - - 2-week, 2- 8-item survey tion type had month, and 6- no impact on month follow- gen- ups eral or specific knowledge 88% (147/ Nagle 2008 Good level - - - - - knowledge 167) in DA group com- was pared to 72% scored higher (123/171) than the mid point of pamphlet the knowledge group. OR 3. scale (greater 43 (95% CI 1. 79 to 6.58) than 4) Ozanne 2007 change 45% to 57% Change 15 48% to 64% Post-test 15 in knowledge in knowledge (in consulta- from baseline score was sig- tion) nificantforde- cision aid (P = 0.01) but not control (P = 0. 13) 10-item P = 0.001 Partin 2004 2 weeks 308 7.44 290 6.9 knowledge in- dex score Rubel 2010 24- to- The - 100 - Immediately 100 post items adapted tal mean stan- dardized from existing prostate can- knowl- edge score was cer knowledge 84.38 (SD 12. measures 38) Ad- Trevena 2008 P = 0.0001 8/137 137 1 month 134 28/134 equate knowl- edge (positive score: under- standing ben- efits/harms) 272 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

276 Table 2. Knowledge ( Continued) 522 12-item true/ Post-test 468 75% (range 0 P < 0.0001 25% (range 0 Watson 2006 to 100) to 100) false/don’t know Mean differ- - 46 14-item-9ad- 52 Immediately Weymiller dressed by de- 2007 ence between post (in con- cision aid; 5 groups 2. sultation) 4 (95% CI 1.5 were not to 3.3) P < 0. 05 (when de- cision aid ad- min- istered during the consulta- tion only - not if prior to the consultation) CI : confidence interval; DA : decision aid; OR : odds ratio; SD : standard deviation. Table 3. Accurate risk perceptions Study N decision Scale used Decision aid - Notes Timing Comparison- N comparison mean mean aid Accuracy of 69 Postinterven- 9.1 (SD 13.3) 66 14.2 (SD 13) P = 0.002 Fraenkel 2012 stroke risk tion (reported by taking the absolute value of the differ- ence between the partic- ipant’s risk as es- timated by the DA and the estimate pro- vided by the par- ticipant - out of 100; lower score indicates more accurate estimation of risk) 273 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

277 Table 3. Accurate risk perceptions ( Continued) 13.1 (SD 12. Accuracy of P = 0.004 69 8.7 (SD 12.5) 66 Postinterven- 2) tion bleeding risk (reported same as above) Post (after re- P = 0.001 2.6 129 2.3 Hanson 2011 Expectation of 127 benefit index viewing DA) 11 items score from 1 to 4 with lower score in- dicating better knowledge Correct es- 208 (59.0%) P < 0.001 Kuppermann 3-6 months 353 263 (73.8%) 357 timate of am- 2014 postinterven- niocente- tion sis miscarriage risk 353 Correct esti- 3-6 months 357 210 (58.7%) P = 0.001 163 (46.1%) postinterven- mate of Down tion syndrome risk - Mann E 2010 - Total knowl- - - 2 weeks post 3 of 8 multiple edge reported choice items only in the knowl- test edge (question 4, 5, 7) Mathieu 2010 Post 113 3.02 189 2.45 P < 0.001 5 item numer- ical questions (max = 5) - 2-week, 2- - Intervention Miller 2005 - - - month, and 6- type had no impact on risk month follow- perceptions ups 2.93 (SD 2. 8 0.58 (SD 1. Smith 2010 - 357 numer- 173 P < 0.001 91) ical questions 28) (max = 8) - Weymiller Difference be- 46 - Immediately 52 - (in con- tween group 2007 sultation) OR 22.4 (95% CI 274 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

278 Table 3. Accurate risk perceptions Continued) ( 5.9 to 85.8) when decision aid admin- istered during the consulta- tion only (not if prior to) OR 6.7 (95% CI 2.2 to 19. 7) when the de- cision aid ad- min- istered prior to or during the consultation DA : decision aid; OR : confidence interval; SD : standard deviation. CI : odds ratio; Table 4. Values congruent with chosen option Scale used Timing N decision Decision aid - Study Notes N Comparison- aid comparison mean mean The interven- - 77 - 75 Postinterven- Arterburn Percent match 2011 tion group ex- procedures de- tion pe- scribed by rienced a more Sepucha et al rapid early im- (2007; 2008). provement in val- For ues items were value concor- most predic- dance imme- tive and used diately to specify lo- after the inter- gis- vention com- pared to con- tic models to trol estimate pre- dicted proba- bility of select- ing surgery > 0.5 209 No difference Berry 2013 6 months - 239 - Con- cordant when postinterven- OR = 0.82; tion 95% CI 0.56 men reported: to 1.2 a) sexual func- tion influ- enced decision 275 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

279 Table 4. Values congruent with chosen option ( Continued) and they had radiation ther- apy; b) bowel function in- fluenced deci- sion and they had surgery; c) all effects in- fluenced deci- sion and they had surveillance 151 - - 155 within weeks Men assigned Concordance Frosch 2008a to the decision between par- aid who chose ticipant’s pref- not to have a erences and values for PSA test rated potential out- their concern comes related about prostate to the decision cancer lower than did and the choice made men who re- quested a PSA test. Men as- signed to usual care provided similar ratings of concern about prostate cancer regard- less of their PSA de- cision. There was no statis- tically signifi- cantdifference between groups Legare 2008a - - - - - - Women valu- ing of non-chem- ical aspect of natural health products was positively as- sociated with their choice of 276 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

280 Table 4. Values congruent with chosen option ( Continued) nature health products, P = 0.006. No dif- ference be- tween groups No difference; - - - Association - Lerman 1997 - between- between val- group differ- ues and choice ences were not reported Congru- Vandemheen Patient - 70 3 weeks 70 - choices were ence between 2009 personal val- consis- ues and deci- tent with their sion values across both random- ized groups : decision aid; DA SD : standard deviation. Table 5. Decisional Conflict Score N N decision Scale used Study Notes Timing Comparison- Decision aid - comparison mean mean aid 11.8 Total de- − Mean 77 Arterburn P = 0.03 Immediately 20SD 75 Mean − post 19.44 cisional con- 2011 SD 22.83 flict- change from base- line (standard- ised values) Berry 2013 Decisional Uncertainty - − 3.61 units - - P = 0.04 conflict scale Uninformed - - - No significant - difference - − 3.57 units - - P = 0.002 Unclear values Unsupported - - - - No significant difference - - - - No significant Ineffective de- cision difference Total - − 1.75 units - - P = 0.07 277 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

281 Table 5. Decisional Conflict Score Continued) ( Decisional - - - - DCS Immediately Fagerlin 2011 post conflict scale was higher in inter- the vention group compared to control, P < 0. 001 Frosch 2008a Decisional Feeling unin- P < 0.05 155 23.37 151 29.68 formed conflict - sub- scales only Feeling 155 32.25 151 37.93 P < 0.05 unclear values 35.21 155 30.51 151 Feeling P < 0.05 supported Feeling uncer- - 151 - No difference 155 tain 155 - 151 - No difference Effective deci- sions Deci- 19 SD 14 No difference 4 months 73 Knops 2014 81 22 SD 17 sional conflict 73 21 SD 17 81 18 SD 17 No difference 10 months (total score) Decisional No difference Immedi- 1.58 Krist 2007 196 1.54 75 ately after of- conflict fice visit LeBlanc 2015 Decision con- P = 0.18 Immediately 22.7 (95% CI 28 10.9 (95% CI 36 flict (overall) post 7.8 to 28.5) (in consult) 1.6 to 26.6) median, IQR 20.8 (95% CI Immediately P = 0.14 28 4.2 (95% CI 0 Informed sub- 36 to 25) post scale 0 to 33.3) 25.0 (95% CI Immediately P = 0.25 28 16.7 (95% CI Values 36 0 to 25) post subscale 8.3 to 33.3) P = 0.35 Support sub- 28 8.3 (95% CI 0 36 16.7 (95% CI Immediately scale post 0 to 25) to 25) Certainty sub- 28 8.3 (95% CI 0 Immediately 36 25 (95% CI 0 P = 0.3 scale to 25) post to 25) Effectiveness P = 0.15 Immediately 18.8 (95% CI 28 12.5 (95% CI 36 0 to 25) post 0 to 25) subscale 278 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

282 Table 5. Decisional Conflict Score ( Continued) 6.3% (95% 4.6% (95% Legare 2012 163 Absolute Immediately De- 165 CI 2.6 to 7.4) CI 0 to 12.8) post difference 1.7; cisional con- (in consult) RR 0.8 (95% flict - propor- CI 0.2 to 2.4) tion who had a value of 2.5 or more on the − 1 5 DCS. (n, %) Leighl 2011 No difference 26 (range 0- 1-2 107 26 (range 0- 100 Decisional weeks postin- 79) 67) conflict scale tervention median (range) Based on ap- Immediately P = 0.24 64% 110 71% Mathieu 2010 91 proaches sug- after interven- tion gested by Marteau et al. (informed choice) Ozanne 2007 Both groups Decisional - Postconsulta- 15 15 - tion conflict showed lower (in consult) de- cisional con- flict postcon- sultation (P < 0.001) but no difference be- tween groups Decisional to- Immediately Rubel 2010 - - - - The conflict tal mean score post was 24.5 with a SD of 15.25 (N = 200) Schwartz Sig- Decisional - 12 of 16 items - - - conflict 2009a nificant longi- of the original scale tudinal im- pact of the de- cision aid was moderated by baseline deci- sion status; de- ci- sion aid led to significant de- 279 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

283 Table 5. Decisional Conflict Score ( Continued) creases in de- cisional con- flict for those who were un- decided at the time of ran- domisation Difference be- - Decisional 56 Postconsulta- Thomson 53 - conflict tween de- 2007 tion (in con- cision aid and sult) control group − 0.18 were − (95% CI 0. − 34 to 0.01). P = 0.036 3-months post 51 - 55 - Difference be- de- tween cision aid and control group were − 0.15 (95% CI − 0. 37to0.06), no significant dif- ference Postinterven- Van 15 item ques- P = 0.76 75 128 72.5 124 tion, pre IVF tionnaire (1- Peperstraten 2010 5) - satisfaction- uncertainty 15 item ques- P = 0.001 87.5 128 77.5 124 Postinterven- tion, pre IVF tionnaire (1- 5) - informed (includes some items from DCS) Weymiller Mean - 46 Decisional - Immediately 52 post 2007 difference in- conflict (in con- dicates statis- sult) tically signif- icantly lower de- cisional con- flict for deci- sion aid com- 280 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

284 Table 5. Decisional Conflict Score ( Continued) pared to usual care Total DCS 10.6 (95% − − CI 15.4 to 5.9) − Un- − 12.8 certain 18. (95% CI − 7.3) − 4 to In- formed − 17.3 (95% CI − 22. − 12.0) if 6 to administered during consult 6.6 (95% CI − − 1. − 14.3 to 1) if adminis- tered prior to consult Values clarity − 8.5 − 15. (95% CI − 7 to 1.3) − 9.4 Support (95% CI 14. − 8 to − 3.9) Effective deci- sion − 10.0 (95% CI − 15. 5.0) 0 to − DA : decision aid; DCS : decisional conflict scale; CI : in vitro fertilisation; SD : standard deviation. : confidence interval; IVF Table 6. Decisional Conflict Score - low literacy version Scale used Timing N decision aid Decision aid - Study N comparison Comparison - Notes mean mean Fraenkel 2012 Informed Immediately 69 13.0 66 24.8 P = 0.01 post Values 69 6.4 66 21.0 P <.001 Immediately post Smith 2010 Total DCS 2 week follow- P = 0.02 357 13.63 (SD 20. 14.91 (SD 18. 173 55) 34) up 281 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

285 Table 6. Decisional Conflict Score - low literacy version Continued) ( Used 8 of 10 Total DCS 80 24.1% high 74 41.9% high Results were Taylor 2006 dichotomized items only (items 1 month post re- moved choos- ing without pressure from others; know what op- tions are avail- able to you) Significant de- Total DCS 2 months post 153 27.5% 136 38.2% Williams crease 2013 for DA group compared to usual care in the home con- dition site 13 months 153 136 31.6% No difference 38.6% post : decision aid; DCS : decisional conflict scale; SD : standard deviation. DA Table 7. Decisional Conflict Score - SURE test Study Scale used Timing N decision Decision aid - N Comparison - Notes comparison mean mean aid 80.3% Stacey 2014a SURE tool Postinterven- No difference 66 72.3% 65 tion; prior to Item: ’Feels sure about the surgical con- sult best choice’ ’Knows the 66 No difference 66.7% 65 92.3% Postinterven- tion; prior to benefits and harms . . .’ surgical con- sult 87.7% 66 No difference ’Clear about 65 74.2% Postinterven- tion; prior to which benefits and harms . . . surgical con- sult ’ ’Has enough No difference 77.3% Postinterven- 66 76.9% 65 tion; prior to support and surgical con- advice . . .’ sult 282 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

286 Table 7. Decisional Conflict Score - SURE test Continued) ( 57.6% Postinterven- Total SURE 65 69.2% 66 No difference tion; prior to score surgical con- sult Table 8. Patient-clinician communication Study N decision Scale used Decision aid - Notes N Comparison- Timing comparison mean mean aid Discussed risk P < 0.001 Immediately Fraenkel 2012 69 71% 66 12% post of stroke 66 Discussed risk P < 0.001 69 69% Immediately 20% post of major bleeding Hanson 2011 P = 0.04 33% 127 Discussed 3 months 126 46% feeding with physician, nurse clini- cian, or physi- cian’s assistant Dis- 3 months 126 64% 127 71% P = 0.42 cussed feeding with other nursing home staff Hess 2012 (in OPTION Signif- Analysis of the Mean 103 Mean 26.6% 101 (95% CI 24.9 consultation 7% (95% CI consult) scale icantly greater using video- 5.9 to 8.1) to 8.2) in the inter- vention arm recordings LeBlanc 2015 P = 0.001 OPTION Mean Analysis of the 13 Mean 25 consultation scale 43% (95% CI 57% (95% CI (in consult) 50 to 64) 37 to 48) using video- recordings Discussed 8.3% P < 0.001 216 Lepore 2012 8 months 15.8% 215 postinterven- PSA testing tion with physician postinterven- tion 283 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

287 Table 8. Patient-clinician communication ( Continued) 49.8 OPTION Montori 2011 P < 0.001 27.3 32 38 Analysis of the consultation 100-point (in consult) using video- scale recorded con- sultations Analysis of the Mean 27.7% OPTION 21. 48 used deci- MD Mean 49.7% 37 usual care Mullan 2009 aid consultation (SD 11.75) scale (SD 17.74) 8 (95% CI 13. sion (in consult) within consul- using video- 0 to 30.5) for tation decision aid vs recorded con- sultations usual care. All but 2 of the 12 items sig- nificantly favoured the decision aid Dis- - Abso- Patient re- Sheridan2006 8/34 usual 16/41 de- - cussed CHD cision aid pre- care ported Imme- lute difference 16% (95% CI consult diately post with doctor − 4 to 37) with summary report to bring to consult 8/34 usual Patient re- - Plan to reduce 15/41 de- - Abso- lute difference cision aid pre- ported Imme- CHD risk and care 13% (95% CI consult diately post discussed with doctor − 7 to 34). with summary report to bring to consult Patient re- Abso- 37/41 de- Plan to reduce - 25/34 usual - CHD risk and cision aid pre- care ported Imme- lute difference not discussed consult diately post 16% (95% CI with summary with doctor − 1 to 33) report to bring to consult Sheridan2011 Had CHD Abso- 58% Patient 78 89% 79 discussion lute difference reported with provider 31% (95% CI Immediately post 15 to 45; P < 0.001) Abso- Patient 35% 78 Patient-raised 79 63% reported discussion lute difference Immediately 28% (95% CI post 9 to 45; P = 0. 02) 284 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

288 Table 8. Patient-clinician communication ( Continued) Abso- Mod- Patient 79 76% 78 91% ified Health- lute difference reported care Climate 15% (95% CI Immediately post Question- 0.1 to 31; P − = 0.02) naire: 1. “My provider pro- vided me with choices and options about lowering my chances of heart disease” Patient 95% Abso- 2. 78 79 86% reported “My provider lute difference 9% (95% CI Immediately under- post 7 to 25; P = stands how I − 0.21) see things with respect to low- ering my chances of heart disease.” 88% 79 78 patient 3. 77% Abso- lute difference reported “My provider Immediately 11% (95% CI conveyed con- − post fidence in my 5 to 27; P = abilitytomake 0.15) changes regarding low- ering my chances of heart disease” 4. Patient 79 78% 78 67% Abso- “My provider reported lute difference 11% (95% CI encour- Immediately post − 4% to 27%; aged me to ask P = 0.13) questions” “My Abso- 71% 78 Patient 5. 92% 79 reported provider lis- lute difference Immediately tened to how I 21% (CI 95% post would like to 6 to 37; P < 0. do things” 01) 78 6. Abso- Patient 69% 84% 79 “My provider reported lute difference Immediately tried to under- 285 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

289 Table 8. Patient-clinician communication ( Continued) post standing how 15% (CI 95% − 0.3 to 31; P I see things be- fore suggest- = 0.05) ing new ways to lower my chances of heart disease.” Greater - Usual care OPTION - Analysis of the Weymiller 1/2 used deci- sion aid prior consultation Scale 2007 patient partic- (in con- to consult and using video- ipation MD 4. sult) 1/2 4 (95% CI 2.9 recorded con- used it during to 6.0) in deci- sultations sion aid com- consult pared to usual care : coronary heart disease; CI : confidence interval; DA : decision aid; DCS : decisional conflict scale; ICC CHD : intraclass correlation coefficient; MD : mean difference; OPTION scale : observing patient involvement scale; RR : risk ratio; SD : standard deviation Table 9. Participation in decision making Study Scale used N decisionaid Decision aid - Timing N Notes Comparison- mean mean comparison 95% No difference Allen 2010 334 Control pref- Postinterven- 291 92% erences - pa- tion tients choos- ing active/col- laborative de- cision making Control pref- No difference 87% Postinterven- 334 291 92% tion erences did not change 334 No difference 5% Postinterven- Control pref- 291 3% tion er- ences changed to passive 3% Control pref- 7% 334 Postinterven- 291 No difference er- tion ences changed to active/ col- laborative 286 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

290 Table 9. Participation in decision making ( Continued) 49 COMRADE Hamann Postconsulta- 79.5 (SD 18. Increased pa- 58 69.7 (SD 20. tient involve- 6) 2006 tion used to mea- 0) sure patients’ 76.8 (SD 20. ment in deci- 73.5 (SD 19. 3) 9) sion aid group perceived in- postinterven- volvement in tion compared decisions to usual care at baseline. At discharge there was no difference be- tween groups Surrogates P = 0.18 77% Hanson 2011 - 83% Postinterven- - tion feeling some- what or very in- volved in deci- sion making Leighl 2011 Achieved de- No difference 35% Postinterven- - - 32% tion cision involve- ment 96 (in Patients’ per- Im- 24.5 pre Loh 2007 26.3 pre 28.0 191 Postconsulta- tion post proved patient ceived consult) in- 25.5 post volvement in participation from baseline decision mak- ing to post expo- sure to the de- cision aid (P = 0.010) and in comparison to the usual care group (P = 0. 003) but there was no change in the control group for the pre-post com- parison Rubel 2010 Adapted from total The - Postinterven- - - - tion the mean Control Pref- scores were: 2. 74 (SD 1.25) erences Scale (N = 99) pre and 2.83 (SD 287 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

291 Table 9. Participation in decision making ( Continued) 1. 16) (N = 199) post, no statis- tically signifi- cantdifference 79% Abso- Immediately 51% Sheridan2011 Patient partic- 78 79 post lute difference ipation: ’Any’ 28% (95% CI 9 to 45; P = 0. 01) ’None’ Immediately 79 21% 78 49% Absolute dif- ference post 28% − (95% CI − 45 to − 9) Van P = 0.33 87.5 Decision Eval- 128 85 124 Postconsulta- tion uation scale Peperstraten (15 item ques- 2010 tionnaire) De- cision control subscale SD : standard deviation. : decision aid; DA Table 10. Proportion undecided Scale used Timing N decisionaid Decision aid - N Study Notes Comparison - mean comparison mean Single - - - - Kasper 2008 No difference - item - ranging from ’0 = com- pletely unde- cided’ to ’100 = made my de- cision’ 37 “I Sawka 2012 - 21.6% Answer 10.8% Im- 37 mediately post don’t know” to - treatment question “I fa- preference vor taking ad- juvant radioactive io- dine” 288 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

292 Table 10. Proportion undecided ( Continued) 37 13.5% 37 8.1% - 6.3 months (mean) post - actual decision “I 37.8% Answer 37 43.2% 37 Im- - don’t know” to mediately post - treatment question “I fa- preference vor not taking adjuvant radioactive io- 6.3 months - 51.4% 37 40.5% 37 dine” (mean) post - actual decision : decision aid DA Table 11. Satisfaction with the choice Scale used Timing N decisionaid Decision aid - Study N Notes Comparison- mean comparison mean Heller 2008 1- P = 0.03 - 55/67 1 month post- - 62/66 surgery item; pleased with treat- ment choice Single 8.54 (SD 1. Legare 2012 8.53 (SD 1. 159 162 Immediately No difference; (in consult) question Lik- 51) post 56) MD 0.0 (95% CI − 0.4 to 0. ert scale to as- 4) sess the qual- ity of the deci- sion made (0 = very low qual- ity; 10 = very high quality) Satisfac- Leighl 2011 No difference 21(15-25) 100 1 22 (13-25) 107 tion with deci- month postin- tervention sion scale: median (range) Marteau 2010 Scale: ranging No difference 91.33 (SD 14. - 91.17 (SD 14) - 4 weeks from 1 7 and − 50) standardized out 100 289 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

293 Table 11. Satisfaction with the choice ( Continued) Over- 6-item Schwartz 100 - 114 - 1, 6, 12 all, no differ- months 2009b ence between groups; deci- sion aid led to significantly increased sat- isfaction com- pared to usual care among those who were un- decidedatran- dom- ization but not among those who had made a decision be- fore random- ization; (only graph in paper with no raw data) Taylor 2006 Single item - - 75.7% 74 79.7% 80 1 month ”Are you satis- fied with your decision about prostate can- cer testing? - Satisfaction No difference Trevena 2008 - 134 137 Immediately with the deci- post (P = 0.56) sion Williams - 6-item Sat- > 95% - Baseline - > 95% 2013 isfaction with Decision Scale DA : decision aid. Table 12. Satisfaction with the decision-making process Study Scale used Timing N decisionaid Decision aid - Notes N Comparison- comparison mean mean Satisfaction with the decision-making process Satisfaction with sion-making process 290 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

294 Table 12. Satisfaction with the decision-making process Continued) ( Patients in DA 103 (in Satisfac- - Hess 2012 - 101 - consult) tion with de- group re- cision process ported greater sat- (0 for strongly isfaction with agree to 5 for strongly the DM pro- cess (strongly disagree) agree, 61% DA vs 40% usual care) 53 Satisfied with High satisfac- 1 week follow- 50 Vodermaier 42 56 tion with no up 2009 process difference by group Satisfaction with participating in decision making Satisfaction with ing in decision making - Com- - - - Kennedy Measured sat- - isfaction 2002 pared to usual with opportu- care, women nities to par- who re- ticipate in de- ceived the de- cision aid fol- cision making by lowed using a single item nurse coaching were significantly more satisfied with the op- portunities to participate in decision mak- ing (OR 1.5, 95% CI 1.1 to 2.0) Satisfaction with the information provided Satisfaction with mation provided LeBlanc 2015 P = 0.69 Amount of in- 34 (92%) 37 Postconsulta- 25 (86%) 29 tion formation was (in consult) just right Informa- P = 0.43 26 (72%) Postconsulta- 36 27 17 (63%) tion tion received was clear 291 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

295 Table 12. Satisfaction with the decision-making process ( Continued) 28 21 (75%) 34 23 (68%) P = 0.53 Postconsulta- Informa- tion received tion was helpful Postconsulta- 35 28 24 (86%) Would recom- 27 (77%) P = 0.52 tion mend method to others P = 0.001 Laupacis 2006 Satisfac- 59 (23.3 SD) 56 Average 10 54 76 (15.5 SD) tion with in- days formation re- ceived sub- scale 4-item (0 to 100; low to high) (7 point Montori 2011 46 6.6 49 Postinterven- P = 0.798 6.3 tion 6 6 P = 0.296 scales) (in consult) 5.8 Participants’ 6 P = 0.624 satisfaction 6.1 5.8 P = 0.248 P = 0.435 6.4 6.2 with knowl- edge transfer Amount • of information • Clarity of information Helpfulness • of the information Would • want other decisions Recommend • to others 5.2 P = 0.006 33 Clinicians’ sat- Postinterven- 39 5.8 6.1 tion isfaction with P < 0.001 4.9 knowledge 4.8 P < 0.001 5.9 transfer Helpfulness • of the information Would • want other 292 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

296 Table 12. Satisfaction with the decision-making process ( Continued) decisions • Recommend to others 10.1 (SD 2.2) Satisfaction 4 months post 16 Oakley 2006 17 10.4 (SD 2.9) No difference with informa- tion about medicines Satisfaction with the clinician Satisfaction with cian P = 0.004 69 (25.3 SD) 54 Average 10 Satisfac- Laupacis 2006 56 54 (26.7 SD) days tionwith prac- titioner treat- ment during decision pro- cess sub- scale 4-item (0 to 100; low to high) 4.37 (0.84 Sat- No difference 4.38 (0.86 Miller 2005 - - 2 weeks SD) isfaction with SD) cancer infor- mation service 6 months No difference 4.51 (0.64 - 4.51 (0.75 - 1-item (1 to 5; SD) SD) low to high) High satisfac- 53 (94.6%) Vodermaier 49 (92.5%) 53 1 week follow- 56 2009 up tion with no 47 50 Physician • difference by 47 51 helped me 44 45 group understand 36 36 Physician • understood important to me • Physician answered questions Satisfied • with involvement Satisfied • with physician’s involvement 293 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

297 DA : standard deviation. SD : decision aid; Table 13. Preparation for decision making Comparison- Decision aid - Study Timing Scale used N decisionaid Notes N mean comparison mean 43 Fraenkel 2007 P < 0.001 Pre- 20.5 (median) Prepa- 35 (median) 40 consultation ration for De- cision Making Scale Postinterven- 4.12 (SD 1. 64 66 Stacey 2014a Prepa- 3.78 (SD 1. No difference 25) 21) ration for De- tion; pre-con- cision Making sultation Scale item (5- point scale from: 1 not at all to 5 a great deal) ’Help recognize de- cision to be made’ 4.48 (SD 0. 66 No difference Prepa- Postinterven- 4.14 (SD 1. 64 10) tion; pre-con- ration for De- 85) sultation cision Making Scale item ’Help know decision de- pends on what matters most’ No difference Prepa- 4.25 (SD 1. 64 4.48 (SD 0. 66 Postinterven- 05) ration for De- 81) tion; pre-con- sultation cision Making Scale item ’Help think about how in- volved you want to be in decision’ No difference 4.23 (SD 1. Postinterven- 4.36 (SD 0. 64 66 Prepa- ration for De- tion; pre-con- 04) 91) sultation cision Making Scale item ’Prepare you 294 Decision aids for people facing health treatment or screeni ng decisions (Review) Copyright © 2017 The Cochrane Collaboration. Published by J ohn Wiley & Sons, Ltd.

Related documents