Impact of Potential Changes to the Treatment of Manufacturer Rebates

Transcript

1 M illima n Client Report Changes to the Impact of Potential Rebate s Treatment of Manufacturer Prepared for: The Assistant Secretary for Planning and Evaluation U.S. Department of Health and Human Services Prepared by: Milliman, Inc. 15800 Bluemound Road Suite 1 00 Jake Klaisner , FSA, MAAA 53005 Brookfield, WI USA Actuary +1 262 784 2250 Tel +1 262 923 3680 Fax Katie Holcomb , FSA, MAAA milliman.com Actuary Consulting Troy Filipek , FSA, MAAA Principal and Consulting Actuary January 31, 2019

2 Milliman Client Report TABLE OF CONTENTS EXECUTIVE SUMMARY ... ... ... ... 1 I. II. BACKGROUND ... ... ... ... 3 ... III. RESULTS ... ... ... ... ... 7 IV. CONSIDERATIONS ... ... ... 18 ... V. METH ODOLOGY AND ASSUMPTI ONS ... ... ... 27 , AND QUALIFICATIONS CAVEATS, LIMITATIONS 28 ... ... ... VI. 1 – Manufacturer Rebates Annual Dollar Impact of Removing APPENDIX A (Billions of Dollars) APPENDIX A 2 – Dollar Impact of Removing Manufacturer Rebates (PMPM) – Removing (Billions of Dollars) Manufacturer Rebates After APPENDIX B 1 Projected Annual Costs APPENDIX B 2 – Projected Costs After Removing Manufacturer Rebates (PMPM) of Removing Cost age Manufacturer Rebates APPENDIX C – Annual Percent Impact U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

3 Milliman Client Report EXECUTIVE SUMMARY I. Assistant Secretary for Planning and Evaluation The ASPE ) for the US Department of Health and Human ( requested that Milliman analyze how the U.S. healthcare market Services (HHS) would be impacted by potential changes disallow ing manufacturer regulatory in their current form . This report analy z es rebates the impact of this potential change , with focus on the impact to stakeholders in the Medicare Part D program. Currently, rebates are protected by safe harbor regulations under the f ederal all Medicare Part D Anti Kickback Statute, which otherwise prohibit payment s in exchange for services under a federal - healthcare program. If this safe harbor protection is alter ed to no longer encompass manufacturer rebates , these could be completely prohibited under Medicare Part D. In this report , we examine the impact rebates of price manufacturer rebates from the Medicare Part D program and replacing them with a removing of his report does not analyze the T . sale (POS) price - concession producing a correspondingly lower point - impact of any potential changes to pharmacy rebates, which are a different form of rebates common in Part impact for a statu s quo scenario ( holding all other assumptions constant ) , as well as We looked at this D. by plan sponsors , members , and pharmaceutical manufacturers layering in potential behavioral changes (e.g., tighter formularies , changes to price concessions , changes in utilization ) . to member, government, and Table 1 below show s the 10 year estimated impact between 2020 and 2029 , assuming the changes manufacturer costs under varying behavior change scenarios in billions of dollars 20 . While several scenarios are presented, we do not intend to imply that any one took place in 20 first outcome is more likely than another. For example, scenarios that may decrease costs are not necessarily more likely than scenarios that may in crease costs . Table 1 ASPE Protections for Manufacturer Rebates Impact of Removing Safe Harbor 2020 to 2029 (Billions of Dollars) Total Total Total Member Program Gov ’ Member t Cost Member Costs* Costs* Costs* Sharing NAR NADS Premium LIPS Scenario CGDP LICS $34.8 $14.5 - $0 - $26.4 - $103.1 $215.4 $40.8 $89.5 $12.0 - - $20.6 No Behavioral Changes Increased Formulary 164.6 - - 56.2 - 78.8 - 139.1 174.7 - 8.4 3.8 - 64.6 - 29.5 118.3 Controls Increased Formulary - 99.6 - 59.5 188.2 - 63.5 4.0 - 29.1 - 163.2 180.3 118.5 1.9 Controls and Increased - - Price Concessions Price Reduced 12.3 135.1 139.9 221.1 - - 32.6 44.9 - 30.2 17.1 - 71.4 20.5 Concessions Decreased Brand Price - - - 38.4 94.3 33.1 - - 46.5 13.3 22.8 6.1 - 154.6 211.4 - 101.3 Trends Increased Brand Use 26.0 69.9 - - - 22.7 and Decreased Brand 21.2 - 143.8 - 214.8 - 101.3 7.6 42.7 16.7 - Trends Price Increased Pharmacy 18.1 17.8 20.9 - - 2 - 2 2 .8 - 20.6 - 11 0 . 20 7 . 1 - 89.5 10.4 40.8 Rebates . due to rounding to individual components *Totals may not sum projected costs and relative The appendices at the end of this report provide further details on the total impacts in each year from 2020 to 2029. takeholders We estimate s affected as follows: could be Members have overall cost savings on average, driven by lower cost sharing, partially offset by  individual members the distribution of savings is not uniform, and With that said, higher premiums. will be impacted differently depending on their pharmacy spend levels and other characteristics. 1 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

4 Milliman Client Report For example, m may have no change to cost sharing, embers not taking any products with rebates but would see an increase in premiums , though depending on changes to benefit designs, these members may also have savings . cost ould have , excluding the impact of any behavioral changes. s increase  The government c W However, we believe it is unlike ly for there to be no behavioral changes. ith plan bid amounts behavior changes result strategic ing and member premiums increas ing , stakeholders may make over time. ) including government savings ( in overall program savings  Manufacturer claim liabilities through the coverage gap discount program (CGDP) would be lower. , this would produce overall savings Assuming no changes to total costs net of manufacturer rebates for manufacturers. However, if plan sponsors push manufacture rs for increased price concessions could produce an increase in as a behavioral impact resulting from these policy changes , it that manufacturer costs our analysis focuses on the impact to manufacturers regarding their . Note ram through the CGDP. We did not model the net impact to role in funding the Part D prog since it was product sales and price concessions manufacturers after accounting for changes in outside of the scope the project timeline would allow. The impact of potential behavioral changes could increase or decrease costs. We estimate  i ncreased formulary controls, higher price concessions, and lower price trends would all reduce . T crease in government he in overall program costs, including greater average member savings costs expected s cenario now becomes a decrease in government in the no behavioral change costs. On the other hand, if price concessions are lower than their current values, total program rease by more and the increase in member government costs would inc ; costs would increase premium could outweigh average savings for member cost sharing. 2 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

5 Milliman Client Report II. BACKGROUND to analyze how the U.S. healthcare market Milliman the removal of ASPE engaged would be impacted by . In particular, we considered the impact to the Medicare Part D manufacturer rebates program if rebates were eliminated and replaced with lower POS costs. We also examined manufacturer completely potential behavioral changes that may occur as a result of this ch of . the effect ange BACKGROUND ON MEDICA RE PART D includes pharmacy coverage for Medicare - Medicare Part D eligible beneficiaries who choose to enroll in individually or through their employer for retiree - : one of two plan types based coverage Medicare Advantage Part  , which provide both medical and pharmacy coverage D (MAPD) plans  Stand - alone Prescription Drug Plans (PDPs), which provide only pharmacy coverage statutorily defined . Plan sponsors are allowed to vary their Part D benefit has multiple phases standard The benefits (with some limitations), as long as the design results in the same or better value to the member, on average. The benefit phases standard cost sharing for 2019 and are as follows:  Deductible Phas e : A $415 deductible during which members pay 100% of allowed claim costs.  Initial Coverage Phase : Members pay 25% of allowed claim costs with plan sponsor s paying the is reached remaining 75% until in total allowed costs $3,820 . the initial coverage limit of  Coverage Gap Phase: In this phase, members pay 37 % of generic costs and 25 % of brand costs . As part of the Affordable Care Act (ACA) , pharmaceutical manufacturers provide a discount for most brand medications filled by non - low income (NLI) members duri ng the coverage gap phase. This is known as the Coverage Gap Discount Program (CGDP). As of 2019, the CGDP was sponsor s pay lan increased from 50% to 70% as part of the Bipartisan Budget Act of 2018 (BBA). P brand costs % of 5 the remaining % of generic costs and 63 Low income (LI) . within this phase subsidy members are not eligible for the CGDP since eligible they receive federal cost sharing - subsidies i n the coverage gap and all other phases of the benefit .  Catastrophic Phase: Once a member ’s true out - of - pocket ( TrOOP ) spending reaches a catastrophic ( $5,10 0 in combined member and CGDP spending ), they enter the final threshold phase, known as the catastrophic or reinsurance phase. A fter this point, members pay roughly 5% coinsurance, plan sponsors pay approximately 15%, and the federal government pays the federal reinsurance. as remaining 80% of claim s costs beyond coverage nhanced also may In addition to varying the standard benefit parameters, plans offer e the benefit ( e.g., by reducing or e liminating the deductible, reducing cost sharing during the initial standard coverage phase , providing additional coverage in the coverage gap, and / or covering medications not typically covered by Part D ). REBATES IN PART D e reduce net costs and in turn plan sponsors to help and pharmacy Manufacturer rebates ar negotiated by member premiums. ebates are a form of direct and indirect remuneration Both manufacturer and pharmacy r (DIR), which is a payment occurring after the POS to help reduce the final price of a medication paid by the plan sponsor. Since rebates occur after the POS and are paid directly to the plan sponsor or pharmacy benefit manager ( , they PBM) only directly reduce costs for the plan sponsor and are not shared directly with members or other stakeholders at the POS. T he plan sponsor shares a portion of rebates with the federal government, proportionate to the a mount of claims paid by the government through reinsurance s rebate (we estimate this proportion to be about 35% in 2019, on average). The plan keeps the remainder of keeping s and use to reduce premiums for all members. This is valuable to a plan sponsor since them enrollment. individual Part D premiums low is typically the primary tool for attracting 3 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

6 Milliman Client Report Plans collect rebates during all phases of the Part D benefit, regardless of what the plan liability is in a given ( anufacturer rebates generally apply only phase. M which tend to cost more than generic to brand products products ) and as such , a large portion of rebates is associated with claims in the coverage gap and catastrophic phases. Plans retain a majority of the rebates collected during these phases, eve n though they will plans are only responsible for paying a small portion of pay the POS claim costs . For example, in 2019 , and % / 0% of brand claims in the gap for NLI / LI members, respectively, 5 approximately 15% of all claims . in the catastrophic phase This dynamic can create a misalignment of incentives between various stakeholders. The other primary form of DIR is price concessions associated with preferred pharmacy networks. In this a pharmacy case, network status, which generally results to the plan sponsor in exchange for pays a rebate in to encourage members to shop at those pharmacies . lower cost sharing at preferred pharmacy chains Just as with manufacturer rebates, th e rebate paid by retail pharmacies is valuable to the plan because a majority of the rebate is used to help reduce member premiums. Pharmacy rebates are typically paid on (unlike manufacturer rebates that tend to be brand only) generic and brand medications are typically , and contracted as either a flat fee per prescription or a pe rcentage of POS costs. While pharmacy rebates are typically associated with preferred network pharmacies, some plans contract DIR arrangements with non - preferred pharmacies as well. This dynamic of rebates offering high value to plan sponsors has led to plans seeking more aggressive rebate contracts with both manufacturers and pharmacies in recent years. In particular , a plan sponsor may prefer a , since the manufacturer rebate product with a high cost is oft en contracted as a percentage of costs ) claim liability is ’s plan sponsor a Since the higher cost leads to a higher dollar value of rebate . (i.e., relatively low in the coverage gap and catastrophic phases of the benefit, where a large portion of rebates a re generated, the rebates can sometimes more than offset the claim costs for which the plan is responsible. which cost product can produce a lower (or even negative) net claim liability for the plan, - This means a high , the plan perspective may make it the preferred option from even if it is not the lowest cost medication when considering all stakeholder payments (i.e., ) . measured as total costs net of rebates PART D REBATE REFORM The Centers for Medicare and Medicaid Services ( CMS ) expressed concern regar ding pharmacy DIR in early 2014, after the first preferred pharmacy networks in Part D emerged. CMS was several years concerned with pharmacy price concessions not being reflected at the POS, thus , not being shared with beneficiaries and affecting price transparency / competition. CMS proposed requiring all pharmacy price concessions be reflected at the POS, but after receiving comments on the proposed rule, ultimately scaled quire price concessions at the POS “except contingent price concessions that cannot back the proposal to re 1 - of - sale.” As a result, many reasonably be determined at the point plans structured contracts such that affecting some contingency existed, even if small or , so that conc essions could continue predictable premium when reflected as DIR rather than at the POS. More recently, there has been a great deal of public attention on the idea of reflecting all types of rebates, ing a n RFI not just pharmacy rebates, at the POS. In November 2017, CMS issued a proposed rule includ on the application of rebates at the POS. CMS proposed requiring plan sponsors to s soliciting comment reflect a minimum percentage of manufacturer rebates RFI and all pharmacy rebates in the POS price. The included an example where the POS rebate for a given therapeutic class would be determined by applying products in that class , a specified minimum percentage to the average manufacturer rebates for rebated . The proposed rule was finalized in April 2018. roaches though the RFI requested feedback on other app While no requirements regarding POS rebates implemented in the April 2018 final rule or to date , were and the responses could be used to help in form CMS noted over 1,400 responses to the RFI were received 2 future policy changes . 1 Medicare Program; Contract Year 2015 Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Dru g Benefit Programs, 79 FR 29843 (May 23, 2014). 2 Medicare Program; Contract Year 2019 Policy and Technical Changes to the Medicare Advantage, Medicare Cost Plan, Medicare Fee - for - Service, the Medicare Prescription Drug Benefit Programs, and the PACE Program – A Rule by the Centers for Medicare & Medicai d Services. April 16, 2018 . technical the - changes - - to - https://www.federalregister.gov/documents/2018/04/16/2018 - 07179/medicare - program - contract - year - 2019 - policy - and - medicare medicare. - advantage - 4 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

7 Milliman Client Report The President’s Fiscal Year 2019 Budget, which was released in February 2018, put forth a similar 3 , in conjunction proposal, which would require plans to share a “substantial portion” of rebates at the POS “d iscourage drug manufacturers’ price and with a number of oth er Par t D program changes designed to ” . rebate strategies that increase spending for both beneficiaries and the Government proposal budget The 4 and price concessions shared at the POS calls for at least one - third of total rebates . to be 2019 when B udget were reinforced in May 2018, The Part D changes outlined in the President’s Fiscal Year The Trump Administration Blueprint to Lower the Trump Administration released “American Patients First: 5 Reduce Out (HHS Blueprint). The HHS Blueprint described several Pocket Costs” - of - Drug Prices and initiatives already taken, proposed, or under consideration for future implementation to reduc pharmacy e deal with challenges in the pharmacy primary reform strategies to costs. The HHS Blueprint identified four supply chain:  Improved competition  Better negotiation  Incentives for lower list prices Lowering out  - pocket costs - of - - of public comments pocket costs and HHS was seeking POS rebates were cited as a means to reduce out on other ways to reduce the impact of the high price / high rebate dynamic existing in the Part D program today, including the idea of eliminating rebates altogether in favor of fixed price contracts. 6 impacting several aspects of Medicare Part D proposed rule . In 2018, CMS released a 26, O n November considering changes to the definition of negotiated price the proposed rule, CMS states it is would , which require plans to pay the lowest possible reimbursement at the POS. This rule change would eliminate all DIR payments from pharmacies to plans, requiring th se discounts to be reflected at the POS. Only o a pharmacy would negative DIR, or payments from plans to pharmacies, would be permitted. For example, be paid the lowest possible amount specified in the contract at the POS, but negative DIR c ould later be if the pharmacy met performance guarantees associated with higher reimbursement. paid FEDERAL ANTI - K STATUTE KICKBAC The federal Anti - , originally enacted in 1972 , prohibits offering any kind of reward or Kickback Statute incentive in exchange for services related to a federal healthcare program. However, safe harbors were established as exceptions to this rule to allow certain typical p ayments in the healthcare system. This includes rebates, defined as “any discount the terms of which are fixed and disclosed in writing to the buyer 7 of the initial purchase to which the discount applies, but which is not given at the time of sa le.” at the time As such, rebates are not considered a kickback and are widespread in Part D. On July 18, 2018, HHS sent a proposed rule to the Office of Management and Budget (OMB) that could ed “ potentially remove or change the rebate safe harbor. The proposal is titl o f Safe Harbor Removal Protection for Rebates to Plans or PBMs Involving Prescription Pharmaceuticals and Creation of New Safe 8 .” The rule has not yet been released at the time of writing, so its content is unknown Harbor Protection to the public owever , if the safe harbor on rebates is fully lifted, rebates could be eliminated from the Part D . H altogether . program 3 ear 2019. https://www.whitehouse.gov/wp Office of Management and Budget; An American Budget; Fiscal Y - - content/uploads/2018/02/budget fy2019.pdf. 4 Department of Health and Human Services; FY 2019 Budget in Brief. February 19, 2018. https://www.hhs.gov/sites/default/files/ fy - 2019 - budget - in - brief.pdf. 5 U.S. Department of Health & Human Services. American Patients First: The Trump Administration Blueprint to Lower Drug Prices and Reduce Out - of - Pocket Costs. May 2018. https://www.hhs.gov/sites/default/files/AmericanPatientsFirst.pdf. 6 CMS (November 26, 2018). Modernizin g Part D and Medicare Advantage to Lower Drug Prices and Reduce Out Pocket Expenses. - of - - - out https://www.federalregister.gov/documents/2018/11/30/2018 - 25945/modernizing - part - d - and - medicare - advantage - to - lower - drug - prices - and - reduce pocket - of expenses - 7 ptions” 42 CFR 1001.952. “Exce 8 Office of Information and Regulatory Affairs. . https://www.reginfo.gov/ 5 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

8 Milliman Client Report MEDICAID REBATES IN In Medicaid, pharmacy costs can be managed by either the state or a managed care organization (MCO). Ingredient co , such as national average drug acquisition cost sts can be based on a variety of benchmarks (NADAC), wholesale acquisition cost (WAC), and more. The net cost to the Medicaid program is the bates. Rebates are specified by the ingredient cost less any patient cost sharing (often zero), less re Medicaid Drug Rebate Program (MDRP), which requires manufacturers to have a national rebate agreement with HHS for their products to be covered by Medicaid. The rebate amount is calculated as follows: differently for certain medicat ion types, but the unit rebate amount (URA) is generally calculated  Basic Rebate Amount: Equal to the greater of 1) 23.1% of average manufacturer price (AMP), etail setting, where AMP is the average price to manufacturers by wholesalers or pharmacies in a r or 2) the difference between AMP and Best Price, which is the lowest price available to any purchaser or insurer in most markets, with some exclusions such as the Part D market.  as a penalty for manufacturers The additional rebate amount serves Additional Rebate Amount: who have increased prices at a rate greater than inflation. It is calculated using a comparison of AMP and a Consumer Price Index (CPI) value in the current period relative to the period in which ed. the product was first launch The basic and additional rebate amounts are combined to form the URA, which is capped at AMP. In Supplemental rebates can be paid . addition to the URA, most states also negotiate a supplemental rebate to either states (who share a portion with the federal government) or MCOs, unless state law prohibits MCOs from collecting rebates. The supplemental rebate is typically provided to the entity responsible for managing the preferred drug list (PDL), in exchange for PDL status and / or lack of prior auth orization or other utilization management criteria. 9 The total value of Medicaid rebates is substantial, with about $24 billion in rebates paid in fiscal year 2015 . A 2015 study by the Office of Inspector General (OIG) found Medicaid rebates medications for top brand 10 were substantially higher than rebates for those same products in Medicare Part D . While this was based on 2012 data and Part D rebates have grown significantly since then, Medicaid rebates have also presumably grown significantly during this period, particularly because of the inflationary component of the URA noted above. In fact, the OIG study noted more than half of Medicaid rebates for the brand studied were attributed to the inflationary component and brand prices have continued to grow (in some cases substantially) since 2012. 9 Medicaid and CHIP Payment and Access Commission. “Policy Options for Controlling Medicaid Spending on Prescriptio n Drugs.” - https://www.macpac.gov/wp - content/uploads/2017/09/Policy - - for Options Controlling - Medicaid - Spending - on - Prescription - Drugs.pdf. 10 - D Rebates by a Name Drugs Exceeded P art Department of Health and Human Services Office of Inspector General. “Medicaid Rebates for Brand - 13 - 03 - Substantial Margin” https://oig.hhs.gov/oei/reports/oei 00650.pdf. 6 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

9 Milliman Client Report RESULTS III. We analyzed the impact of Part D manufacturer rebates no longer being excluded from the Federal Kickback Statute. o compensate for the lost rebates, we assumed manufacturers would produce T - Anti equivalen t price concessions, whether through explicit reductions to list prices or through negotiated Assuming no behavioral impacts, the total net cost of the Medicare discounts reflected at the POS. marketplace. However, individual Part D program would remain unchanged from the current even stakeholder costs vary considerably and are markedly different from current contributions without behavior changes ( . will undoubtedly occur) though some changes REMOVING MANUFACTURER REBAT E S GOVERNMENT IMPACT OF below show Tab le 2 s how government costs for the Part D program c ould change over the 10 - year period s 2029 to rebates were removed and replaced with equivalent price concession if from 2020 manufacturer remain in place with no change in expected e assumed pharmacy rebates would W reflected at the POS. or move to the POS value . s out the The table break following components of government spending , which combined account for total funding of Part D government : National average reinsurance (NAR):  T he average projected portion of claims paid by the federal government in the catastrophic phase . This is equal to 80% of POS claims costs and is paid for all member types. National average direct subsidy (NADS):  subsidy paid from the federal direct he estimated T government to plans . This amount is paid prospectively to all Part D plan sponsors, and is risk - adjusted based on the health status of each plan’s actual enrolled population. T d by the federal he portion of total cost sharing pai  Low income cost sharing (LICS) subsidies: . specified benefit government for LI members - This is equal to the difference between the plan design and a nominal amount paid by the LI member. The amount paid by the LI member varies ally equal to a few dollars per prescription and by income level and institutional status, but is typic zero in the catastrophic phase. Low income premium subsidies (LIPS): T  he portion of total member premiums paid by the federal government based on a me on a sliding scale varies also The magnitude of LIPS mber’s . income level and institutional status. . he sum of NAR, NADS, LICS, and LIPS T Total government costs:  Table 2 ASPE 1 for Manufacturer Rebates Government Cost Changes - Impact of Removing Safe Harbor Protections 2020 to 2029 (Billions of Dollars) Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact NAR $103.1 - - - - $8.6 - $9.2 - $9.9 $10.4 $8.0 $11.2 - $11.9 - $12.8 - $13.8 $7.3 - - 215.4 NADS 25.1 23.4 29.0 27.0 15.1 16.4 17.8 21.9 19.1 20.6 89.5 LICS - - 6.3 - 7.0 - 7.7 - 8.4 - 9.2 - 10.0 - 10.8 - 11.8 - 12.6 - 5.8 LIPS 12.0 1.3 0.9 1.0 1.1 1.1 1.4 1.5 1.6 0.8 1.2 4.2 34.8 3.8 3.6 3.5 4.0 Government Costs 2.8 3.0 3.2 3.3 3.5 1 . rebates are replaced with POS price concessions resulting in the same total cost net of rebates manufacturer Assumes 7 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

10 Milliman Client Report These results assume no corresponding behavioral changes. While we believe it very unlikely that no stakeholder would modify behaviors if a major change like this took place, we initially present the results s manufacturer without behavioral changes to demonstrate the impact of the elimination of rebates alone. The impact of potential b ehavioral changes are presented in the follow ing section of the report. Overall, if manufacturer rebates were removed, we estimate government costs c ould in crease , on average , by about $3. billion per year between 2020 and 2029 . This assumes manufacturer rebates are replaced 5 with equivalent price c oncession s producing the same total net cost . In other words, the POS cost underlying the estimates in Table 2 is equal to the POS cost in the current environment less the full value of manufacturer rebates. Pharmacy rebates are assumed to still exist at t he same levels as in the baseline , functioning as DIR . Part D on government impacts : ing fund Table 2 illustrates the following The NAR and LICS value s  steadily decrease throughout the 10 - year period due to the decreased tastrophic phase of the POS costs. The lower POS costs cause fewer members to reach the ca , benefit lower than in the current rebate reinsurance amount for each claim is and if they do, the This reduces NAR costs to the government by over $100 billion over ten years . LICS environment. is reduced as well , as the government - funded portion of cost sharing would be calculated as a percentage of a smaller POS cost . year 10 Offsetting the lower NAR and LICS costs, NADS and LIPS both increase over the  - timeframe. , rebates environment In the current function of the total plan liability in Part D. NADS is a s are used as a tool to keep plan cost price to converted low. If manufacturer rebates are concessions producing lower costs , the plan liability increases because rebate s would be POS shared with all stakeholders involved in paying POS claims (members, manufacturers, and the government), meaning fewer rebate dollars can be used to offset plan claim liability. This causes a stead il y growing NADS funded by the government. 35  Sin ce 2012, NADS has decreas ed by over $ PMPM , while NAR has increased by over $41 . and immediate C rebates would produce a stark manufacturer hanging the treatment of reversal in this pattern , with the government funding a greater portion of Part D costs through direct subsidy payments for all members and a smaller portion of Part D costs through reinsurance for high cost members . - We note the government may also be impacted by changes in risk corridor payments to the extent plan profitability is impacted by this change. Rebates are particularly impactful in the risk corridor calculations and as such, the elimination or reduction of rebates may lead to unpredictable impacts at the individual pla n level. Modeling plan level impacts was outside the scope of our analysis. The appendices at the end of this report provide further detail on these results, including the overall rebates are removed. projected costs in the baseline and after manufacturer RER IMPACTS OF REMOV ING MANUFACTURER REBATES MEMBER AND MANUFACTU If manufacturer rebates were replaced with equivalent POS price concessions, on average b oth members and manufacturers would have decreased funding responsibilities to offset the governmen t’s larger cost the burden. Table s 3 illustrate s to estimated impacts the following member and manufacturer funded elements of the Part D program :  Member cost sharing : T he portion of total cost sharing paid out - of - pocket by members (exclusive of LICS ) of pocket by members - he portion of total member premiums paid out T -  Member premium: (exclusive of ) LIPS  he sum of member cost sharing and member premium T Total member costs: 8 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

11 Milliman Client Report  Coverage gap discount program (CGDP): T he portion of claims paid by pharmaceutical manufacturers for NLI members in the coverage gap phase of the Part D benefit Table 3 ASPE 1 for Manufacturer Rebates Cost Changes – Member / Impact of Removing Safe Harbor Protections Manufacturer 2020 to 2029 (Billions of Dollars) Ten Year 2026 2021 2022 2023 2024 2027 2028 2029 2020 Impact 2025 - $40.8 - - $2.8 - $3.0 - $3.3 - $3.6 - $3.9 Member Cost Sharing $4.2 - $4.4 - $4.8 - $5.2 - $5.6 26.4 2.0 2.2 2.3 2.5 2.7 1.8 2.9 3.1 3.3 3.6 Member Premium - 14.5 - 1.0 1.1 - 1.2 - 1.3 - 1.4 - 1.5 - 1.6 - 1.7 - 1.8 - 2.0 Total Member Costs - 20.6 - 2.1 2.2 - 2.2 - 2.1 CGDP - 1.9 - 2.0 - 2.0 - - 2.0 - 2.1 - 2.1 - 1 manufacturer rebates are replaced with POS price concessions resulting in the same total cost net of rebates . Assumes If manufacturer rebates are removed, total member costs and manufacturer claim costs are billion, respectively, over the 10 expected to decrease by $14 billion and $21 - year period. As with Table 2, these results assume no corresponding behavioral changes. While we believe it likely that stakeholder behaviors would change in some way , we initially present the results without behavioral rebates alone. The impacts including changes to demonstrate the impact of eliminating manufacturer potential behavioral chan ges are presented in the following section of the report. lower on average, due to a  Total member costs are expected to be about $ 1.4 billion per year Note reduction in average cost sharing, only partially offset by an increase in premiums. that the impact will vary for members taking different types of products, with the greatest savings associated with members taking very high - cost products subject to coinsurance, while cos ts through members taking low - cost or no medications may have an increase in total . s higher premium members enrolled — These results assume a defined standard benefit design in enhanced plans or employer group waiver plans (EGWPs) may not see as much savings if their benefit design uses copayments rather than coinsurance, though high utilizing members would still - see meaningful savings as nearly all plans use defined standard coinsurance on brands in the gap . and catastrophic phases expected 2020 plan year m  s (the portion paid by members, exclusive of LIPS) are ember premium rebates to be $3. 2 0 higher per month as a result of the replacement of manufacturer with POS over 5% annually to We project this premium to increase by 4% year - . 10 the price concessions member PMPM, relative to a projected 2029 premium of $46.20 projection. This equates to a 2029 . The member member premium of $41.30 PMPM assuming no changes to manufacturer rebates premium increase be the largest challenge with implementing this change since it is could bers focus on when selecting their plan annually a highly visible feature of the plan that mem most and impacts members (except for those with fully subsidized premiums). However, as have a decrease in premium as a result actually demonstrated later in this report, some plans could of this change.  Manufacturer payments through the CGDP are expected to be about $2 billion per year lower on average. When POS costs are lower, it takes longer for costs to accumulate toward the initial lower as fewer members would be coverage limit (ICL) and the TrOOP threshold. CGDP payments and because CGDP reach the coverage gap or spend less time in the gap before the year ends, would be calculated as a percentage of a lower POS cost, but the impact would vary by and by plan sponsor manufacturer . 9 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

12 Milliman Client Report As noted above, the overall average impact to members is a small increase in premium and larger decrease in cost sharing, but the results would vary for different members in different plans. While all members in a PS), different plans will have different given Part D plan pay the same premium (not accounting for LI changes in premium if manufacturer rebates are removed. Changes in premium would vary based on the following:  The change in premium will be directly correlated to the current magnitude of manufacturer rebates. rebate levels may experience the highest manufacturer sponsors with very high Plan (presumably market These plans increases in premiums. , leading plans with high enrollment and - therefore high negotiation leverage) would have the largest reduction in POS costs, causing the , largest increase in plan costs because fewer rebate dollars can be used to offset plan claims liability. However, because all plans receive the same direct subsidy amoun t (prior to risk adjustment), the NADS would not be enough to compensate for the increase in plan cost, resulting in a higher than average premium increase.  On the other hand, plan sponsors with lower than average rebate levels may in fact experience a d ecrease in premiums. Similar to the effect described above, these plans would not have as large of a change in net plan liability, but they would receive the same NADS, which would be higher to reflect the average change in bid amounts across all plans. As a result, the increased NADS may more than offset the increase in bid amount for these plans, causing a decrease in premiums.  In line with the dynamics described above, PDPs may have higher premium increases than MAPD plans, s inc e PDPs tend to have highe r manufacturer rebate levels than MAPD plans. This could MAPD plans relative to PDPs . cause a higher trends for enrollment n acceleration in the  Shifts in enrollment could cause further changes in plan premiums over time. Members may seek continue negotiating the lowest cost sharing available for the products they use. If plan sponsors (as opposed to manufacturers uniformly reducing list price concessions distinctly for themselves - prices), members taking high st price concessions. cost products may shift to plans with the highe This adverse selection could cause premiums to increase in the following year. The results presented to this point represent no change in total program costs or in stakeholder  behaviors. Plan behavioral changes would certainly future premiums, as discussed in the influe nce next section. Member cost sharing would also vary in different situations, including but not limited to the following considerations:  Assuming price concessions are reflected at the POS at a product level or aggre gated across products in the same therapeutic class or (as opposed to aggregated across all rebated products manufacturer across all products), only members taking a product with a rebate would see cost sharing savings. A majority of brand utilization is f or products with manufacturer rebates, and about Part NLI of D members used a brand medication in 2017. embers not m All else constant, 4 0 % , however, reductions taking any rebated medications would have no change in cost sharing to cost sharing for generics may be needed to meet actuarial equivalence testing, which . utilizers - members other than non NLI could produce savings for all In particular,  be more likely to save than others. may Members with certain benefit designs save may members with a deductible and a coinsurance benefit during the initial coverage phase Members with copayments . a greater percentage of costs than members with other benefit designs the copayment remains the same and does not may not see cost sharing savings if the value of , though the value of the copayment may in fact change exceed the new POS cost of the medication due to actuarial equivalence requirements in Part D . However, a large portion of brand utilizers gap or catastrophic phases, during which nearly all plans have a coinsurance reach the coverage design. As a result, members with copayments during the initial coverage phase could still have some cost sharing savings if they reach the gap and catastrophic phases. 10 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

13 Milliman Client Report  Members in plans with the highest premium increases might have the highest offsetting cost – as noted above, plans with the highest sharing savings rebate levels may have the manufacturer largest premium increases, but those high rebate levels would produc e greater cost sharing savings if the rebate is converted to a lower POS cost. or most LI members would not be significantly impacted by this (  other) changes to Part D costs because their cost sharing and premiums are already highly subsidized by the gove rnment. However, to the extent some LI members pay a percentage of cost sharing and / or premium, they would experience similar impacts as NLI members, but the dollar change would be smaller. cost sharing and premium, some members When taking into account total member costs, including both would have savings and some members would have increases in costs. The overall average impact shown is a savings to total member costs, though the distribution of changes to member costs is far in Table 3 from uniform the “break even” point With that said, no members would have an increase in cost sharing, so . decreasing is equal to the total annual increase in premium. of member spending increasing or T he estimated average premium increase in 2020 (exclusive of LIPS) is $3.20 PMPM. As such, if manufacturer rebates were eliminated, a member with a reduction in cost sharing of more than $3.20 money from this change per month ( about $38 per year ) , would save . could brand utilizers any M The average projected gross cost . or much more see this level of savings specialty products and over $800 for brand and specialty per brand prescription in 2020 is over $400 for non - products combined. A member taking a $400 brand with a 30% manufacturer rebate would save $120 if hey were in the deductible phase, and would continue to save $30 per month while in the initial coverage t phase. given these amounts are LI members are largely shielded from changes to both cost sharing and premium, heavily subsidized by the government. Only partial subsidy LI members, which we estimate account for see direct cost sharing savings (these members have an about 15% of all LI members, could meaningful $85 deductible and 15% coinsurance Other LI members would have minimal impact up to TrOOP in 2019). subsidized. because their cost sharing and premiums are already highly As such, the impact of this change is concentrated among NLI members, who account for about 70 % of Based on actual 2017 Part D claim records for about 4 million NLI lives, we estimate total Part D enroll ment. ( if manufacturer rebates are removed NLI members would be affected as follows though the actual results will vary depending on benefit design, rebate levels, and pre mium increases ): According to Milliman manual rate data, about 7% of NLI members do not use any medications.  , These members have no cost sharing change in premium would only be affected by the , and thus (which would increase on average, but . by plan) significantly c ould vary could  directly benefit from the changes in POS costs by enough to Up to 30% of NLI members increase in premium. average make up for the  Of the remaining 63% of NLI members, total costs may increase or decrease. All else constant, these members generally do not have enough cost sharing savings to fully offset the increase in premium. However, even members not taking any brand medications could experience copayment rands decreases, the overall reductions as plans alter their benefit designs. As the POS cost of b value of the benefit will decrease, and plans may need to reduce cost sharing to maintain the As a result, generic or other copays minimum required value of 25% in the initial coverage phase. . savings for these members, but is difficult to quantify may decrease, which could generate 11 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

14 Milliman Client Report BEHAVIORAL EFFECTS The results in Table s 2 and 3 reflect the and 10 - year estimated impact of eliminating manufacturer rebates list replacing them with commensurate POS price reductions (whether through lower prices or negotiated While it is difficult to predict concessions with separate plan sponsors), but assumes no other changes. likely what other kinds of changes may result from this potential change , all stakeholders would change behavior as a result of this change , so it is critical to consider possible behavioral impacts as well. However, we do not intend to imply that any one outcome modeled is more likely than , and it is difficult to predict the magnitude of change in all scenarios another . scenarios: eight the impact to stakeholder costs from 2020 to 2029 under the following We modeled i  Scenario 1 – No Behavioral Changes : Th s is the same scenario presented in Table s 2 and 3 . PBMs and plans may favor less rich formulary : Increased Formulary Controls Scenario 2  – lower POS costs. rebates were removed from designs , instead focus ing more on manufacturer If costs after the POS. reduce the Part D program, plans would lose a large portion of their ability to Therefore, discount contracting (lowest POS costs) c ould become the main avenue to keep costs profit While changes in plan costs may not directly affect plan to the plan as low as possible. , it is plan sponsors to keep costs low to maintain competitive premiums. could This important for e formulary designs that would move members to the lowest cost alternatives. These iz incentiv formulary design changes may include: ̶ uicker adoption of new generics Q ̶ H igher tier placement (or removal) of brands with generic equivalents ore robust prior authorization and step therapy programs for high cost brands ̶ M this s cenario , we assume plans will react to the elimination of manufacturer rebates by placing In stronger emphasis on formulary management, resulting in a gradual reduction in costs over time a relative to the baseline, brought about by a higher generic dispensing rate (GDR). We assumed GDR would increase the most in the second year of implementation , a ssuming plans may need additional time to formulate strategies and react to competitor changes in year one. We assumed continue to increase by a small amount would also four years. over GDR harmacy rebates would continue to exist as DIR and with GDRs ap proaching or exceeding 90%, P a large portion of these rebates are for generic medications . I ncreasing GDR has an added bonus for plan sponsors of not just reducing POS costs , but maintaining the same level of pharmacy rebates (as opposed to manufacturer rebates, which decrease when GDR increases because they lead are primarily for brand medications) , which c ould plan sponsors to pursue this type of strategy . Scenario 3 As formulary :  ssions Increased Formulary Controls and Increased Price Conce – price concessions to PBMs or plan sponsors additional control tightens, manufacturers may offer ould be which c greater than the equivalent rebate arrangements today even ( ) to keep their products rebates to on formulary. In the current supply chain, manufacturers can negotiate with PBMs using and the value of the rebate can sometimes outweigh any difference positioning maintain formulary . However, in a market without re bates, manufacturers in POS cost relative to competing products may need to negotiate larger price concessions with PBMs to secure of higher - cost products - formulary access or risk losing formulary status to a lower cost alternative . keep the net plan sponsor Manufacturers may be pushed to provide higher price concessions to net allowed the maintaining , rather than current to the current liability after rebates closer liability . In dollars, the value of the price concession would need to be greater than the after rebates cost . keep the plan sponsor net neutral value of the rebate to 12 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

15 Milliman Client Report Scenario 3 s upon Scenario 2 by assuming price concessions negotiated with manufacturers build gradually increase. We assumed brand costs would initially be 0.5% lower, with concessions will to a maximum difference of 2.5% less than baseline costs. This reaction further increasing over time may occur as plans fight to keep premiums down, but no longer can do so with manufacturer rebates, so they may gradually push manufacturers for deeper price concessions in order to intain formulary status. Scenario ma assume s GDR will still increase relative to the baseline due 3 to a shift in formulary strategies focused on lowest cost products, but smaller GDR assumes a . Relative to Scenario 2, brand may utilization increase than Scenario 2 be higher if more brands are covered due to larger . price concessions  Scenario 4 – Reduced Price Concessions : As mentioned above, Tables 2 and 3 (Scenario 1) assume the impact of the elimination of manufacturer rebates in Medicare Part D has no net impact new price on overall costs. However, in the transition between negotiations for rebates and the , manufacturers may take the opportunity to improve POS costs , but not to the point at concessions fully ma which it . In particular, this could happen if manufacturers kes up for the former rebate choose to reduce POS costs through lower list prices as opposed to negotiated price concessions with Part D plans. This is because list prices are the basis of manufacturer rei mbursement in all markets — if a manufacturer reduced list price by the full value of the Part D rebate on a per script basis, it would lose money due to lower reimbursement in the commercial and other markets. For example, assume a $100 list price, 30% Part D rebate, 50% Part D market share, 20% commercial rebate, and 50% commercial market share. The average net reimbursement for the manufacturer is $75 [= ($100 - $30) * 50% + ($100 - $20) * 50% ] . However, if th e manufacturer reduces the list price to $70 and rebates continue to exist in the commercial market, the new net $14) * 50% - $70 ( $0) * 50% + - ($70 [= reimbursement would only be $63 . In this example, the ] manufacturer would only be able to reduce the li st price to $83.33 for the overall average reimbursement to remain neutral at $70. Scenario 4 examines this possibility by assuming that manufacturers adjust prices to result in POS costs reduced by an amount equivalent to only 80% of the lost rebate rat her than 100%. Note that the 80% assumption is for illustrative purposes and may be plausible, though it is difficult to predict how much manufacturers would be able to reduce price concessions by given the pressures they This may face from plan sponsors. scenario, as with others, might occur in combination with other ) downstream behavioral changes increased focus on lower - cost generic medications (e.g., . However, we present this scenario to show how costs could increase if plans were unsuccessful in offset ting increases to net costs. scenario assume manufacturers will not s  Scenario 5 – Decreased Brand Unit Cost Trend : Th is quickly increase list prices as rebates. To the extent manufacturers manufacturer in a world without currently list price increases in reaction to plan strategies preferring high list implement high annual prices with high corresponding rebates, there may be less need for high future price increases. In particular, assuming formulary strategies evolve to favor lower POS cost products, m anufacturers may need to respond by limiting price increases over time. We assumed a 1% lower annual trend on brand medications, beginning in 2020. This could take the form of large drops in trend, or even negative trends, in competitive therapeutic classe s, combined with smaller or no changes in trend in other classes. Scenario 6 –  Increased Utilization and Decreased Brand Unit Cost Trend : In addition to the lower brand price trends assumed in Scenario 5, there may also be changes in utilization. The discounts will decrease member cost sharing at the POS transfer of manufacturer rebates to POS they are now more affordable given s brand embers may use more Therefore, m in many cases. . We kept the same price trends as in Scenario 5 , but assumed brand u tilization trends would be 0.5% greater in this scenario than in the baseline for NLI members only (LI members would not be influenced by the change because their cost sharing is already subsidized). Increases in brand utilization will increase total costs , all else constant, and partially offset savings generated by reduced brand price trends in this scenario. 13 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

16 Milliman Client Report : 7 – Increased Pharmacy Rebates This scenario assumes PBMs and plans will more  Scenario Pharmacy aggressively contract for pharmacy rebates in the absence of manufacturer rebates. would still have strong value through their treatment as DIR rebates there may be opportunities and to offset increases to member premium by negotiating for increased pharmacy rebates. the This scenario is similar to Scenario 3 in that both assume there will be pressure to provide ario 7 assumes alternative avenues to keep plan liability closer to its current levels. However, Scen this will increase through pharmacy rebates (which we assume will remain as DIR and not at the POS rather than manufacturer price concessions at the POS . ) Table 4 below shows the estimated 10 - year impact in billions of dollars of the scenarios descr ibed above on each stakeholder. The following appendices detail the 10 year projections across the same categories: -  Appendix A 1 details the annual change for each scenario if manufacturer rebates are removed , in billions of dollars x A2 details the same annual change as Appendix A1, but on a PMPM basis Appendi   1 details the total annual costs of each scenario Appendix B if manufacturer rebates are removed , in billions of dollars Appendix B2 details the same annual costs as Appendix B1, but on a PMPM basis   Appendix C details the percentage change relative to the baseline scenario if manufacturer rebates are removed Table 4 ASPE Impact of Removing Manufacturer Rebates – Including Behavioral Changes 2020 to 2029 (Billions of Dollars) Total Total Total Member Member Gov’t Program Cost Member Costs* Costs* Costs* NADS LIPS Premium Sharing CGDP Scenario NAR LICS $34.8 $0 $14.5 - - $26.4 $20.6 - $103.1 $40.8 $215.4 - $89.5 $12.0 - Scenario 1 - 164.6 - 78.8 - 56.2 118.3 64.6 8.4 - 139.1 174.7 - - 3.8 - 29.5 Scenario 2 188.2 - - 59.5 - 99.6 - 63.5 - 163.2 180.3 29.1 118.5 1.9 - 4.0 - Scenario 3 135.1 139.9 12.3 - 32.6 44.9 17.1 - - 30.2 221.1 - 71.4 20.5 Scenario 4 - 38.4 - 94.3 - 33.1 6.1 - 101.3 - - 46.5 13.3 22.8 - 154.6 211.4 Scenario 5 - 22.7 69.9 - 26.0 - - 7.6 - 42.7 16.7 - 21.2 - 143.8 214.8 101.3 Scenario 6 20.9 - 17.8 - 18.1 10.4 - 40.8 2 2 .8 - 20.6 - 11 0 . 2 20 7 . 1 - 89.5 Scenario 7 *Totals may not sum to individual components due to rounding. Depending on the type and degree of behavioral changes, . increase or decrease total program costs could  The results in Scenario 2 show that tighter formularies and increased generic utilization could cause 7 9 year period from 2020 to 2029. While the - 10 the government to save about $ billion in the direct subsidy cause s an increase in total government costs in Scenario 1, the direct increase in subsidy is not as high in Scenario 2, and reinsurance and LICS savings are even greater, contributing to the overall government savings. Both member and government savings increase further in Scenario 3, in which additional price  year projection period. concessions could bring down POS costs by $70 billion over the 10 - or more ions, through increased price concess Manufacturers would fund the majority of this cost reduction 4. s though this would be partially offset by the savings in CGDP seen in Table 14 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

17 Milliman Client Report an opposite impact, with total program costs increasing because total price has Scenario 4  concessions are lower than in the baseline. This would cause costs for all stak eholders to increase. on average because the price concession, while Members would still have cost sharing savings lower than in the baseline, is now reflected at the POS when in the baseline it was not, and some members would still have overall savings. H owever, on average across all members, the increase in premium would outweigh the cost sharing savings. Scenarios 2 and 3, but it is still close to double the  less than member savings is In Scenario 5, to estimated otal government costs are still T . impact without any behavioral changes (Scenario 1) decrease rather than increase. The lower rate of brand price increases lower total produces program cost and the increases in member premium relative to the baseline shrink over time. s  In Scenario 6, total savi ngs is less than Scenario 5, with increases in brand utilization offset by decreases in price trends. The assumption of increased brand utilization is in direct contrast to , which assumed less brand utilization he elimination that if t Scenarios 2 and 3 . This illustrates of manufacturer rebates corresponds with lower increases in brand prices over time, savings government may be generated even if formularies do not change.  The basis of Scenario 7 is similar to Scenario 3, assuming plans are able to gain increased price concessions from pharmacies. We assumed pharmacy rebates would increase but utilization would . not change one woul member claims not impacted by rebates, With medication choices d expect may not change. As a result, claim payments (member cost sharing, LICS, CGDP) do not change relative to Scenario 1. While total costs are lower than in Scenario 1, government costs still increase pharmacy eline, $8.6 billion relative to the bas not great enough to offset the loss of DIR is as manufacturer rebates. OTHER CHANGES There certainly could be numerous other behavioral changes across different stakeholders and further nuances surrounding the effects of the potential ch anges described here. We did not attempt to capture every possible combination of changes, but illustrate potential effec ts. Enrollment In addition to the changes modeled in Table 4, the re could also be significant changes in enrollment. Premium is a very important factor for members shopping for a Part D plan, so if premiums increase, This could result in a large number of members members may choose to “shop around” for a new plan. shopping for a new plan, especially considering the highe st concentration of members are in the largest plans, which presumably have the largest manufacturer rebate levels (and , therefore , may have the largest premium increases). This could result in the following types of enrollment sh ifts: .  Members shifting to plans with leaner coverage Typically, plans with tighter formularies or less the total as in the enhanced benefit designs have lower premiums, ote cost of coverage is lower. N um than a leaner current environment, high rebates could allow a richer plan to have lower premi largely plan with low rebates , but this dynamic would be eliminated if manufacturer rebates were . If members react to premium increases by moving to leaner plans, total program costs removed ward lower cost medications. steered to potentially This option would may decrease with members two be largely limited to members enrolled in enhanced plan designs, which account for about of all thirds members. Medicare Advantage and stand - alone Part D increase, or even decrease, in premium. While  Members shifting to similar plans with a smaller manufacturer premiums are expected to increase on average, plans with lower than average rebate levels may in fact have lower premiums because the plan bid amount increases by a average bid , but the plan still collects the higher national the national smaller amount than potentially even the playing field for Part D premiums and This could . average direct subsidy enrollment, such that smaller carriers with less negotiation leverage may now have premiums o larger carriers closer t . them to gain more membership , allowing 15 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

18 Milliman Client Report - enrolled LI members. Basic plans with premium below the low income benchmark  Shifts in auto LIB ) in a given region ( are eligible for auto - enrolled LI members. The LIB is calculated from the average bids across all plans in the region, weighted by LI membership. A plan with premium not currently below the LIB typically has fewer LI members and might not face as large of a premium rebates are increase , which could cause the plan’s premium to be below the LIB if manufacturer - enrollment, particularly in the first year This could result in large shifts in LI auto eliminated. of the change .  Continued enrollment shifts over t ime. The elimination of manufacturer rebates would be a continue to adjust strategies over time. As significant change in Part D, and we expect plans to such, there could be swings in plan premiums not just in the first year of implementation, but for seve ral years to follow. Some plans may take longer than others to find ways to offset the increases in premiums and some plans might choose to wait and see how their competitors react to the change. As noted above, members may initially shift to plans with lo wer premium increases or premium decreases. However, if this was caused by lower manufacturer rebates, those plans may still have less negotiatin g leverage , and thus , may lose ground over time as larger carriers potentially negotiate greater price concessions in the new environment. As a result, any enrollment shifts away from current market leaders may only be temporary, unless the shift is significant shift leverage from the curren enough to change negotiation dynamics and t market leaders to other carriers. Benefit Design Benefit designs may also change if rebates are eliminated. Enhanced plans may be less likely manufacturer to enhance the benefit for brand tiers than for generic tiers if their focus is strictly on achie ving the lowest POS cost. As described previously, all plans are subject to actuarial equivalence testing, which requires before reaching the ICL. If the POS cost to pay 25% or less of costs (on average across all tiers) members for brands decreases, a cop ayment will become a larger percentage of costs, which may cause a plan to a fail the equivalence test unless it reduces cost sharing on that tier or another tier. To illustrate, consider design and average costs the following benefit plan with : ) tier formulary, for simplicity - hree t (assuming a Table 5 ASPE 1 Impact of Removing Manufacturer Rebates Changes to Actuarial Equivalence Testing Remove Manufacturer Rebates Baseline 2 2 Generic Brand Specialty Total Total Generic Brand Specialty $23 $25 $85 $5 $625 $500 Member Cost Sharing $5 $85 $102 $85 Allowed Cost per Script $20 $350 $2,500 $2,000 $20 $280 27 % % 25 25% 25% Actuarial Value 30% 25% 24% 25% 1 Brand and specialty manufacturer rebates assumed to be replaced by 20% POS price concessions in this example . 2 Specialty cost sharing assumed to be 25% coinsurance . In this example, the plan offers a $5 copay on generics and an $85 copay on brands, both of which are equivalent to approximately 25% coinsurance (a 25% coinsurance applies to specialty medications as well). specialty POS costs are assumed to decrease by When manufacturer rebates are removed, brand and 20%. The $85 brand copay now has an effective value of 30%, and the total effective coinsurance increases from 25% to 27%. The plan would need to reduce cost sharing on at least one tier to pass the actuarial equivalence test requiring 25% or lower cost sharing. This could be achieved in this case by reducing the generic copay from $5 to $3.25 or reducing the brand copay from $85 to $70, among other options. 16 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

19 Milliman Client Report - specific benefit design chang In addition to plan - es, it is possible the CMS defined Part D benefit parameters (deductible, ICL, TrOOP) may also change as a result of removing manufacturer rebates. These parameters - are set each year using an annual percentage increase (API) methodology, reflective of year year - over Part D costs used in this approach are changes in average Part D costs observed in actual claims data. so it is possible there could be a one - gross of rebates, time drop in the Part D benefit parameters in the year following the elimination of m anufacturer rebates. As POS costs decrease, there could be a similar percentage decrease in the deductible, ICL, and TrOOP when those lower costs are compared to the prior a different methodology is currently used for calculating TrOOP, but th , is is expected to revert year. (Note to the API approach beginning in 2020.) retroactively back The impact to each component of Part D spending could change if the Part D benefit parameters change ater average member savings and would in this way. A reduction in the Part D deductible would result in gre result in a larger number of members embers would However, we estimate m . with cost sharing savings have larger average premium increases would increase, largely due to a lower government costs and TrOOP decreases, members will reach the catastrophic phase more quickly and the . I f TrOOP amount government will pay more claims through federal reinsurance. This may nearly offset the savings in federal the change in benefit parameters. reinsurance related to lower POS costs, depending on the magnitude of 17 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

20 Milliman Client Report IV. CONSIDERATIONS manufacturer There are many possible outcomes if rebates are eliminated. We do not attempt to discuss all possible consequences , but provide discussion in this section of important considerations and pot ential impacts. LAN TYPES DIFFERENCES ACROSS P Different plans will likely have different strategies in reaction to the elimination of manufacturer rebates. There is variability among formulary strategies today, and some formularies do already target the lowest not be uniform across all plans. will POS costs in most or all classes, so changes in formulary strategies It is difficult to predict how plans will react, but we observe the following potential distinctions among different types of plans: :  MAPD / PDP MAPD plans on average have lower rebate levels than PDP plans, according to Milliman Part D bid survey results. As a result, MAPD plans may not experience as large of client an increase in bid amount, which would lead to a lower than average increase in premium (or even a decrease in premium). In the last several years, there has been a shift in enrollment from PDP to MAPD plans, with MAPD enrollment growing nearly 7% from 2017 to 2018, while PDP enrollment grew only 1%. This movement from PDP to MAPD could increase if PDP premium increases drive members to plans often able to offset increases in Part D costs with lower increases. Additionally , MAPD plans are through benefit changes on the Part C side as well, which may be less visible or less important to some members. ower average Similarly, plans with a large proportion of NLI members may have l : NLI / LI  manufacturer rebate levels than plans with a large proportion of LI members. This is because LI members on average use a greater proportion of brand rebated medications. As a result, plans with a large proportion of LI members may face greater premium increases. This could cause an interesting dynamic with the LIBs. Basic plans with premium below the LIB are eligible for auto - enrolled LI members, which is an easy way to gain volume and is often a strategy used by basic PDPs. The LIB is calculated from the average bids across all plans in the region, weighted by LI me mbership. A plan with premium not currently below the LIB typically has , his could cause T . might not face as large of a premium increase therefore fewer LI members and , the the plan’s premium to be below the LIB once manufacturer . Across rebates are eliminated market, it could enrollment, particularly in the first year of the - result in large shifts in LI auto change. /  EGWP , we modeled a national average Individual : In the results presented in this report EGWP and Individual plan types. However, population, assuming total enrollment across both EGWP plans have unique benefit designs. Average EGWP member spending can be greater than , including greater utilization of brands driven average individual member spending (all else equal) by richer cost sharing. As a result, an EGWP with strong manufacturer rebate levels might see a large increase in premium and could possibly react by reducing the level of benefit enhancements sly. or increasing retiree contributions in addition to formulary strategies discussed previou have limited ability to offset any premium increases through benefit / Basic :  Basic plans Enhanced they are required to offer a plan with the same average value as the defined standard as reductions, ssibly make benefits less rich to offset premium plan. Enhanced plans, on the other hand, could po increases, though they would still be required to meet meaningful difference requirements relative to the carrier’s basic plans. 18 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

21 Milliman Client Report To further illustrate the variability by plan type, Table 6 below shows our average estimated member premium in 2020 for the following hypothetical Part D plans:  all This plan is assumed to have claim costs, rebates, income mix, and : National average plan tions underlying our other assumptions equal to the national average. These are the same assump national average impacts in the tables throughout this report. 100% NLI plan In this case, we applied the same assumptions as for the national average plan,  : LI members, we While most plans have some mix of NLI and but assuming 100% NLI enrollment. provide this sample plan for illustration.  100% LI plan : Similarly, this plan has the same assumptions as for the national average plan, but assuming 100% LI enrollment.  - ome mix as the national This example assumes the same claim costs and inc : Low rebate plan average plan, but assumes the plan has lower than average rebate levels (15% of total allowed costs). Table 6 ASPE Impact of Removing Manufacturer Rebates Monthly Premium Across Sample Plan Types Baseline Manufacturer Rebates Remove $39 $34 National average plan $37 100% NLI plan $34 100% LI plan $30 $45 Low - rebate plan $60 $51 rebates could have a decrease in premium As seen in Table 6, manufacturer a plan with lower than average of about $ 9 PMPM . Based on the assumptions underlying this illustration, the low - rebate plan still has higher premium than the other plans, but it is a much more level playing field relative to the baseline, in highest plan. which the low - rebate plan premium was at least $20 higher than the next In the examples in Table 6, average premiums for a 100% NLI plan are higher than the national average in rebates are eliminated and transferred to POS the baseline, but decrease by about $3 when manufacturer price concessions. On the other premium increases by — hand, a 100% LI plan has the opposite impact 1 about $ such that the LI plan now has higher premium than the national average plan. Note this income 5 , mix effect assumes no change to the risk adjustment model, though as discussed below is likely a change to be needed. RISK ADJUSTMENT Risk scores may become more important to plans if manufacturer rebates are eliminated. The direct subsidy - paid to plans is risk adjusted based on the average risk score across the members enrolled in the p lan. In Tables 4 above, the national average direct subsidy is expected to increase by $175 billion or more. As the direct subsidy increases, the value of the multiplicative risk score adjustment increases as well. MAPD plans may increase focus on risk sco re coding efforts to capture as many diagnoses as possible and for PDP plans though, which could possibly lead PDPs to push maximize risk scores. This is not as easy greater control over risk scores or the inclusion of pharmacy data in the Part D risk score model. for rebates were eliminated. manufacturer The CMS risk adjustment model would likely need to be updated if Currently, the risk adjustment model is calibrated to gross Part D costs, prior to accounting for rebates. dramatically rebates, which could manufacturer POS costs would now reflect change how certain conditions 19 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

22 Milliman Client Report - It is possible are high evaluated. rebate therapeutic classes currently profitable for plan sponsors cou ld receive a reduction in the risk adjustment factor for the associated condition. REBATES THE REIMBURSEMENT MODELS IN ABSENCE OF MANUFACTURER If manufacturer rebates are eliminated completely, manufacturers may take some other kind of action to et costs in Part D. There are two primary mechanisms by which this could be done: reduce n As discussed previously, incentives exist today to encourage formulary  Reducing List Prices : - and high rebates over lower list prices preference of products with high priced products. Without rebates, these incentives would be removed. Further, we estimate manufacturer manufacturer rebates account for as much as 30% of gross costs for brand medications, so without rebates, o would presumably be able t significantly manufacturers reduce list prices. However, this has implications for other markets, as discussed previously, so manufacturers may not necessarily reduce list prices by the full amount of average rebates in the Part D market, particul arly if rebates continue to exist in the commercial or other market s . market function Negotiated Price Concessions : O ne possible way the could  without rebates would be for the manufacturer to negotiate directly with PBMs or plan sponsors for an additional he wholesaler would still price concession reflected at the POS. In terms of the supply chain flow , t relative to discount a purchase products from manufacturers at and sell to pharmacies the list price prescription, the pharmacy would collect the copayment and be a dispensing Upon on a spread. POS final reimbursed by the PBM for the by negotiated price concessions cost after accounting for the PBM with both the pharmacy and manufacturer. For each Part D claim , pharmacies c ould request a “ chargeback ” payment to be rei mbursed for the differential between the manufacturer’s price concession and pharmacy ’s price concession to make the pharmac y whole . key difference between these two options is that negotiated price concessions will result in Note one variable pricing across different plan sponsors. In particular, plans with the lowest POS costs may continue to be plans with large volume / negotiating power (tho ugh these are the plans who will also need to make up the most ground and will face the largest premium increases if manufacturer rebates are eliminated , as acturers discussed previously ). Plans with small enrollment and little negotiation leverage may prefer manuf list prices reduce potentially result in a lower POS cost for them and make enrollment . gains more likely to Of the two reimbursement options outlined above, reducing list prices is the simplest in that it requires fewer changes to supply chain stakeholders and less administrative burden. The chargeback model would involve administrative and logistical challenges and would require a significant change to the supply chain to take place in a short period of time. Additionally, the chargeback payment introduces a new step in the supply chain that could bring along an additional fee related to administering the progra m. Further, pharmacies may have a cash flow issue depending on the timing of payments from the manufacturer to true up to the manufacturer negotiated PBM - . The impact of this cash flow issue is dependent on how price concessions frequently the chargeback However, the chargeback model results in more similar net yment is made. pa prices as the current environment , with POS costs varying by plan, so some plans may push for maintaining this type of structure. APPLICATION OF PRICE CONCESSIONS rebates The CMS RFI on reflecting a portion of rebates at the POS suggested reflecting POS manufacturer using an average across all rebated products in a given class - specific level. rather than at a product For his classes with multiple rebated products, t approach has the advantage of reducing member cost sharing while still protecting proprietary contracting information associated with a single product. However, spreading rebates across the entire class would result in one manufacturer subsidizing POS subsidization could make manufacturers - sts for another if the rebate values are not equal. This cross co . less willing to give strong concessions Manufacturers offering high rebates today may be opposed to spreading rebates at any level other than product cific, while other manufacturers could be in favor of spe - 20 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

23 Milliman Client Report Additionally, contracting terms may not be this approach to the extent it reduces costs for their products. protected if there is only one brand in a given therapeutic class. Spreading rebates across a therapeutic class may be less relevant to the concept of eliminating through, as proposed in the CMS RFI) manufacturer rebates altogether (as opposed to partial rebate pass - s inc ey are today. However, e it hinges on rebate contracts continuing to be contracted in the same way th if rebates are eliminated completely, they could be transferred to lower POS costs through manufacturer w or through the chargeback model described above. While specific, - be product lower list prices, which ould the price concessions in this model could in theory be spread by plan sponsors, manufacturers may This challenge this approach and could word their contracts to prohibit applying rebates to other products. would both eliminate the gap between POS and net costs and increase transparency on product - and level pricing. - plan - Product level POS price concessions would be particularly transparent in the Part D market because the D CMS Plan Finder tool, available online, provides easy access to information on POS prices for every Part Price t has the advantage of improving clarity on medication costs for ransparency plan in the market. reveal previously confidential information as well as the general public. However, it would also beneficiaries , — if a competitor’s on manufacturer contracting. This could possibly lead manufacturers to outbid each other price concession is known, a manufacturer can offer a slightly higher concession to gain formulary status — and could also give Part D plans, or even commercial plans, to push for a known rebate level offered to another plan. These additional pressures could lead to lower total costs. Alternatively, transparency could ake manufacturers less likely to want to weaken manufacturers’ bargaining power and m negotiate Manufacturers with high rebates today would likely not want their contract terms aggressive contracts. exposed such that others could view those terms and take them into account in their own contracting. EGOTIATIONS IMPACT TO CONTRACT N Plans and manufacturers may initially have competing incentives if . rebates are eliminated manufacturer P rebate strategies to focus simply on lowest cost - price / high - lans would shift focus away from high products. This allows for better alignment with manufacturers wishing to compete on list price alone. lans and manufacturers will naturally both be motivated However, during the initial shift away from rebates, p to remain whole relative to prior years. For the manufacturer, this means the POS cost would be equal to the net cost (POS cost less rebates) e e plan’s point of view, this would not produc rebates were in existence. But from th manufacturer when neutral financials. Instead, plan liability would increase if the same magnitude of rebates were translated to POS discounts (whether through reduced onsors or POS price concessions). As a result, plan sp list prices may push for stronger price concessions to keep their net liability neutral, rather than only keeping the net total cost neutral. On the other hand, manufacturers may be less willing to negotiate the same rebate levels. Some may manufacturers might ch oose to reduce list prices to benefit all markets, not just Part D and , therefore , not be willing to give additional concessions in Part D. As mentioned above, how price concessions are shared is an important factor. If price concessions are spread across multiple products, manufacturers may nts s inc e they would be subsidizing other products. Also, as be less willing to give additional discou negotiated rates become more transparent, it is possible manufacturers are less willing to give strong discounts, which could reveal previously confidential information and possibly imply strateg ies used in other markets where contracts are still confidential. This could be particularly true for manufacturers offering aggressive rebate levels today. Alternatively, transparency could cause manufacturers to concede additional discounts, since w ould presumably they know a certain level is necessary to beat a competitor’s price. Plans may seek stronger price concessions from manufacturers, but they may also negotiate additional price concessions from pharmacies . I f manufacturer rebates are eliminated, i t is possible plans focus on pharmacy price concessions first , since each dollar of DIR is able to reduce premiums by a greater amount high price / high - ed by rebate dynamics may - be m itigat than each dollar of POS cost reductions. While the in n of manufacturer rebates, plans may still have a strong focus on rebates if they still exist the eliminatio 21 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

24 Milliman Client Report Plans could even structure DIR contracts to provide a greater dollar amount per . POS - any format post the focused formulary strategies discussed generic script than brand script, which would align with - generic in Scenarios 2 and 3. With t hat said, $1 PMPM of increased pharmacy DIR may be more difficult to achieve than $1 PMPM of . ates today, so a harmacy DIR only comprises about 15% of total reb P manufacturer price concessions large percentage increase would be needed to offset changes in the treatment of manufacturer rebates. This was seen in Scenario 7, in which increased pharmacy DIR did less to offset member premium increases than in most other scenarios. IMPACT TO MANUFACTURERS Manufacturers may have differing strategies if manufacturer rebates are eliminated. Manufacturers currently not offering any Part D rebates on a given product may be less likely to change pricing or contracting strategies when facturer manu rebates are eliminated. Products with rebates, especially those in competitive therapeutic classes with multiple alternatives, will face the most pressure to offer the , negotiated price lowest POS cost, whether through lower list prices concessions, or some combination of these two approaches. When considering whether to reduce list prices, manufacturers will need to balance the impact in Part D other markets with the impact to all markets, . List prices are generally the basis of reimbursement across could D if result in a higher net cost in Part so a reduction in list price (with no separate price concession) the manufacturer is attempting to remain neutral across all markets. That said, a lower list price could ter formulary placement in the commercial market, which could improve sales potentially result in bet volume. s discussed later in this section, lower list prices could have significant implications for rebates paid by A ion in list price could increase manufacturer manufacturers to state Medicaid programs, such that a reduct brand y, Additionall revenue in Medicaid. prices have received a great deal of public attention and scrutiny in recent years, so manufacturers may take the opportunity to explicitly reduce list prices, if even o nly by a small amount, to generate a positive public message. Price concessions could be used as the balancing item to then achieve a similar net cost in the Part D market. Manufacturers not applying an explicit reduction to list prices in the short - might still reduce prices over term - the long term by implementing smaller annual price increases than they otherwise would if rebates still hen launching a new product intended to be used by the Medicare existed in Part D. Additionally, w the value of rebates and may launch at a high list price and rs population, manufacture generally recognize offer a rebate right away to get to the targeted net cost in the current system . Without rebates, to launch new products at a lower list price. choose manufacturers could manufacturer rebates is the creation One strategy manufacturers might use in reaction to the elimination of . An authorized generic is the same treatment as the brand, but marketed without the of a uthorized generics brand label, and sometimes sold at a low er cost. If the brand manufacturer produced an authorized generic satisfy Part D plans by offering a at a lower list price and only contracted it in the Part D market, it could lower price without disrupting pricing in other market s . This could be an option if negotiated price concessions are also prohibited. rebates are Manufacturers may have more success marketing b iosimilars manufacturer in Part D if because they were eliminated. H , biosimilars have been strongly disadvantaged in Part D istorically not eligible for the CGDP, which made plans much less likely to prefer them. The BBA changed this dynamic such that biosimilars are now eligible for the CGDP, though they are still often at a disadvantage to - rebate strategies are , high ebate. Without rebates cost biologics if the biologic offers a r - / high - higher price and biosimilar manufacturers could potentially have greater success in Part D. mitigated 22 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

25 Milliman Client Report PHARMACIES RETAIL IMPACT TO As modeled in Scenario 7, if manufacturer rebates are eliminated, retai l pharmacies could receive additional pressure to provide greater rebates, given how valuable rebates are in reducing premiums. However, the results of Scenario 7 showed that greater savings could be achieved through other strategic changes. could potentially cut into retail pharmacy revenue and could also make it harmacy rebates Increases in p difficult for smaller, less profitable pharmacy chains to gain network access. On the other hand, if there are any increases in utilization as member cost sharing decreases, such as modeled in Scenario 6, the added volume could help pharmacy profits. Overall, pharmacies may be impacted less since they are not as affected by changes to brand costs and are less impacted by reimbursement pressures. IMPACT TO PBMS y large health plans own their PBMs, so Man strategies, particularly align with plan PBM reactions may UnitedHealth Group, Humana, and CVS Health, the three largest Part D carriers . among those carriers wn their own PBM. The next largest Part D carrier, accounting for more than half of total enrollment, each o Express Scripts. As a result, formulary and Aetna , recently merge d with CVS Health, and Cigna acquired , though those PBMs may n ot PBMs rebate strategies for health plans could align with those of their owned necessarily share the same strategies with other Part D customers for competitive reasons. Manufacturer administrative fees (MAF), which are paid from manufacturers to PBMs for formulary status, not passed through to Part D plans. The MAF , therefore , are often are not considered rebates by PBMs and is one component of PBM revenue, particularly in Part D where spread pricing is effectively not permitted based on pricing and reporting requirements . If list prices decrease, MAF ( typically contracted as a would also decrease. Moreover, if the safe harbor protection on rebates is lifted, ) percentage of list prices therefore be prohibited. While PBMs do not currently classify , MAF may also be considered a kickback and , Kickback Statute. This could - tion of discount under the federal Anti MAF as a rebate, it may meet the defini PBM revenue may shift to fixed costs paid by plans on a cause a significant reduction to PBM revenues. , per script basis e fee would be to which would require PBMs to determine what a reasonable alternativ Alternatively, PBMs might move to pass through all rebates and MAF to the POS even maintain profitability. if rebates are not prohibited — PBMs have been moving in this direction in recent years, and doing so would allow the current structu re of negotiated rebates to rema in largely in tact. and therefore are able to generate , Most PBMs own their own specialty and / or mail order pharmacies , more revenue when prescriptions are dispensed at these locations. PBMs may try to offset potential This revenue decreases by pushing Part D members to owned spec ialty and mail order pharmacies. incentive rebates and may be manufacturer already exists today, but could be boosted by the loss of owned PBMs who can set lo wer mail order cost sharing to incentivize particularly aligned within carrier - members to use mail. manufacturer rebate contracts, providing an Price protection rebates are common in current Part D additional rebate to the extent list prices increase above a pre - otections are a defined rate. While price pr part of rebates and are required to be passed through from PBMs to plans, in a new environment without manufacturer rebates, perhaps PBMs could leverage similar price protection arrangements as a means to further reduce POS costs. COMMERCIAL MARKET O IMPACT T - Kickback Statute would not directly affect commercial rebates. Recent Changes to the federal Anti interpretation suggests that commercial plans offered under the ACA health insurance exchanges are not 11 , Kickback Statute and thus , are excluded from the Anti - considered federal health care pro grams . While s the current administration has given no indication they intend to revise this treatment, hould the administration revise both their interpretation of the ACA as a federal health care program and the rebate safe harbor, this could significantly alter practices in the commercial market that have heretofore not been 11 Radick, Robert. “The Anti - Kickback Statute and the Affordable Care Act: A Law Enforcement Tool Suddenly Goes Missing.” Forbes. goes a - - enforcement - - tool - suddenly - law https://www.forbes.com/sites/inside r/2013/11/13/the - anti - kickback - statute - and - the - affordable - care - act - missing/#771910f11cea. 23 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

26 Milliman Client Report If that were to occur, the impact to the commercial market would largely tatute. S ickback K - nti A subject to the mirror the impact to the Part D market. Member cost sharing savings could potentially be higher, to the from lower POS costs. ting nefit directly be deductible health plans - extent commercial members have high Premiums would increase in the commercial market as well, though plans might potentially not react as sensitive as in the Part D market, or the standalone - strongly as not all commercial plans are as premium . ticular, which does not exist in commercial PDP market in par Assuming any changes to the rebate safe harbor do not directly impact the commercial market, there could rebates still be some spillover impact to this market. Some plans in the commercial market already reflect For example, . UnitedHealth Group and Aetna both announced they will be passing through at the POS rebates at the POS for fully insured commercial group customers beginning in 2019. Rebates are not as valuable to plans in the commercial market as in Part D, which in some ways makes it easier and less impactful to use POS discounts instead of rebates. For carriers and PBMs operating in both the commercial and Part D markets (which is the case for a majority of companies), it may be difficult to mainta in two different reimbursement strategies, particularly in terms of contracting with manufacturers. Given this and lowering the advantage of costs for members with high deductibles and / or coinsurance, plans may choose ollow suit with Part D. On the other hand, it is also possible plans and to adjust commercial strategies to f shifting focus toward this market PBMs react in the commercial market by valuing rebates even more , where manufacturer rebates can still be used to keep premiums low. time reductions hanges to list prices, whether through explicit one - As mentioned pr eviously, manufacturer c price trends over time or changes to commercial market will directly impact . If list prices go down, , the o keep rebates as well, which could improve commercial plans may benefit while still being permitted t overall revenue. MEDICAID IMPACT TO rebates in Part D, which manufacturer The analysis presented in the previous section focused on removing S could be done by the safe harbor protection on rebates under the federal A nti - K ickback changing tatute. It is important to note the rebate safe harbor currently applies to all federal healthcare programs, including Medicaid. However, it is uncertain whether lifting the safe harbor completely would eliminate all rebates in Medicaid, since the URA is statutorily defined. Under a very broad definition of eliminating manufacturer rebates, one could assume legislation would be adjusted such that all rebates in all federal healthcare pro this grams, including both the URA and supplemental Medicaid rebates, would be prohibited. However, may be unlikely because it would cause a significant increase in both state and federal government costs for Medicaid , at a time when budget pressure in thi s program is enormous . Eliminating rebates in Part D has advantages of reducing member cost sharing and potentially manufacturer rebate strategies and focus more on the lowest price / high - - changing plan behaviors to shift away from high ould possibly lead to lower overall program costs and lower government spending over net cost, which c In Part D, the CGDP and do not apply in the Medicaid market. However, m time. ost of these effects reinsurance complicate the impact of rebates POS costs, since rebate sharing er does not low relative to ember cost sharing in Medicaid is already impact all stakeholders in the same way. Additionally , m influence their decisions. and may not negligible, so a lower POS cost would not affect member spending States and M POS to manage to the lowest are motivated COs already costs, but also have limited tools to to reduce total Medicaid member behavior, both of which mean influence there is much less potential lue of supplemental rebates is fairly Further, the va pharmacy spending to offset the loss of rebates. smal ), digits as a percent of total pharmacy costs - which makes it less typically in the low single l ( , simply due to the lower magnitude of likely for strategies to change if these rebates are altered . impact Eliminating Supplemental Rebates in Medicaid Most states collect some level of supplemental rebates. In many cases, multiple states take part in a purchasing pool, in which the PBM helps negotiate supplemental rebates for the entire pool. MCOs sometimes a though typically only one of the state lso negotiate supplemental rebates (if permitted by law), 24 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

27 Milliman Client Report or the MCO will collect supplemental rebates for a given brand, depending on which entity is responsible DL, as some states manage only certain classes). for managing the PDL (or the relevant portion of the P the state manages the PDL and collects supplemental rebates, the PDL may be more In cases where - brand heavy and also may favor higher cost brands because the rebates more than make up for the higher starti ng cost. The net cost to the state after rebates is often lower for a brand, sometimes even in comparison to a generic alternative. If all supplemental rebates were eliminated, PDL strategies might shift may still be strong the federal URA still exists, there toward medications with lower list prices, though if incentive to cover brands . ) given the URA is many multiples higher than most supplemental rebates ( Moreover, a shift toward products with lower list prices would not necessarily save the government . Without priced brand - or lower money, since the net cost is often lower for a brand than for a generic rebates, a generic may become the lower net cost optio n compared to a brand, but it may still have a higher net cost than the brand had when rebates were permitted. In cases where MCOs currently manage the PDL, PDL strategies could shift towards more generic and he MCO s only concern w ’ ould be lowest POS cost and the MCO does not collect . less brand coverage T any of the federal URA. However, it is uncertain whether there would be overall savings in costs to offset ernments. this would result in lower URA collected by the state and federal gov the loss of rebates, given The URA can average 50% or more of total pharmacy spending, so shifting toward generic products, which Additionally, because MCOs not compensate for lower gross costs. may have a significantly lower URA, already are incentivized to favor generics collected rebates are - and there are likely few cases where MCO manufacturer rebates great enough to offset the cost differential relative to a generic alternative, eliminating may not impact PDL strategies in a material way. Impact to Medicaid of Eliminating Part D Rebates K Even if all Medicaid rebates continue to be excluded from the A - nti ickback S tatute, the removal of manufacturer rebates in Part D could potentially still impact the Medicaid program. Part D rebates do not directly affect the Medicaid program s inc e they are excluded from the definition of Best Price. would have a direct impact if changes to Part D rebates result in , this changes in list prices However , As list price changes, on Medicaid prices and rebates. , as well as it affects the Medicaid gross claim cost the rebate. These changes have opposing effects – if list price decreases, gross claims decrease in turn. However, rebates will decrease, which could more than offset the reduction in gross claims. The i mpact to the URA is twofold : would be he basic rebate amount T lower because it is calculated as 23.1% of a lower AMP value  (or because the difference between AMP and Best Price is also correspondingly lower) lower because the relativity between current would he inflationary component of the rebate also be  T AMP and AMP at time of product launch decreases s We estimate t e URA impacts would result in an increase in total government costs for the e h example: hypothetical o illustrate, consider the following T Medicaid program. 25 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

28 Milliman Client Report Table 7 ASPE Impact to Medicaid of Lower Brand List Prices 15% Lower List Price Baseline Baseline AMP 1.00 1.00 Current AMP 1.47 1.25 Baseline CPI - U 151.60 151.60 175.00 U - Current CPI 175.00 0.29 0.34 Basic Rebate 0.31 Inflationary Rebate 0.09 Total Rebate 0.38 0.65 Net 0.87 0.81 Cost In this example, the brand unit price was $1 when it launched, bu t increased over time to $1.47. The basic rebate is assumed to be based on 23.1% of AMP and there is an additional inflationary rebate of nearly equal magnitude. If the list price is reduced by 15%, AMP would presumably decrease by a similar amount. The basic rebate component is also 15% lower, but the inflationary rebate is significantly lower. This results ease in rebate of more than 40%, which more than outweighs the savings in list price. The net cost in a decr to the government (equal to the current AMP less total rebates, assuming POS reimbursement is equivalent to AMP) increases by 7% in this example. 7 Table illustrates that a decrease in list prices could actually cause an increase in government costs in Medicaid. It is uncertain whether list prices will be reduced as a result of changes to rebates in Part D. prices may remain the same, but negotiated POS Under the chargeback model described above, list discounts would replace the rebate. Whether this model affects AMP may be dependent on how the new reimbursement system is structured. If the chargeback amount is handled by the pharmacy or wholesaler, would ossible the chargeback would be considered in AMP and consequently it is p affect AMP and URA as described above. However, if the chargeback is handled differently, it may not affect AMP based on the current definitions. HHS could rethink the and / or URA to avoid unintended definitions of AMP, Best Price, consequences in the Medicaid market. 26 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

29 Milliman Client Report V. MPTIONS METHODOLOGY AND ASSU odel is We used Milliman’s Part D Analysis and Rating Tool (DART) to complete this analysis. This m data forward for the purpose of creating Medicare Part D bids. The designed to project historical claims Milliman underlying data reflects manual rate s . The manual rates, adjustment factors, assumed ’s demographics, and risk scores in the Milliman Medicare Part D pricing models are based on individual D experience Medicare Part including over 4 0 million member months across 34 U.S. regions and , c NLI and LI relies on separate istributions providing allowed spend d robability p Puerto Rico. Our model laim ications split by product type and distribution channel. levels based on the average price for med W national average amounts, using the manual rate data described e calibrated to the published 2019 Part D o . We adjusted the data t average above with trend and generic pipeline adjustments to project to 201 9 contracting / non - benefit expense assumptions from Milliman’s Medicare Part D Contract Survey, which Part surveyed D plan sponsors on the assumptions underlying their Part D bid development. For simplicity, changes to CGDP as a our baseline scenario assumed the 2019 defined standard benefit, which includ es result of the BBA . For 2020 and future values, we began with our 2019 projection, and we applied annual trends to gross Medicare Trustees report. using information from the 2018 Note , costs and the Part D benefit parameters, the Trustees report includes the anticipated large increase in TrOOP expected to occur in 2020. We used public enrollment files from CMS to estimate the proportion of LI and NLI members nationwide. To estimate LIPS, we ass umed that on average, 95% of LI premiums are paid by the government through premium subsidies. To convert manufacturer rebates to lower POS costs , we calculated the average wholesale price (AWP) discount required to achieve the same allowed cost net of re bates as in the baseline scenario. Manufacturer rebates were assumed to apply to specialty and brand prescriptions only . In our 2020 baseline scenario, we assumed total rebates (manufacturer and pharmacy) were equal to about 27 .5 % of allowed costs and that 15% of total rebates were attributable to pharmacy rebates. In future years, we held the AWP 2020 discount constant. In Scenario 2, we assumed GDR would increase within the first several years. We assumed GDR would % higher an in the baseline by 0. 5 % in 20 20 , and gradually continue to increase until it was 2 .5 be higher th and all future years. 3 than the baseline in 202 GDR with the ultimate 2029 GDR coming in assumptions increase more gradual In Scenario 3, we applied 2.0 % high er than the baseline. We also assumed price concessions would increase within the first several for brands . We per year for each of the 0.5% reduced brand POS costs by years of the 10 - year projection . first 5 years of the projection In Scenario 4, we adj usted the conversion of manufacturer rebates to lower POS costs described above the resulting discounts offset 80% of the lost manufacturer rebates (Scenario 1 assumes the such that discounts offset 100% of the lost manufacturer rebates). In Scenario 5, we decreased brand unit cost trend by 1.0% per year, beginning in 2020. 6 In Scenario , we increased brand use trend by 0.5% and decreased brand unit cost trend by 1.0% per , beginning year in 2020. 5 . in each year from 2020 to 2024 % In ates by , we increased pharmacy reb 7 Scenario 27 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

30 Milliman Client Report VI. CAVEATS, LIMITATIONS , AND QUALIFICATIONS ASPE better understand the This report was developed to help potential changes to the Medicare impact of Part D program . This information may not be appropriate, and should not be used, for other purposes. This . ASPE may share this information with outside entities with ASPE for the use of report is provided , and assumes no duty or liability to, other parties permission. Milliman do es not intend to benefit Milliman’s who receive this work product. Any third party recipient of this work product who desires professional guidance should not rely upon Milliman’s work product, but should engage qualified professionals for advice appropriate to its own specific needs. Any re leases of this report to a third party should be in its entirety. Milliman does not endorse any public policy or advocacy position on matters discussed in this report. D a Milliman database of national Medicare Part Please note that in preparing our estimates, w e relied upon Medicare Trustees Report 2018 and the , public information from CMS , claims . Actual results will certainly vary for specific health plans due to differences in trends, discount arrangements, formulary, demographics, utilization patt erns, and rebate arrangements, among other factors . Note that we did not attempt to evaluate every possible change in stakeholder behavior that may result from these potential program changes. Results will vary based on how members and other stakeholders react to the changes if implemented. for Milliman, member The authors are actuar ies s of the American Academy of Actuaries, and meet the our Qualification Standards of the Academy to render the actuarial opinion contained herein. To the best of knowledge a nd belief, this information is complete and accurate and has been prepared in accordance with generally recognized and accepted actuarial principles and practices. This report outlines the review and opinions of the author s and not necessarily that of Mil liman. The terms , effective of Milliman’s subcontractor agreement with the RAND Corporation to perform work for ASPE . apply to this information and its use , January 30, 2019 and amended , 2018 6 2 October 28 and Human Services Page U.S. Department of Health Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

31 Milliman Client Report APPENDICES U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

32 Milliman Client Report Appendix A1 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Annual Dollar Impact of Removing Manufacturer Rebates (Billions of Dollars) Scenario 1 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -$2.8 -$3.0 -$3.3 -$3.6 -$3.9 -$4.2 -$4.4 -$4.8 -$5.2 -$5.6 -$40.8 Member Premium 1.8 2.0 2.2 2.3 2.5 2.7 2.9 3.1 3.3 3.6 26.4 Total Member Costs -1.0 -1.1 -1.2 -1.3 -1.4 -1.5 -1.6 -1.7 -1.8 -2.0 -14.5 NAR -7.3 -8.0 -8.6 -9.2 -9.9 -10.4 -11.2 -11.9 -12.8 -13.8 -103.1 NADS 15.1 16.4 17.8 19.1 20.6 21.9 23.4 25.1 27.0 29.0 215.4 LICS -5.8 -6.3 -7.0 -7.7 -8.4 -9.2 -10.0 -10.8 -11.8 -12.6 -89.5 LIPS 0.8 0.9 1.0 1.1 1.1 1.2 1.3 1.4 1.5 1.6 12.0 Total Government Costs 2.8 3.0 3.2 3.3 3.5 3.5 3.6 3.8 4.0 4.2 34.8 CGDP -1.9 -2.0 -2.0 -2.0 -2.1 -2.1 -2.1 -2.1 -2.2 -2.2 -20.6 Total Gross Allowed Costs -$43.4 -$48.2 -$53.2 -$58.2 -$63.8 -$68.8 -$75.0 -$81.6 -$89.5 -$97.9 -$679.7 Scenario 2 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -$3.1 -$4.4 -$5.3 -$6.1 -$6.6 -$6.9 -$7.3 -$7.8 -$8.3 -$8.8 -$64.6 Member Premium 1.5 0.9 0.6 0.5 0.6 0.6 0.7 0.8 0.9 1.0 8.4 Total Member Costs -1.6 -3.5 -4.7 -5.6 -6.0 -6.3 -6.6 -6.9 -7.3 -7.8 -56.2 NAR -7.9 -10.1 -11.6 -12.8 -13.7 -14.5 -15.5 -16.4 -17.6 -19.0 -139.1 NADS 14.5 14.1 14.4 14.9 16.1 17.2 18.5 19.9 21.6 23.5 174.7 LICS -6.2 -8.0 -9.3 -10.4 -11.4 -12.3 -13.4 -14.5 -15.8 -17.0 -118.3 LIPS 0.7 0.4 0.3 0.2 0.3 0.3 0.3 0.4 0.4 0.5 3.8 Total Government Costs 1.1 -3.6 -6.2 -8.1 -8.7 -9.3 -10.0 -10.6 -11.4 -12.1 -78.8 CGDP -2.0 -2.6 -2.8 -3.0 -3.1 -3.1 -3.1 -3.2 -3.3 -3.3 -29.5 Total Gross Allowed Costs -$45.9 -$57.8 -$66.9 -$74.8 -$81.6 -$87.4 -$94.7 -$102.3 -$111.4 -$121.1 -$843.8 Scenario 3 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -$3.0 -$4.1 -$5.0 -$5.8 -$6.4 -$6.9 -$7.4 -$7.8 -$8.3 -$8.8 -$63.5 Member Premium 1.6 1.0 0.6 0.3 0.1 0.0 0.1 0.1 0.1 0.2 4.0 Total Member Costs -1.4 -3.1 -4.5 -5.5 -6.3 -6.8 -7.3 -7.7 -8.1 -8.7 -59.5 NAR -8.0 -10.4 -12.5 -14.5 -16.4 -17.5 -18.7 -20.1 -21.7 -23.5 -163.2 NADS 14.8 14.7 15.0 15.6 16.7 17.7 19.0 20.5 22.3 24.2 180.3 LICS -6.1 -7.8 -9.1 -10.4 -11.5 -12.5 -13.5 -14.6 -15.9 -17.1 -118.5 LIPS 0.7 0.5 0.3 0.1 0.0 0.0 0.0 0.0 0.1 0.1 1.9 Total Government Costs 1.5 -3.0 -6.3 -9.2 -11.2 -12.2 -13.2 -14.2 -15.3 -16.4 -99.6 CGDP -2.0 -2.5 -2.7 -2.9 -3.0 -3.1 -3.2 -3.2 -3.3 -3.4 -29.1 Total Gross Allowed Costs -$45.3 -$56.7 -$66.7 -$75.8 -$84.3 -$90.9 -$98.7 -$106.7 -$116.2 -$126.4 -$867.7 Page A - 1 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

33 Milliman Client Report Appendix A1 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Annual Dollar Impact of Removing Manufacturer Rebates (Billions of Dollars) Scenario 4 Ten Year 2028 2024 2025 2026 2020 2023 2029 2022 Year 2021 2027 Impact -$2.4 -$2.7 -$2.9 -$3.1 -$3.3 -$3.6 -$3.8 -$4.1 -$4.4 -$32.6 Member Cost Sharing -$2.2 3.2 Member Premium 3.9 4.2 4.6 5.0 5.4 5.8 6.3 44.9 2.9 3.6 1.6 0.7 1.0 1.1 1.2 1.4 0.9 1.7 1.9 12.3 Total Member Costs 0.8 -3.0 -3.1 NAR -3.2 -3.1 -3.1 -3.0 -2.9 -2.8 -2.8 -30.2 -3.1 NADS 16.9 18.3 19.7 21.2 22.5 24.1 25.7 27.6 29.6 221.1 15.6 -8.7 -4.4 -6.0 -6.7 -7.4 -8.0 -5.5 -9.5 -10.2 -71.4 LICS -4.9 1.4 1.5 1.6 1.8 1.9 2.1 LIPS 2.4 2.6 2.9 20.5 2.2 Total Government Costs 10.4 11.3 12.3 13.3 9.5 15.3 16.4 17.9 19.4 139.9 14.1 -1.8 -1.8 -1.7 -1.7 -1.7 -1.5 -17.1 CGDP -1.8 -1.6 -1.7 -1.7 Total Gross Allowed Costs -$34.7 -$42.6 -$46.6 -$51.0 -$55.0 -$60.0 -$65.3 -$71.5 -$78.3 -$543.6 -$38.6 Scenario 5 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Impact -$5.2 Member Cost Sharing -$3.6 -$4.0 -$4.4 -$4.7 -$3.2 -$5.6 -$6.2 -$6.8 -$46.5 -$2.8 Member Premium 1.7 1.7 1.7 1.6 1.5 1.4 1.3 1.1 0.8 0.5 13.3 -2.8 Total Member Costs -1.2 -1.5 -1.9 -2.4 -3.3 -3.9 -4.6 -5.3 -6.2 -33.1 NAR -7.8 -10.5 -12.0 -13.7 -15.4 -17.5 -19.9 -22.7 -26.0 -154.6 -9.1 22.9 NADS 17.6 18.9 20.2 21.5 16.3 24.5 26.3 28.3 211.4 15.0 LICS -5.9 -6.7 -7.5 -8.4 -9.4 -10.4 -11.4 -12.6 -13.9 -15.2 -101.3 0.6 LIPS 0.8 0.8 0.7 0.7 0.8 0.6 0.5 0.4 0.2 6.1 -3.7 -2.1 -5.5 -7.5 -9.9 -12.7 -38.4 Total Government Costs 2.1 1.3 0.4 -0.7 -2.1 -2.0 -1.9 CGDP -22.8 -2.7 -2.6 -2.4 -2.4 -2.3 -2.3 -2.2 Total Gross Allowed Costs -$44.4 -$63.5 -$71.0 -$78.1 -$86.8 -$96.2 -$107.4 -$119.6 -$774.3 -$50.4 -$56.8 Scenario 6 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Impact -$2.8 -$3.1 -$3.4 -$3.8 -$4.1 -$4.4 -$4.7 -$5.0 -$5.5 -$5.9 -$42.7 Member Cost Sharing 1.7 1.7 1.8 1.8 1.8 1.8 1.8 1.6 1.5 1.3 16.7 Member Premium -2.3 -1.1 -1.3 -1.6 -2.0 -4.6 -2.6 -3.0 -3.4 -4.0 Total Member Costs -26.0 NAR -7.7 -10.1 -11.4 -12.9 -14.4 -16.2 -18.2 -20.6 -23.5 -143.8 -8.9 23.3 NADS 17.7 19.1 20.5 21.8 16.4 25.0 26.9 29.0 214.8 15.1 LICS -5.9 -6.7 -7.5 -8.4 -9.4 -10.4 -11.4 -12.6 -13.9 -15.2 -101.3 0.8 LIPS 0.8 0.8 0.8 0.8 0.8 0.8 0.7 0.7 0.6 7.6 -22.7 Total Government Costs 2.2 1.7 1.0 0.1 -0.9 -2.1 -3.5 -5.1 -6.9 -9.0 -2.1 -2.3 -2.3 -2.2 -2.1 -2.2 -21.2 CGDP -1.9 -2.0 -2.0 -2.1 Total Gross Allowed Costs -$44.2 -$55.9 -$62.2 -$69.2 -$75.7 -$83.8 -$92.4 -$102.7 -$113.9 -$749.8 -$49.9 Scenario 7 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Impact -$4.4 Member Cost Sharing -$3.3 -$3.6 -$3.9 -$4.2 -$3.0 -$4.8 -$5.2 -$5.6 -$40.8 -$2.8 Member Premium 1.7 1.8 1.9 2.0 2.1 2.3 2.4 2.6 2.8 3.0 22.8 -18.1 Total Member Costs -1.0 -1.2 -1.4 -1.6 -1.8 -1.9 -2.0 -2.2 -2.3 -2.5 NAR -7.5 -9.0 -9.8 -10.7 -11.3 -12.1 -12.9 -13.8 -15.0 -110.2 -8.2 22.4 NADS 17.2 18.4 19.6 20.9 16.1 24.0 25.8 27.8 207.1 14.9 LICS -5.8 -6.3 -7.0 -7.7 -8.4 -9.2 -10.0 -10.8 -11.8 -12.6 -89.5 1.1 LIPS 0.9 0.9 1.0 1.0 0.9 1.2 1.3 1.4 10.4 0.8 17.8 2.5 2.4 2.2 1.9 1.5 1.4 1.4 1.5 1.5 1.6 Total Government Costs -2.2 -2.1 -20.6 -2.1 -2.1 -2.1 -2.2 CGDP -1.9 -2.0 -2.0 -2.0 -$81.6 -$75.0 -$68.8 -$63.8 -$58.2 -$53.2 -$48.2 -$43.4 Total Gross Allowed Costs -$679.7 -$97.9 -$89.5 Page A - 2 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

34 Milliman Client Report Appendix A2 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Dollar Impact of Removing Manufacturer Rebates (PMPM) Scenario 1 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing -$4.85 -$5.12 -$5.46 -$5.79 -$6.06 -$6.27 -$6.53 -$6.85 -$7.19 -$7.54 -$6.23 Member Premium $3.15 $3.33 $3.51 $3.70 $3.89 $4.06 $4.24 $4.45 $4.67 $4.85 $4.03 Total Member Costs -$1.70 -$1.80 -$1.95 -$2.09 -$2.18 -$2.21 -$2.28 -$2.40 -$2.52 -$2.69 -$2.20 NAR -$12.84 -$13.43 -$14.04 -$14.62 -$15.24 -$15.66 -$16.37 -$17.07 -$17.82 -$18.75 -$15.73 NADS $26.35 $27.64 $28.99 $30.37 $31.77 $32.89 $34.38 $35.99 $37.65 $39.36 $32.86 LICS -$10.06 -$10.69 -$11.35 -$12.15 -$12.96 -$13.80 -$14.62 -$15.54 -$16.43 -$17.15 -$13.66 LIPS $1.48 $1.54 $1.61 $1.69 $1.77 $1.84 $1.92 $2.02 $2.12 $2.20 $1.84 Total Government Costs $4.93 $5.06 $5.21 $5.29 $5.34 $5.27 $5.32 $5.41 $5.52 $5.66 $5.32 CGDP -$3.24 -$3.28 -$3.26 -$3.23 -$3.19 -$3.09 -$3.08 -$3.05 -$3.05 -$3.02 -$3.14 Total Gross Allowed Costs -$75.90 -$81.18 -$86.77 -$92.40 -$98.45 -$103.28 -$110.08 -$117.27 -$124.88 -$132.93 -$103.70 Scenario 2 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing -$5.44 -$7.46 -$8.68 -$9.64 -$10.12 -$10.40 -$10.77 -$11.14 -$11.54 -$11.95 -$9.85 Member Premium $2.70 $1.58 $1.05 $0.79 $0.87 $0.96 $1.06 $1.18 $1.31 $1.42 $1.27 Total Member Costs -$2.74 -$5.88 -$7.63 -$8.85 -$9.26 -$9.44 -$9.71 -$9.96 -$10.23 -$10.53 -$8.58 NAR -$13.78 -$16.97 -$18.92 -$20.33 -$21.15 -$21.72 -$22.67 -$23.63 -$24.63 -$25.81 -$21.22 NADS $25.35 $23.70 $23.40 $23.68 $24.83 $25.82 $27.18 $28.64 $30.19 $31.84 $26.65 LICS -$10.83 -$13.48 -$15.09 -$16.50 -$17.56 -$18.54 -$19.64 -$20.83 -$22.03 -$23.09 -$18.04 LIPS $1.26 $0.73 $0.48 $0.36 $0.39 $0.44 $0.48 $0.54 $0.59 $0.64 $0.58 Total Government Costs $2.01 -$6.02 -$10.12 -$12.79 -$13.48 -$14.00 -$14.64 -$15.28 -$15.88 -$16.41 -$12.03 CGDP -$3.58 -$4.33 -$4.60 -$4.77 -$4.76 -$4.62 -$4.61 -$4.59 -$4.57 -$4.53 -$4.51 Total Gross Allowed Costs -$80.19 -$97.31 -$109.03 -$118.71 -$125.85 -$131.24 -$138.94 -$147.00 -$155.46 -$164.32 -$128.73 Scenario 3 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing -$5.22 -$6.96 -$8.23 -$9.21 -$9.83 -$10.33 -$10.81 -$11.18 -$11.56 -$11.97 -$9.68 Member Premium $2.77 $1.70 $0.96 $0.42 $0.09 $0.07 $0.10 $0.14 $0.18 $0.21 $0.61 Total Member Costs -$2.44 -$5.26 -$7.27 -$8.80 -$9.74 -$10.26 -$10.71 -$11.04 -$11.38 -$11.76 -$9.07 NAR -$13.89 -$17.45 -$20.37 -$22.98 -$25.31 -$26.21 -$27.48 -$28.84 -$30.28 -$31.94 -$24.89 NADS $25.80 $24.73 $24.47 $24.75 $25.70 $26.52 $27.89 $29.43 $31.06 $32.81 $27.51 LICS -$10.61 -$13.08 -$14.88 -$16.49 -$17.76 -$18.73 -$19.83 -$21.02 -$22.21 -$23.26 -$18.08 LIPS $1.30 $0.79 $0.44 $0.19 $0.04 $0.03 $0.04 $0.06 $0.08 $0.09 $0.28 Total Government Costs $2.60 -$5.01 -$10.33 -$14.53 -$17.32 -$18.39 -$19.38 -$20.37 -$21.34 -$22.30 -$15.19 CGDP -$3.45 -$4.14 -$4.43 -$4.59 -$4.65 -$4.60 -$4.65 -$4.62 -$4.60 -$4.55 -$4.45 Total Gross Allowed Costs -$79.19 -$95.53 -$108.75 -$120.26 -$130.13 -$136.51 -$144.79 -$153.28 -$162.20 -$171.55 -$132.39 Page A - 3 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

35 Milliman Client Report Appendix A2 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Dollar Impact of Removing Manufacturer Rebates (PMPM) Scenario 4 Ten Year 2028 2024 2025 2026 2020 2023 2029 2022 Year 2021 2027 Average -$4.08 -$4.34 -$4.62 -$4.84 -$5.02 -$5.22 -$5.45 -$5.74 -$6.03 -$4.97 Member Cost Sharing -$3.86 $5.45 Member Premium $6.17 $6.54 $6.87 $7.26 $7.69 $8.13 $8.56 $6.84 $5.11 $5.80 $2.23 $1.25 $1.55 $1.69 $1.86 $2.04 $1.47 $2.39 $2.53 $1.87 Total Member Costs $1.37 -$5.26 -$5.22 NAR -$5.00 -$4.85 -$4.63 -$4.45 -$4.21 -$3.95 -$3.80 -$4.61 -$5.13 NADS $28.50 $29.86 $31.25 $32.67 $33.82 $35.30 $36.86 $38.50 $40.17 $33.73 $27.19 -$12.53 -$7.77 -$9.57 -$10.27 -$11.06 -$11.75 -$8.90 -$13.29 -$13.89 -$10.89 LICS -$8.31 $2.39 $2.52 $2.66 $2.82 $2.98 $3.12 LIPS $3.49 $3.69 $3.89 $3.12 $3.30 Total Government Costs $17.49 $18.50 $19.51 $20.53 $16.55 $22.39 $23.61 $24.94 $26.37 $21.35 $21.24 -$2.52 -$2.55 -$2.56 -$2.57 -$2.65 -$2.68 -$2.60 CGDP -$2.48 -$2.71 -$2.70 -$2.68 Total Gross Allowed Costs -$60.70 -$69.39 -$73.89 -$78.73 -$82.60 -$88.03 -$93.78 -$99.87 -$106.30 -$82.93 -$64.92 Scenario 5 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Average -$7.58 Member Cost Sharing -$5.84 -$6.32 -$6.75 -$7.12 -$5.37 -$8.10 -$8.64 -$9.20 -$7.09 -$4.95 Member Premium $2.92 $2.84 $2.72 $2.57 $2.38 $2.14 $1.86 $1.55 $1.18 $0.72 $2.04 -$4.37 Total Member Costs -$2.03 -$2.52 -$3.12 -$3.75 -$4.98 -$5.72 -$6.54 -$7.46 -$8.48 -$5.05 NAR -$13.71 -$17.08 -$18.98 -$21.11 -$23.11 -$25.71 -$28.55 -$31.70 -$35.31 -$23.58 -$15.31 $33.61 NADS $28.69 $29.93 $31.23 $32.21 $27.45 $35.14 $36.73 $38.37 $32.26 $26.26 LICS -$10.31 -$11.22 -$12.18 -$13.27 -$14.44 -$15.60 -$16.79 -$18.12 -$19.44 -$20.63 -$15.46 $0.97 LIPS $1.31 $1.25 $1.18 $1.09 $1.37 $0.84 $0.70 $0.54 $0.33 $0.93 -$5.52 -$3.23 -$8.05 -$10.82 -$13.88 -$17.24 -$5.85 Total Government Costs $3.60 $2.23 $0.67 -$1.14 -$3.45 -$3.40 -$3.31 CGDP -$3.48 -$3.62 -$3.57 -$3.52 -$3.46 -$3.42 -$3.49 -$3.46 Total Gross Allowed Costs -$77.64 -$100.78 -$109.57 -$117.25 -$127.37 -$138.23 -$149.88 -$162.38 -$118.14 -$84.88 -$92.69 Scenario 6 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Average -$4.89 -$5.23 -$5.61 -$5.99 -$6.32 -$6.57 -$6.87 -$7.24 -$7.62 -$8.01 -$6.51 Member Cost Sharing $2.47 $2.98 $2.92 $2.86 $2.76 $2.64 $2.96 $2.29 $2.08 $1.78 $2.55 Member Premium -$3.55 -$1.91 -$2.26 -$2.69 -$3.13 -$6.23 -$3.94 -$4.40 -$4.95 -$5.55 Total Member Costs -$3.97 NAR -$13.53 -$16.45 -$18.08 -$19.89 -$21.55 -$23.76 -$26.15 -$28.79 -$31.83 -$21.94 -$14.93 $34.25 NADS $28.90 $30.24 $31.65 $32.74 $27.58 $35.89 $37.61 $39.41 $32.78 $26.32 LICS -$10.31 -$11.22 -$12.18 -$13.27 -$14.44 -$15.60 -$16.79 -$18.12 -$19.44 -$20.63 -$15.46 $1.19 LIPS $1.37 $1.34 $1.31 $1.26 $1.40 $1.12 $1.04 $0.94 $0.81 $1.16 -$3.46 Total Government Costs $3.87 $2.80 $1.60 $0.19 -$1.42 -$3.21 -$5.18 -$7.33 -$9.68 -$12.24 -$3.28 -$3.16 -$3.18 -$3.19 -$3.19 -$3.18 -$3.24 CGDP -$3.27 -$3.32 -$3.32 -$3.31 Total Gross Allowed Costs -$77.21 -$91.21 -$98.66 -$106.73 -$113.65 -$122.89 -$132.77 -$143.33 -$154.62 -$114.39 -$83.96 Scenario 7 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Average -$6.53 Member Cost Sharing -$5.46 -$5.79 -$6.06 -$6.27 -$5.12 -$6.85 -$7.19 -$7.54 -$6.23 -$4.85 Member Premium $3.04 $3.10 $3.15 $3.20 $3.23 $3.39 $3.56 $3.75 $3.94 $4.10 $3.47 -$2.76 Total Member Costs -$1.80 -$2.02 -$2.31 -$2.59 -$2.83 -$2.88 -$2.97 -$3.10 -$3.25 -$3.44 NAR -$13.02 -$14.68 -$15.54 -$16.48 -$16.95 -$17.73 -$18.50 -$19.33 -$20.35 -$16.82 -$13.83 $32.83 NADS $28.12 $29.17 $30.23 $31.34 $27.08 $34.43 $36.08 $37.78 $31.60 $26.08 LICS -$10.06 -$10.69 -$11.35 -$12.15 -$12.96 -$13.80 -$14.62 -$15.54 -$16.43 -$17.15 -$13.66 $1.61 LIPS $1.45 $1.46 $1.47 $1.54 $1.43 $1.70 $1.79 $1.86 $1.59 $1.43 $2.71 $4.42 $3.99 $3.53 $2.95 $2.26 $2.13 $2.09 $2.09 $2.11 $2.14 Total Government Costs -$3.02 -$3.09 -$3.14 -$3.08 -$3.19 -$3.05 -$3.05 CGDP -$3.24 -$3.28 -$3.26 -$3.23 -$117.27 -$110.08 -$103.28 -$98.45 -$92.40 -$86.77 -$81.18 -$75.90 Total Gross Allowed Costs -$103.70 -$132.93 -$124.88 Page A - 4 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

36 Milliman Client Report Appendix B1 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Projected Annual Costs After Removing Manufacturer Rebates (Billions of Dollars) Baseline Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing $24.8 $26.8 $28.9 $30.9 $33.0 $35.0 $37.1 $39.3 $41.9 $44.7 $342.4 Member Premium 13.3 14.8 16.4 17.9 19.7 21.2 23.2 25.3 27.7 30.5 209.9 Total Member Costs 38.0 41.6 45.3 48.8 52.7 56.3 60.3 64.5 69.6 75.1 552.4 NAR 46.9 52.7 58.9 65.1 72.2 78.4 86.9 95.8 106.5 118.4 781.8 NADS 10.1 10.5 10.9 11.2 11.4 11.7 11.7 11.5 11.3 11.0 111.3 LICS 31.6 34.4 37.2 40.2 43.5 46.6 50.1 53.9 58.3 62.9 458.7 LIPS 6.2 6.8 7.5 8.2 9.0 9.6 10.5 11.5 12.6 13.8 95.8 Total Government Costs 94.9 104.5 114.5 124.7 136.1 146.4 159.2 172.6 188.7 206.1 1,447.6 CGDP 6.6 7.1 7.5 7.9 8.3 8.7 9.2 9.6 10.2 10.9 85.9 Total Gross Allowed Costs $182.9 $201.2 $220.2 $239.1 $260.2 $279.2 $302.6 $327.2 $356.7 $388.5 $2,757.6 Scenario 1 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing $22.0 $23.8 $25.6 $27.3 $29.1 $30.9 $32.7 $34.5 $36.7 $39.1 $301.6 Member Premium 15.1 16.8 18.5 20.2 22.2 23.9 26.1 28.4 31.1 34.0 236.3 Total Member Costs 37.1 40.6 44.1 47.5 51.3 54.8 58.8 62.9 67.8 73.2 537.9 NAR 39.5 44.7 50.3 55.9 62.4 68.0 75.7 84.0 93.8 104.5 678.8 NADS 25.2 27.0 28.7 30.3 32.0 33.6 35.1 36.5 38.3 40.0 326.7 LICS 25.9 28.0 30.3 32.6 35.1 37.4 40.1 43.0 46.5 50.2 369.2 LIPS 7.1 7.8 8.5 9.3 10.1 10.9 11.8 12.9 14.1 15.4 107.8 Total Government Costs 97.7 107.5 117.7 128.0 139.6 149.9 162.8 176.4 192.7 210.2 1,482.5 CGDP 4.7 5.1 5.5 5.8 6.3 6.6 7.1 7.5 8.1 8.6 65.3 Total Gross Allowed Costs $139.5 $153.0 $167.0 $180.9 $196.4 $210.4 $227.5 $245.6 $267.2 $290.5 $2,077.9 Scenario 2 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing $21.6 $22.4 $23.6 $24.8 $26.4 $28.1 $29.8 $31.5 $33.6 $35.9 $277.8 Member Premium 14.8 15.7 17.0 18.4 20.2 21.9 23.9 26.1 28.7 31.5 218.3 Total Member Costs 36.5 38.2 40.6 43.2 46.7 50.0 53.7 57.6 62.3 67.4 496.1 NAR 39.0 42.6 47.3 52.3 58.5 64.0 71.4 79.4 88.9 99.3 642.7 NADS 24.6 24.6 25.3 26.1 27.5 28.9 30.2 31.4 32.9 34.5 286.0 LICS 25.4 26.4 28.0 29.8 32.1 34.2 36.7 39.4 42.5 45.9 340.4 LIPS 6.9 7.3 7.8 8.4 9.2 9.9 10.9 11.8 13.0 14.3 99.6 Total Government Costs 96.0 100.9 108.3 116.7 127.4 137.0 149.2 162.0 177.3 194.0 1,368.8 CGDP 4.5 4.5 4.7 4.9 5.2 5.6 6.0 6.5 7.0 7.5 56.4 Total Gross Allowed Costs $137.0 $143.4 $153.3 $164.3 $178.6 $191.8 $207.9 $224.9 $245.3 $267.4 $1,913.9 Scenario 3 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing $21.8 $22.7 $23.9 $25.1 $26.6 $28.1 $29.7 $31.5 $33.6 $35.9 $278.9 Member Premium 14.9 15.8 17.0 18.2 19.7 21.3 23.3 25.4 27.9 30.6 213.9 Total Member Costs 36.6 38.5 40.8 43.3 46.4 49.4 53.0 56.9 61.5 66.5 492.9 NAR 38.9 42.3 46.4 50.6 55.8 61.0 68.1 75.8 84.8 94.8 618.7 NADS 24.9 25.2 25.9 26.8 28.0 29.4 30.7 32.0 33.6 35.2 291.6 LICS 25.6 26.6 28.1 29.8 32.0 34.1 36.6 39.2 42.4 45.7 340.2 LIPS 7.0 7.3 7.8 8.3 9.0 9.7 10.6 11.5 12.7 13.9 97.6 Total Government Costs 96.3 101.5 108.2 115.6 124.9 134.1 146.0 158.5 173.4 189.6 1,348.1 CGDP 4.6 4.6 4.8 5.0 5.3 5.6 6.0 6.4 7.0 7.5 56.8 Total Gross Allowed Costs $137.6 $144.5 $153.5 $163.3 $175.9 $188.3 $203.9 $220.5 $240.5 $262.1 $1,889.9 Page A - 5 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

37 Milliman Client Report Appendix B1 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Projected Annual Costs After Removing Manufacturer Rebates (Billions of Dollars) Scenario 4 Ten Year 2028 2024 2025 2026 2020 2023 2029 2022 Year 2021 2027 Impact $24.4 $26.3 $28.0 $29.9 $31.7 $33.5 $35.5 $37.8 $40.2 $309.8 Member Cost Sharing $22.5 18.0 Member Premium 21.8 23.9 25.8 28.2 30.6 33.6 36.8 254.8 16.2 19.9 66.1 38.8 49.8 53.8 57.5 61.7 46.2 71.4 77.0 564.6 Total Member Costs 42.5 43.9 49.6 NAR 62.0 69.1 75.3 83.8 92.9 103.7 115.6 751.6 55.8 NADS 27.5 29.2 30.9 32.6 34.3 35.7 37.1 38.9 40.6 332.4 25.7 45.1 27.2 34.2 36.9 39.2 42.1 31.8 48.8 52.6 387.3 LICS 29.4 7.6 8.3 9.1 10.0 10.9 11.7 LIPS 13.9 15.2 16.7 116.2 12.8 Total Government Costs 114.8 125.9 137.0 149.4 104.3 174.4 189.1 206.6 225.5 1,587.6 160.5 8.4 7.9 7.4 7.0 6.6 5.0 68.8 CGDP 9.0 5.5 5.8 6.2 Total Gross Allowed Costs $148.2 $177.6 $192.5 $209.2 $224.2 $242.6 $261.9 $285.1 $310.1 $2,214.1 $162.6 Scenario 5 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Impact $31.9 Member Cost Sharing $25.3 $26.9 $28.6 $30.3 $23.7 $33.6 $35.7 $37.9 $296.0 $21.9 Member Premium 15.0 16.5 18.0 19.5 21.2 22.7 24.5 26.3 28.6 31.0 223.3 49.8 Total Member Costs 36.9 40.1 43.4 46.5 53.0 56.4 60.0 64.3 68.9 519.2 NAR 39.0 48.4 53.1 58.6 63.0 69.3 76.0 83.8 92.3 627.3 43.6 34.6 NADS 28.5 30.0 31.6 33.2 26.8 35.9 37.6 39.3 322.7 25.1 LICS 25.7 27.7 29.7 31.9 34.2 36.2 38.7 41.3 44.4 47.7 357.4 10.3 LIPS 7.6 8.3 8.9 9.7 7.0 11.1 12.0 13.0 14.1 101.9 142.7 134.0 153.7 165.1 178.8 193.4 1,409.3 Total Government Costs 96.9 105.8 114.9 124.0 5.4 5.0 4.7 CGDP 63.1 8.2 7.7 7.2 6.8 6.4 6.1 5.7 Total Gross Allowed Costs $138.5 $175.6 $189.2 $201.1 $215.8 $231.0 $249.3 $268.8 $1,983.3 $150.8 $163.3 Scenario 6 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Impact $22.0 $23.7 $25.5 $27.1 $28.9 $30.7 $32.4 $34.2 $36.4 $38.8 $299.7 Member Cost Sharing 24.9 15.0 18.2 19.7 21.4 23.0 16.6 26.9 29.2 31.8 226.6 Member Premium 50.4 36.9 40.3 43.6 46.8 70.6 53.6 57.3 61.1 65.7 Total Member Costs 526.4 NAR 39.1 48.8 53.7 59.3 64.1 70.7 77.6 85.9 94.9 638.0 43.8 35.0 NADS 28.6 30.2 31.9 33.5 26.9 36.5 38.3 40.0 326.2 25.2 LICS 25.7 27.7 29.7 31.9 34.2 36.2 38.7 41.3 44.4 47.7 357.4 10.4 LIPS 7.7 8.3 9.0 9.8 7.0 11.3 12.2 13.3 14.4 103.4 1,425.0 Total Government Costs 97.1 106.1 115.5 124.8 135.2 144.2 155.6 167.5 181.8 197.0 6.2 8.5 8.0 7.0 6.6 7.4 64.7 CGDP 4.7 5.1 5.4 5.8 Total Gross Allowed Costs $138.7 $164.2 $176.9 $191.0 $203.5 $218.8 $234.8 $254.0 $274.5 $2,007.9 $151.3 Scenario 7 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Impact $32.7 Member Cost Sharing $25.6 $27.3 $29.1 $30.9 $23.8 $34.5 $36.7 $39.1 $301.6 $22.0 Member Premium 15.0 16.6 18.3 19.9 21.8 23.5 25.6 27.9 30.6 33.5 232.7 534.3 Total Member Costs 37.0 40.4 43.9 47.2 50.8 54.3 58.3 62.4 67.3 72.6 NAR 39.4 49.9 55.3 61.6 67.1 74.8 83.0 92.7 103.4 671.6 44.5 34.0 NADS 28.1 29.6 31.0 32.6 26.6 35.4 37.2 38.8 318.4 25.0 LICS 25.9 28.0 30.3 32.6 35.1 37.4 40.1 43.0 46.5 50.2 369.2 11.6 LIPS 8.4 9.1 9.9 10.7 7.7 12.7 13.9 15.2 106.2 7.0 1,465.4 97.4 106.8 116.7 126.6 137.6 147.8 160.6 174.1 190.2 207.6 Total Government Costs 8.6 6.6 65.3 7.1 6.3 7.5 8.1 CGDP 4.7 5.1 5.5 5.8 $245.6 $227.5 $210.4 $196.4 $180.9 $167.0 $153.0 $139.5 Total Gross Allowed Costs $2,077.9 $290.5 $267.2 Page A - 6 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

38 Milliman Client Report Appendix B2 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Projected Costs After Removing Manufacturer Rebates (PMPM) Baseline Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing $43.26 $45.19 $47.15 $49.05 $50.94 $52.60 $54.44 $56.44 $58.48 $60.65 $52.24 Member Premium $23.21 $24.93 $26.71 $28.44 $30.33 $31.88 $34.04 $36.30 $38.72 $41.33 $32.03 Total Member Costs $66.46 $70.11 $73.86 $77.49 $81.27 $84.48 $88.48 $92.73 $97.20 $101.99 $84.27 NAR $81.92 $88.72 $96.05 $103.34 $111.47 $117.76 $127.45 $137.69 $148.69 $160.64 $119.28 NADS $17.63 $17.74 $17.79 $17.74 $17.57 $17.63 $17.12 $16.49 $15.80 $14.93 $16.98 LICS $55.28 $57.87 $60.69 $63.86 $67.17 $69.93 $73.51 $77.39 $81.36 $85.33 $69.98 LIPS $10.87 $11.51 $12.25 $13.01 $13.84 $14.46 $15.44 $16.48 $17.58 $18.76 $14.61 Total Government Costs $165.71 $175.84 $186.78 $197.95 $210.04 $219.78 $233.53 $248.04 $263.42 $279.67 $220.86 CGDP $11.46 $11.89 $12.20 $12.52 $12.86 $13.05 $13.45 $13.86 $14.30 $14.73 $13.11 Total Gross Allowed Costs $319.55 $338.72 $359.05 $379.51 $401.52 $419.19 $443.92 $470.11 $497.85 $527.22 $420.73 Scenario 1 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing $38.41 $40.06 $41.69 $43.27 $44.87 $46.33 $47.91 $49.58 $51.29 $53.11 $46.01 Member Premium $26.36 $28.26 $30.22 $32.14 $34.22 $35.94 $38.28 $40.75 $43.39 $46.19 $36.06 Total Member Costs $64.76 $68.32 $71.91 $75.40 $79.09 $82.27 $86.20 $90.34 $94.68 $99.30 $82.07 NAR $69.08 $75.29 $82.01 $88.72 $96.23 $102.10 $111.08 $120.62 $130.87 $141.89 $103.56 NADS $43.99 $45.39 $46.78 $48.11 $49.34 $50.52 $51.51 $52.48 $53.45 $54.30 $49.84 LICS $45.22 $47.18 $49.34 $51.71 $54.20 $56.13 $58.89 $61.85 $64.93 $68.18 $56.33 LIPS $12.35 $13.05 $13.86 $14.70 $15.61 $16.30 $17.37 $18.50 $19.70 $20.96 $16.45 Total Government Costs $170.64 $180.91 $191.99 $203.25 $215.38 $225.05 $238.85 $253.45 $268.94 $285.33 $226.18 CGDP $8.22 $8.60 $8.93 $9.28 $9.67 $9.96 $10.37 $10.81 $11.25 $11.71 $9.97 Total Gross Allowed Costs $243.65 $257.55 $272.27 $287.11 $303.07 $315.91 $333.85 $352.85 $372.98 $394.30 $317.03 Scenario 2 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing $37.82 $37.73 $38.47 $39.41 $40.82 $42.19 $43.67 $45.30 $46.94 $48.70 $42.39 Member Premium $25.90 $26.50 $27.76 $29.23 $31.20 $32.85 $35.10 $37.48 $40.03 $42.75 $33.30 Total Member Costs $63.72 $64.23 $66.23 $68.64 $72.01 $75.04 $78.77 $82.77 $86.97 $91.45 $75.69 NAR $68.14 $71.75 $77.13 $83.01 $90.32 $96.04 $104.78 $114.06 $124.06 $134.83 $98.06 NADS $42.98 $41.44 $41.19 $41.43 $42.40 $43.45 $44.31 $45.13 $45.99 $46.78 $43.64 LICS $44.46 $44.40 $45.61 $47.36 $49.61 $51.40 $53.87 $56.56 $59.33 $62.24 $51.94 LIPS $12.13 $12.24 $12.74 $13.37 $14.23 $14.89 $15.93 $17.01 $18.17 $19.41 $15.20 Total Government Costs $167.72 $169.82 $176.66 $185.16 $196.56 $205.78 $218.89 $232.76 $247.55 $263.25 $208.83 CGDP $7.88 $7.56 $7.59 $7.75 $8.10 $8.43 $8.84 $9.28 $9.73 $10.20 $8.60 Total Gross Allowed Costs $239.36 $241.41 $250.01 $260.81 $275.67 $287.95 $304.99 $323.11 $342.39 $362.90 $292.00 Scenario 3 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Average Member Cost Sharing $38.04 $38.22 $38.92 $39.84 $41.10 $42.26 $43.63 $45.26 $46.92 $48.68 $42.56 Member Premium $25.98 $26.63 $27.67 $28.85 $30.43 $31.96 $34.14 $36.44 $38.91 $41.54 $32.64 Total Member Costs $64.02 $64.86 $66.59 $68.69 $71.53 $74.22 $77.77 $81.69 $85.83 $90.22 $75.20 NAR $68.03 $71.27 $75.68 $80.36 $86.16 $91.55 $99.97 $108.85 $118.41 $128.70 $94.39 NADS $43.43 $42.47 $42.26 $42.50 $43.28 $44.15 $45.01 $45.92 $46.86 $47.75 $44.49 LICS $44.68 $44.80 $45.81 $47.37 $49.40 $51.20 $53.68 $56.37 $59.15 $62.07 $51.90 LIPS $12.17 $12.30 $12.70 $13.20 $13.88 $14.49 $15.49 $16.54 $17.66 $18.85 $14.89 Total Government Costs $168.31 $170.83 $176.45 $183.42 $192.72 $201.39 $214.15 $227.68 $242.08 $257.37 $205.67 CGDP $8.01 $7.75 $7.77 $7.93 $8.21 $8.45 $8.80 $9.24 $9.70 $10.18 $8.66 Total Gross Allowed Costs $240.36 $243.20 $250.29 $259.26 $271.39 $282.69 $299.13 $316.83 $335.66 $355.68 $288.34 Page A - 7 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

39 Milliman Client Report Appendix B2 ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Projected Costs After Removing Manufacturer Rebates (PMPM) Scenario 4 Ten Year 2028 2024 2025 2026 2020 2023 2029 2022 Year 2021 2027 Average $41.10 $42.81 $44.43 $46.09 $47.58 $49.21 $50.98 $52.74 $54.62 $47.27 Member Cost Sharing $39.39 $30.38 Member Premium $34.60 $36.87 $38.76 $41.31 $43.98 $46.86 $49.90 $38.87 $28.32 $32.51 $94.97 $67.71 $79.04 $82.97 $86.34 $90.52 $75.33 $99.60 $104.51 $86.14 Total Member Costs $71.48 $76.66 $83.50 NAR $98.34 $106.62 $113.13 $123.00 $133.48 $144.74 $156.84 $114.67 $90.92 NADS $46.24 $47.65 $49.00 $50.24 $51.45 $52.43 $53.35 $54.29 $55.11 $50.71 $44.82 $64.86 $47.51 $54.29 $56.89 $58.87 $61.76 $51.80 $68.06 $71.44 $59.09 LICS $49.56 $13.26 $14.03 $14.92 $15.83 $16.82 $17.57 LIPS $19.96 $21.27 $22.65 $17.73 $18.74 Total Government Costs $193.33 $205.28 $217.46 $230.57 $182.26 $255.92 $271.66 $288.37 $306.03 $242.21 $241.02 $11.78 $11.31 $10.89 $10.48 $10.21 $8.78 $10.50 CGDP $12.25 $9.18 $9.49 $9.84 Total Gross Allowed Costs $258.85 $289.65 $305.62 $322.79 $336.59 $355.89 $376.33 $397.99 $420.92 $337.80 $273.80 Scenario 5 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Average $46.86 Member Cost Sharing $41.31 $42.73 $44.18 $45.48 $39.82 $48.34 $49.84 $51.45 $45.15 $38.30 Member Premium $26.13 $27.77 $29.43 $31.01 $32.71 $34.03 $35.90 $37.85 $39.91 $42.05 $34.07 $76.90 Total Member Costs $64.43 $67.59 $70.74 $73.74 $79.51 $82.76 $86.19 $89.75 $93.51 $79.22 NAR $68.21 $78.97 $84.36 $90.36 $94.65 $101.74 $109.14 $116.99 $125.33 $95.70 $73.41 $50.73 NADS $46.47 $47.68 $48.81 $49.84 $45.19 $51.63 $52.53 $53.31 $49.24 $43.89 LICS $44.97 $46.65 $48.51 $50.59 $52.73 $54.33 $56.72 $59.27 $61.91 $64.70 $54.52 $15.43 LIPS $12.82 $13.50 $14.18 $14.92 $12.24 $16.29 $17.18 $18.11 $19.09 $15.54 $214.25 $206.81 $225.48 $237.22 $249.55 $262.43 $215.01 Total Government Costs $169.31 $178.07 $187.46 $196.81 $8.75 $8.48 $8.16 CGDP $9.63 $11.11 $10.73 $10.35 $9.99 $9.63 $9.37 $9.05 Total Gross Allowed Costs $241.91 $278.73 $291.95 $301.94 $316.55 $331.88 $347.97 $364.85 $302.59 $253.84 $266.35 Scenario 6 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Average $38.37 $39.96 $41.54 $43.07 $44.62 $46.03 $47.56 $49.20 $50.86 $52.64 $45.73 Member Cost Sharing $36.51 $26.19 $29.63 $31.29 $33.10 $34.52 $27.89 $38.59 $40.80 $43.12 $34.58 Member Premium $77.72 $64.55 $67.85 $71.17 $74.36 $95.76 $80.54 $84.08 $87.79 $91.66 Total Member Costs $80.31 NAR $68.39 $79.60 $85.26 $91.58 $96.21 $103.69 $111.54 $119.90 $128.81 $97.34 $73.79 $51.37 NADS $46.68 $47.98 $49.22 $50.37 $45.32 $52.38 $53.41 $54.34 $49.76 $43.96 LICS $44.97 $46.65 $48.51 $50.59 $52.73 $54.33 $56.72 $59.27 $61.91 $64.70 $54.52 $15.65 LIPS $12.88 $13.59 $14.31 $15.10 $12.27 $16.56 $17.52 $18.52 $19.57 $15.78 $217.40 Total Government Costs $169.58 $178.64 $188.39 $198.15 $208.62 $216.56 $228.35 $240.71 $253.74 $267.43 $9.57 $11.57 $11.12 $10.26 $9.86 $10.68 $9.87 CGDP $8.20 $8.56 $8.87 $9.20 Total Gross Allowed Costs $242.34 $267.84 $280.85 $294.79 $305.54 $321.03 $337.34 $354.52 $372.60 $306.34 $254.76 Scenario 7 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Average $47.9 Member Cost Sharing $41.7 $43.3 $44.9 $46.3 $40.1 $49.6 $51.3 $53.1 $46.0 $38.4 Member Premium 26.2 28.0 29.9 31.6 33.6 35.3 37.6 40.0 42.7 45.4 35.5 81.5 Total Member Costs 64.7 68.1 71.6 74.9 78.4 81.6 85.5 89.6 94.0 98.5 NAR 68.9 81.4 87.8 95.0 100.8 109.7 119.2 129.4 140.3 102.5 74.9 50.0 NADS 45.9 46.9 47.8 49.0 44.8 50.9 51.9 52.7 48.6 43.7 LICS 45.2 47.2 49.3 51.7 54.2 56.1 58.9 61.9 64.9 68.2 56.3 17.1 LIPS 13.7 14.5 15.3 16.0 12.9 18.2 19.4 20.6 16.2 12.3 223.6 170.1 179.8 190.3 200.9 212.3 221.9 235.6 250.1 265.5 281.8 Total Government Costs 11.7 10.0 10.0 10.4 9.7 10.8 11.3 CGDP 8.2 8.6 8.9 9.3 $352.8 $333.8 $315.9 $303.1 $287.1 $272.3 $257.5 $243.7 Total Gross Allowed Costs $317.0 $394.3 $373.0 Page A - 8 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

40 Milliman Client Report Appendix C ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Annual Percentage Cost Impact of Removing Manufacturer Rebates Scenario 1 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -11% -11% -12% -12% -12% -12% -12% -12% -12% -12% -12% Member Premium 14% 13% 13% 13% 13% 13% 12% 12% 12% 12% 13% Total Member Costs -3% -3% -3% -3% -3% -3% -3% -3% -3% -3% -3% NAR -16% -15% -15% -14% -14% -13% -13% -12% -12% -12% -13% NADS 149% 156% 163% 171% 181% 187% 201% 218% 238% 264% 193% LICS -18% -18% -19% -19% -19% -20% -20% -20% -20% -20% -20% LIPS 14% 13% 13% 13% 13% 13% 12% 12% 12% 12% 13% Total Government Costs 3% 3% 3% 3% 3% 2% 2% 2% 2% 2% 2% CGDP -28% -28% -27% -26% -25% -24% -23% -22% -21% -21% -24% Total Gross Allowed Costs -24% -24% -24% -24% -25% -25% -25% -25% -25% -25% -25% Scenario 2 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -13% -17% -18% -20% -20% -20% -20% -20% -20% -20% -19% Member Premium 12% 6% 4% 3% 3% 3% 3% 3% 3% 3% 4% Total Member Costs -4% -8% -10% -11% -11% -11% -11% -11% -11% -10% -10% NAR -17% -19% -20% -20% -19% -18% -18% -17% -17% -16% -18% NADS 144% 134% 132% 133% 141% 146% 159% 174% 191% 213% 157% LICS -20% -23% -25% -26% -26% -27% -27% -27% -27% -27% -26% LIPS 12% 6% 4% 3% 3% 3% 3% 3% 3% 3% 4% Total Government Costs 1% -3% -5% -6% -6% -6% -6% -6% -6% -6% -5% CGDP -31% -36% -38% -38% -37% -35% -34% -33% -32% -31% -34% Total Gross Allowed Costs -25% -29% -30% -31% -31% -31% -31% -31% -31% -31% -31% Scenario 3 Ten Year Year 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Impact Member Cost Sharing -12% -15% -17% -19% -19% -20% -20% -20% -20% -20% -19% Member Premium 12% 7% 4% 1% 0% 0% 0% 0% 0% 0% 2% Total Member Costs -4% -7% -10% -11% -12% -12% -12% -12% -12% -12% -11% NAR -17% -20% -21% -22% -23% -22% -22% -21% -20% -20% -21% NADS 146% 139% 138% 139% 146% 150% 163% 178% 197% 220% 162% LICS -19% -23% -25% -26% -26% -27% -27% -27% -27% -27% -26% LIPS 12% 7% 4% 1% 0% 0% 0% 0% 0% 0% 2% Total Government Costs 2% -3% -6% -7% -8% -8% -8% -8% -8% -8% -7% CGDP -30% -35% -36% -37% -36% -35% -35% -33% -32% -31% -34% Total Gross Allowed Costs -25% -28% -30% -32% -32% -33% -33% -33% -33% -33% -31% Page A - 9 U.S. Department of Health and Human Services Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

41 Milliman Client Report Appendix C ASPE Impact of Potential Changes to the Treatment of Manufacturer Rebates and Behavioral Changes Annual Percentage Cost Impact of Removing Manufacturer Rebates Scenario 4 Ten Year 2028 2024 2025 2026 2020 2023 2029 2022 Year 2021 2027 Impact -9% -9% -9% -10% -10% -10% -10% -10% -10% -10% Member Cost Sharing -9% 22% Member Premium 22% 22% 22% 21% 21% 21% 21% 21% 22% 22% 2% 2% 2% 2% 2% 2% 2% 2% 2% 2% Total Member Costs 2% -6% -6% NAR -5% -4% -4% -3% -3% -3% -2% -4% -5% NADS 161% 168% 176% 186% 192% 206% 224% 244% 269% 199% 154% -16% -14% -15% -15% -16% -16% -15% -16% -16% -16% LICS -14% 22% 22% 22% 22% 22% 22% LIPS 21% 21% 21% 21% 21% Total Government Costs 10% 10% 10% 10% 10% 10% 10% 9% 9% 10% 10% -20% -17% -18% -18% -19% CGDP -23% -23% -22% -21% -21% -20% Total Gross Allowed Costs -19% -19% -19% -20% -20% -20% -20% -20% -20% -20% -19% Scenario 5 Ten Year 2021 2022 2023 2024 2020 2026 2027 2028 2029 Year 2025 Impact -14% Member Cost Sharing -12% -13% -13% -14% -12% -14% -15% -15% -14% -11% Member Premium 13% 11% 10% 9% 8% 7% 5% 4% 3% 2% 6% -5% Total Member Costs -3% -4% -4% -6% -5% -6% -7% -8% -8% -6% NAR -17% -18% -18% -19% -20% -20% -21% -21% -22% -20% -17% 196% NADS 161% 169% 178% 183% 155% 213% 232% 257% 190% 149% LICS -19% -19% -20% -21% -21% -22% -23% -23% -24% -24% -22% 7% LIPS 11% 10% 9% 8% 13% 5% 4% 3% 2% 6% -3% -4% -3% -5% -6% Total Government Costs 2% 1% 0% -1% -2% -3% -28% -29% -29% CGDP -27% -25% -25% -25% -26% -26% -27% -28% Total Gross Allowed Costs -24% -27% -27% -28% -29% -29% -30% -31% -28% -25% -26% Scenario 6 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Impact -11% -12% -12% -12% -12% -12% -13% -13% -13% -13% -12% Member Cost Sharing 7% 13% 11% 10% 9% 8% 12% 6% 5% 4% 8% Member Premium -4% -3% -3% -4% -4% -6% -5% -5% -5% -6% Member Costs -5% NAR -17% -17% -17% -18% -18% -19% -19% -19% -20% -18% -17% 200% NADS 162% 170% 180% 186% 155% 218% 238% 264% 193% 149% LICS -19% -19% -20% -21% -21% -22% -23% -23% -24% -24% -22% 8% LIPS 12% 11% 10% 9% 13% 7% 6% 5% 4% 8% -2% Government Costs 2% 2% 1% 0% -1% -1% -2% -3% -4% -4% -24% -21% -23% -22% -25% CGDP -29% -28% -27% -26% -26% -24% Total Gross Allowed Costs -24% -26% -27% -27% -28% -28% -29% -29% -27% -25% -25% Scenario 7 Ten Year 2027 2023 2024 2025 2026 Year 2028 2029 2020 2022 2021 Impact -11% -11% -12% -12% -12% -12% -12% -12% -12% -12% -12% Member Cost Sharing Member Premium 12% 12% 11% 11% 13% 10% 10% 10% 10% 11% 11% Member Costs -3% -3% -3% -3% -3% -3% -3% -3% -3% -3% -3% NAR -16% -16% -15% -15% -15% -14% -14% -13% -13% -13% -14% NADS 148% 158% 164% 172% 178% 192% 209% 228% 253% 186% 153% -20% LICS -19% -19% -19% -20% -18% -20% -20% -20% -20% -18% LIPS 13% 12% 12% 11% 11% 11% 10% 10% 10% 10% 11% 1% Government Costs 2% 2% 1% 1% 3% 1% 1% 1% 1% 1% -22% -21% -21% -24% -23% CGDP -28% -28% -27% -26% -25% -24% -25% -25% -25% -25% -24% -24% -24% -24% Total Gross Allowed Costs -25% -25% -25% U.S. Department of Health and Human Services Page A - 10 Impact of Potential Changes to the Treatment of Manufacturer Rebates January 31, 2019

Related documents