Comparison of U.S. and InternationalPrices for Top Medicare Part B Drugsby Total ExpendituresOctober 25, 2018Executive SummaryThe prices charged by drug manufacturers to wholesalers and distributors (commonly referred to as exmanufacturers prices) in the United States are 1.8 times higher than in other countries for the top drugsby total expenditures separately paid under Medicare Part B. U.S. prices were higher for most of thedrugs included in the analysis, and U.S. prices were more likely to be the highest prices paid among thecountries in our study.1. IntroductionRecently there has been increased interest inhow U.S. drug prices compare to those of otherdeveloped countries. Much of this interestfocuses on pricing for pharmacy-dispenseddrugs, which account for about 72 percent oftotal prescription drug spending.1 This paper,instead, focuses on prices for non-retail drugs,which are generally physician-administered.In the fee-for-service Medicare program,outpatient prescription pharmaceuticals arecovered under two separate voluntary benefits.Drugs dispensed by retail and specialtypharmacies to patients for self-administrationare typically covered under the Medicare Part Dprogram. Part D is operated by commercialinsurance companies that negotiate formularyplacement and prices with drug manufacturersand payment rates with pharmacies. This1U.S. Department of Health and Human Services, Officeof the Assistant Secretary for Planning and Evaluation(ASPE). Observations on Trends in Prescription DrugSpending. March 2016. Available gspending.pdf.approach is one reason why spending growthunder Part D has remained below its initialspending projections.Drugs more typically administered to patientsby healthcare practitioners, however, arecovered and paid under Medicare Part B, whichis part of the fee-for-service traditionalMedicare benefit.2 Under Part B, providers andsuppliers “buy and bill” these types of drugs.Since 2005 for physicians, and 2006 for hospitaloutpatient departments, Medicare has paidsuppliers and providers based upon the AverageSales Price (ASP) for each product, as reportedby manufacturers to the Centers for Medicare &Medicaid Services (CMS).3 Physician officesthat buy and bill Part B drugs are paid 106% ofthe drug’s ASP. Depending on a hospitaloutpatient department’s participation in a safety2Medicare Part B covers some self-administered drugsthat were added to the benefit by Congress prior to thecreation of Part D. These self-administered drugs are notthe subject of this paper.3 Section 1847A of the Social Security Act governspayments to physicians for certain Part B drugs. Section1833(t) governs payments to HOPDs, and allows the useof 1847A payment rates. By 2006, CMS cited thisauthority, and by 2014, was paying HOPDS based upon it.Page 1

Comparison of U.S. and International Prices for Top Spending Medicare Part B DrugsU.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluationnet drug pricing program, hospitals arereimbursed either 106 or 77.5 percent of ASP.4ASP is reported by manufacturers to CMS as thetotal sales to all purchasers minus the priceconcessions granted to these purchasers andeventual end users, i.e. physicians and hospitals,with certain exceptions. For example,manufacturers may offer a rebate to physicianspecialists to prescribe and administer theirproduct over a competitor’s. The sale of theproduct to a wholesaler and the price concessiongranted to the physician are both accounted forin the ASP. Purchases and price concessions orrebates offered under federal discount programs(such as the Veterans Health Administration, theMedicaid Drug Rebate Program and StateSupplemental Rebate Agreements, and the 340BDrug Discount Program) are excluded from theASP calculation.5Unlike the situation with traditional, pharmacydispensed drugs, payers are not typicallyinvolved in the prescribing, purchasing, ordispensing decision for physician-administereddrugs, and there is therefore more limitedprivate-payer negotiation for formularycoverage. Specifically, the Medicare programhas not applied the types of formularymanagement practices that are commonly usedto achieve better value for self-administereddrugs by commercial insurers, including thosesponsoring Medicare Advantage or Part Dplans, which were recently granted newauthority by CMS to use formulary managementpractices such as step therapy for Part B drugs.6However, this flexibility does not extend to thefee-for-service Medicare Part B. Many have4HOPDs and off-campus locations that participate in the340B Drug Discount Program are reimbursed a lower rateto account for significantly reduced acquisition costs. See82 Fed. Reg. 52356.5 See SSA 1847A(c)(2) for exclusions.6 Source: CMS, “Prior Authorization and Step Therapy forPart B Drugs in Medicare Advantage,” available PlansGenInfo/Downloads/MA Step TherapyHPMS Memo 8 7 2018.pdf.also suggested that the 6 percent add-onpayment currently in place for physician officesand some hospitals may incentivize the use ofthe highest priced clinically beneficial product.Differences in coverage for drugs under Part Bcompared to Part D may have contributed to anacceleration in spending for physicianadministered drugs, relative to spending growthunder the approach taken under Part D.Specifically, spending for Part B drugs hasdoubled since 2006, despite overall low FFSenrollment growth.7 In Part D, althoughenrollment continues to grow, annual and perbeneficiary expenditure growth rates are lowerthan in Part B. Put another way, per-beneficiaryspending under Part B rose 7 percent and then11 percent annually over two five-year periods(2006-2011 and 2011-2016) while Part D perbeneficiary spending increased only 3 percentper year in the same five-year intervals. SeeFigure 1.Figure 1: Part B and D Per-Beneficiary NetExpenditures Growth Rates, 2006-20168Administration of coverage and payment of PartB drugs is delegated to regional Medicare7While overall Medicare enrollment has grown, a growingproportion of beneficiaries have enrolled in MedicareAdvantage plans, whose spending is not reflected inFigure 1.8 Source: Medicare Trustees Report from 2016 (for 2006)and 2018 (for 2011 and 2016); Part B annual NationalSummary Files; OPPS Final Rules from 2008, 2013, and2018. Percent changes reported are annually over thefive-year periods shown. Net payments excludebeneficiary cost-sharing.Page 2

Comparison of U.S. and International Prices for Top Spending Medicare Part B DrugsU.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and EvaluationAdministrative Contractors (MACs). Broadrulemaking by CMS and ASP-based paymentlimits are applicable at the national level, whileeach MAC can determine for each patient if aPart B drug claim is reasonable and necessary.While the Medicare program and MACs do notuse formulary decision-making to restrict thecoverage and payment of Part B drugs, anumber of other economically comparablecountries do for these types of drugs. Thoughthese countries use their national health systemsto negotiate lower prices in exchange for marketaccess, drug manufacturers retain the choicewhether to offer price concessions beyond thoseavailable to payers in the United States. Tobetter understand the effect of these negotiationson prices paid for physician-administered drugs,ASPE compares in this paper the prices paid forphysician-administered drugs in the U.S. toother selected countries.those in the U.S. Prices in Japan were lowerthan U.S. prices in the 2004 Department ofCommerce study but higher in Danzon andFurukawa (2003, 2005, 2006). This is likely dueto methodological differences that result indifferent products being included in the study.For instance, package sizes in Japan differsignificantly from elsewhere particularlybecause doses tend to be lower in Japan.Figure 2: Reported Brand Drug Price Differentialsfrom Price Index-Based Studies (U.S. 100)2. national drug prices has significantlimitations, which we sought to address in ouranalysis. Namely, few of these analyses use datafrom after 2007, and there are no specificanalyses of the exact set of drugs that we areinterested in comparing.Drug prices are generally higher in the U.S.based on price comparisons in the literature. Intheir recent systematic literature review,Kesselheim and Avorn (2016) estimate that U.S.prices were more than twice as high as those inother, similar countries. However, they includebilateral comparisons combined into a metaanalysis, which may overstate price differences.Instead, below and in Figures 2 and 3, wecompare seminal original research publications.Comparing sample baskets of branded andgeneric prescription drugs in the U.S., allcompared countries except Japan and Mexicohad prices that were at least 20 percent less thanNarrowing to branded drugs, the literaturedemonstrates similar results, with prices higherin the U.S. than in all countries except Mexicoamong the three sources comparing brandeddrug prices. Two of the selected studiescompared differences among biologics; thesestudies demonstrated mixed results for the drugclass (Figure 3). We would note that Kanavos etal. (2013) compared a different price (averageretail price per standard unit) than Danzon &Furukuwa (2006), which used ex-manufacturerprices, likely explaining the divergent results.Some of the variability may be related toproduct availability, per the authors’conclusions.Page 3

Comparison of U.S. and International Prices for Top Spending Medicare Part B DrugsU.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and EvaluationFigure 3: Reported Price Differentials from StudiesComparing Biologics (U.S. 100)There are important challenges in comparingdrug prices across countries, includingambiguity in actual U.S. prices, assumptionsand limitations related to available data on drugssales and volume, and mismatches betweendrugs and dosage forms available in differentmarkets. Despite these challenges, updatedestimates of price differentials are needed.3. MethodsIn this paper, we calculate the price per gram ofeach included product in each selected country.We aggregate sales outside the U.S. andcompare an average volume-weightedinternational price to the U.S. price. Below wedescribe each source underlying each aspect ofthis calculation in more detail.Data SourcesInternational and Domestic Acquisition CostData. ASPE purchases licenses to several dataproducts maintained by IQVIA (formerlyknown as Quintiles-IMS Institute for HealthcareInformatics or QIIHI). For this study, we usedtwo products that contain acquisition pricingand volume information. First, MIDAS isIQVIA’s international sales and volumedatabase, which contains sales information(price and quantity) for more than 50 countriesthrough the second quarter of 2018, from asearly as 2013. Sales are stated in local and U.S.currency, as of the transaction date or currentdate, as desired.For our analysis, we use ex-manufacturer prices9(sometimes called the ex-factory price) stated inU.S. currency on the transaction date. IQVIAalso provides sales and volume information onU.S. domestic sales in its National SalesPerspective (NSP) database. We used thisdatabase to facilitate the accurate comparison ofdrug quantities in different package sizes and toaccount for how overfill is treated across thedatabase. We describe how we make thesecorrections later in this section.Medicare Program Data. In order to identifystudy drugs, we used two files that summarizedMedicare program spending on Part B drugs.First, for physician offices, we used the Part BNational Summary File for 2016, summingallowable charges and payments across allHealth Care Procedural Coding System(HCPCS) J-codes and select Q-codes asappropriate. Second, for hospital outpatientdepartments, we used the CY 2018 MedicareHospital Outpatient Prospective System’sunderlying cost statistics files to identifyutilization and spending for separately payabledrugs in 2016, and applied the listed paymentrates to these. From each file we identified the20highest-spendingproducts.Furtherdiscussion regarding drug selection is below.We use Medicare’s quarterly HCPCS ASPpayment allowances for the third quarter of2018 to compare prices paid in the U.S. andabroad through the first quarter of 2018. SinceASP is calculated based on the most recentlyavailable quarter’s manufacturer’s sales and isassociated with a two quarter lag, third-quarter9Ex-manufacturer price is the price received bymanufacturers of a product, including discounts applied atthe point of sale. In comparison, invoice price is the pricepaid by the dispenser of a product, including on-invoicediscounts. To the extent that a product is sold throughwholesalers, this price will differ by the wholesaler’smarkup.Page 4

Comparison of U.S. and International Prices for Top Spending Medicare Part B DrugsU.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation2018 ASP is the best temporal approximation tothe actual purchase prices paid in the firstquarter of 2018. We also use these quarterlyfiles to identify exactly which products areincluded in each selected drug’s HCPCS code,which we describe in the section to follow. In allcases, the Medicare ASP payment files we useare publicly available. No manufacturerconfidential information was collected for use inthese analyses.Drug SelectionASPE compiled data on the top 20 drugs basedon total Medicare reimbursement to ts (HOPDs), or overall underMedicare Part B in 2016, which is the mostrecently available publicly accessible data (asdescribed above). Drugs are defined in thisstudy as each unique HCPCS code assigned byCMS. We inc