Assessment of Price Changes of Existing Tumor Necrosis Factor Inhibitors After the Market Entry of Competitors

Abstract

This study uses wholesale medication acquisition costs and Medicare claims data to assess how prices of existing tumor necrosis factor inhibitors changed in response to the market entry of new tumor necrosis factor inhibitors.

Before 2009, etanercept, infliximab, and adalimumab were the only tumor necrosis factor (TNF) inhibitors approved by the US Food and Drug Administration for rheumatoid arthritis. Subsequently, 3 therapies have gained US Food and Drug Administration approval: subcutaneous golimumab (April 2009),¹ certolizumab pegol (May 2009),² and intravenous golimumab (July 2013).³ All 6 agents are brand-name drugs. We assessed how the prices of existing TNF inhibitors changed in response to the market entry of new TNF inhibitors.

Methods

Using 2006-2016 wholesale acquisition costs obtained from Analysource (reprinted with permission from First Databank),⁴ we calculated monthly estimates of the annual costs of TNF inhibitor treatment. Using claims data from a 5% random sample of Medicare beneficiaries, we also calculated monthly estimates of annual costs of TNF inhibitor treatment (gross drug costs for drugs typically reimbursed under Medicare Part D [etanercept, adalimumab, subcutaneous golimumab, and certolizumab pegol] and total claim payment amounts for drugs typically reimbursed under Part B [infliximab and intravenous golimumab]). Estimates based on wholesale acquisition costs and Medicare Part D payment data were adjusted for increases in manufacturer rebates reported for Medicare Part D.⁵ This study was approved by the University of Pittsburgh Institutional Review Board as exempt because it used unidentifiable data.

To test how new product entry affected the prices of existing TNF inhibitors, we constructed an interrupted time-series analysis with a linear model. This model regressed the annual cost of treatment of existing TNF inhibitors against a continuous variable for month, 2 indicator variables for each period after market entry of new drugs, and the interactions between them. Using estimates from this model, we estimated trends in costs that would have been expected in the absence of new agents’ market entry. All P values were from 2-sided tests, and results were deemed statistically significant at P < .05.

To understand how changes in costs affected different stakeholders, we evaluated trends in Medicare payments, out-of-pocket costs, coverage gap discounts, and other payments toward total costs of treatment with TNF inhibitors mostly reimbursed under Medicare Part D. Owing to lack of data, we were unable to assess how purchasing prices for drugs typically reimbursed under Medicare Part B changed over time.

Results

The trend in annual costs of treatment estimated with wholesale acquisition costs significantly increased after market entry of new products (Figure 1).⁵ When estimates were based on Medicare payment data, the trend increased significantly after market entry of intravenous golimumab.

Figure 1. Observed and Expected Trend for the Annual Costs of Treatment With Tumor Necrosis Factor (TNF) Inhibitors, 2006-2016.

A, Observed annual costs of treatment with all TNF inhibitors, based on wholesale acquisition costs (WAC). B, Observed annual costs of treatment with all TNF inhibitors, based on Medicare payment data. C, Expected annual costs of treatment with existing TNF inhibitors compared with January 2006, based on WAC. D, Expected annual costs of treatment with existing TNF inhibitors compared with January 2006, based on Medicare payment data. Expected annual costs were estimated from regression models described in the Methods. Estimates based on WACs and Medicare Part D payment data were adjusted for increases in manufacturer rebates reported for Medicare Part D.⁵ Period 1 denotes the period before the entry of new drugs (January 2006–April 2009). Period 2 denotes the period between April 2009 (approximately when subcutaneous golimumab and certolizumab pegol entered the market) and July 2013, when intravenous golimumab entered the market. Period 3 denotes the period between the entry of intravenous golimumab in July 2013 and the end of the study period (December 2016). The dotted lines represent the market entries of new TNF inhibitors. All estimates for annual costs of treatment were based on dosing recommendations for a standard 80-kg patient with rheumatoid arthritis. IV indicates intravenous; SQ, subcutaneous.

^aMostly reimbursed under Medicare Part B, and whose annual costs of treatment based on Medicare payment data were estimated using total claim payment amounts under Medicare Part B.

^bMostly reimbursed under Medicare Part D, and whose annual costs of treatment based on Medicare payment data were estimated using gross drug costs under Medicare Part D.

Using wholesale acquisition cost data, annual treatment costs with existing TNF inhibitors increased by 144% from April 2009 to December 2016 after new drug entry (from $15 809 to $38 574), compared with a 34% increase expected in the absence of new drugs’ entry (from $15 809 to $21 184). Using Medicare data, annual treatment costs increased by 139% (from $14 901 to $35 613), compared with a 43% increase expected in the absence of new drugs’ entry (from $14 901 to $21 308). Medicare spending increased in parallel with increases in annual treatment costs (Figure 2); however, out-of-pocket costs and manufacturer coverage gap discounts remained relatively constant over time.

Figure 2. Observed Trends for the Contribution of Medicare Payments, Out-of-Pocket Costs, Manufacturer Coverage Gap Discounts, and Other Payments Toward Total Annual Costs of Treatment With Tumor Necrosis Factor Inhibitors Covered Under Medicare Part D, 2006-2016.

Each panel shows the trend in annual costs of treatment with each tumor necrosis factor inhibitor (A-D) reimbursed under Medicare Part D, which were estimated on the basis of gross drug costs (dark-blue series). Each panel also shows the contribution of Medicare payments, out-of-pocket costs, manufacturer discounts in the coverage gap, and other payments toward total costs. Other payments include Medicare Part D low-income subsidy, payments made by the Part D plan for benefits beyond the standard Part D benefit, payments made by third-party payers (eg, group health plans, worker’s compensation, and governmental programs such as the Veterans Administration and TRICARE), and payments made by qualified state pharmacy assistance programs or charities. Estimates were adjusted for increases in manufacturer rebates reported for Medicare Part D.⁵ SQ indicates subcutaneous.

Discussion

Annual treatment costs with existing TNF inhibitors increased after the entry of 3 new agents. If cost trends had not changed after the entry of new products, costs of etanercept, infliximab, and adalimumab in December 2016 would have been 40% to 45% lower than they actually were. These increases were born solely by Medicare, while patient out-of-pocket spending remained flat. In addition, these increases were not offset by manufacturer discounts in the Medicare Part D coverage gap.

The rising costs of existing products may reflect manufacturers’ opportunism in response to payers’ increased willingness to pay for TNF inhibitors after market entry of new, more expensive agents. Intravenous and subcutaneous drugs followed different trends. For instance, observed costs of infliximab fell under the expected range without market entries.

Using TNF inhibitors as a case study, we showed that increased competition in the pharmaceutical market does not necessarily translate into price reductions. Our findings illustrate a market failure contributing to the rising costs of prescription drugs.