Value-Based Pricing and Its Implications for the Newly Announced Medicare Negotiated Price Under the Inflation Reduction Act Free

The Centers for Medicare & Medicaid Services (CMS) recently announced negotiated prices for the first 10 medications selected under the Inflation Reduction Act (IRA). Among these, four medications primarily used to treat diabetes, sitagliptin (Januvia), dapagliflozin (Farxiga), empagliflozin (Jardiance), and insulin aspart (Novolog and Fiasp), saw the largest price cuts, ∼70% from their 2023 list prices. While these reductions are projected to yield significant Medicare savings, concerns have been raised about the transparency and methodology of the pricing decisions. CMS’s pricing process significantly diverged from the value-based models used in other high-income nations and prohibited the use of quality-adjusted life-year (QALY) in the process (1).

This study aims to contextualize how the IRA-negotiated prices for the four diabetes medications compare with those estimated within the framework of value-based pricing.

We employed BRAVO, a well-validated and widely used microsimulation diabetes model (2), to project lifetime QALY gains and to estimate the corresponding price ceilings for the four IRA-listed diabetes medications under various willingness-to-pay (WTP) thresholds (i.e., $0/QALY, $25,000/QALY, $50,000/QALY, and $100,000/QALY). Our simulation population was based on a nationally weighted sample of 1.55 million Medicare beneficiaries with diabetes from the National Health and Nutrition Examination Survey. Sulfonylurea was selected as the common low-cost comparator for Januvia, Farxiga, and Jardiance for glycemic control, a glucagon-like peptide 1 receptor agonist (GLP-1RA) was selected as the comparator for Farxiga and Jardiance for cardiovascular and renal risk management, and regular human insulin was used as the comparator for Novolog and Fiasp. Medication and service costs from a health care system perspective were estimated using Medicare data, while QALYs were calculated using the Health Utilities Index diabetes equation. Both were adjusted with a 3% annual discount rate.

The pricing scale is presented in Fig. 1. Farxiga was associated with an additional 0.077 QALY compared with sulfonylureas and an additional 0.030 QALY compared with GLP-1RA, corresponding to a 30-day price ceiling of $359, $416, $474, and $589 under WTP thresholds of $0, $25,000, $50,000, and $100,000 per QALY, respectively. Jardiance achieves an additional 0.083 QALY compared with sulfonylureas and an additional 0.038 QALY compared with GLP-1RA and had price ceilings of $373, $435, $496, and $620 under the same thresholds. Januvia accrued an additional 0.037 QALY with corresponding price ceilings of $78, $127, $176, and $274. Novolog and Fiasp, associated with a 0.046 QALY gain compared with regular human insulin, had price ceilings of $213, $222, $230, and $248 under the four WTP thresholds.

The scale provides value-based price benchmarks for diabetes medications, offering stakeholders a clear and transparent reference point for evaluating how negotiated prices compare with a value-based pricing framework. In 2023, the list prices of Novolog, Fiasp, and Januvia exceeded their value-based prices, even at a $100,000/QALY threshold, suggesting overpricing (3). In contrast, the prices for Jardiance and Farxiga were just below this threshold, indicating cost-effectiveness at a higher WTP threshold. The IRA negotiations significantly reduced drug prices, achieving a cost-saving level.

The implementation of these negotiated prices creates both opportunities and challenges. While these lower prices could save Medicare billions, it is crucial to ensure that pricing does not hinder pharmaceutical innovation, especially when the added therapeutic value of new treatments may not be fully compensated. In most cases, strict cost-saving levels are not required in price negotiation, as health systems globally accept some level of incremental health care costs if they lead to improved health outcomes.

Although the BRAVO (Building, Relating, Assessing, and Validating Outcomes) model is extensively validated, modeling studies are inherently influenced by the parameters and assumptions applied, and their results should be interpreted within the context of those limitations. Additional sensitivity analyses are warranted to explore the robustness of our findings. Note that behavioral aspects of diabetes management, such as adherence to treatment or lifestyle changes, are not comprehensively integrated into the simulation. Additionally, the risk equations used in the BRAVO model are based on data from the ACCORD (Action to Control Cardiovascular Risk in Diabetes) cohorts, which may not fully represent evolving population health dynamics.

Using QALY in pricing has faced criticism for potential discrimination against the elderly and disabled in the U.S. (4). However, such concerns can be overcome by choosing different WTP thresholds across population subgroups (i.e., equity weights) rather than dismissing the utility of QALY (5). The Medicare pricing process could benefit from incorporating QALYs, or similar approaches, as one of the factors and providing more transparency on how the final prices are determined. This would help ensure that the pricing decisions are fair, equitable, and well-understood by all stakeholders.