Value-Based Pricing for Drugs With Uncertain Clinical Benefits

This is the title of a recent paper by Jiao et al. (2025). Previous studies (e.g., Goldman et al. 2013, Goldman et al. 2023) have argued for a 3-part pricing model. Under a 3-part pricing model, drugs would come to market quickly under accelerated approval with a relatively lower price. As clinical trials and other evidence produced more (less) robust evidence, the price would rise (fall). Then, when the medication’s market exclusivity ended, prices would fall as generics (or biosimilars) entered the market.

The paper by Jiao et al. (2025) provides the theoretical foundations for why this three-part pricing approach is so sensible. At a high level, the approach combines expected utility theory with value of information (VOI) metrics to inform value-based prices (VBP) of new medical technologies. To implement this approach, the authors rely on a five step methodology. First, the authors apply a standard, CEA-based approach which assumes that payers are risk neutral and pay based on the value-based price of the new therapy (VBP). Second, the authors loosen the assumption that payers are risk neutral. One would think that most payers are risk neutral, but the authors justify potential payer risk aversion as follows:

Current policies reflect a prevailing risk aversion among policymakers toward AA [accelerated approval] drugs. For example, Medicaid requires higher rebates for AA drugs to counterbalance uncertainty (Rome and Kesselheim 2021).

Third, the authors apply the expected value of perfect information (EVPI) in a way that incorporates potential payer risk aversion to create a risk‐adjusted value‐based price (rVBP). Fourth, the author recognize that there is no such thing as perfect information. In the real world, even large clinical trials still leave some uncertainty. Thus, in a fourth step the authors argue for the use of Expected Value of Sample Information (EVSI). EVSI assesses the expected value of conducting a future study with a finite sample size; the larger the sample size, the closer EVSI comes to EVPI. Finally, in the fifth step, rVBP is set based on the following decision rule:

…the insurance payer applies the decision rule that the VBPNT should be set so that the expected utility in this scenario equals the expected utility in the benchmark scenario. In other words, the expected utility with current uncertainty should align with the utility without uncertainty.

The authors also conduct a case study for a hypothetical treatment that would receive accelerated approval for the treatment of triple‐negative breast cancer (TNBC). The case study showed that drug prices would receive a material price reduction when confirmatory trials are not yet available.

For the hypothetical AA [accelerated approval] drug, the traditional VBP was calculated at $2000 per month, based on the drug’s expected cost‐effectiveness without accounting for uncertainties in its clinical benefits. However, when applying rVBP with the idealized benchmark of no uncertainty, the rVBP was set at $1900, $1400,and $1000 for [absolute] risk aversion coefficients (α) of 0.0001, 0.0005, and 0.001, respectively. These adjusted prices reflect the need to lower the price as decision‐makers’ risk aversion increases, compensating for the uncertainty they face. When the uncertainty level was based on the expected results of a future confirmatory trial, the rVBP was slightly higher, with adjusted prices of $1900, $1700, and $1500 per month for the same [absolute] risk aversion coefficients.

For more details, including all the mathematical derivations and more details on the case study, you can read the paper here.

Innovation in the pharmaceutical industry: New estimates of R&D costs

Originally posted at Healthcare Economist.?

The views expressed herein are those of the author and not necessarily the views of?FTI Consulting, Inc., its management, its subsidiaries, its affiliates, or its other professionals.