Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Dr. Gregory S. Zaric


Many new and expensive drugs have been introduced in the past 10 years. However, at the time of introduction, the effectiveness of these drugs outside of clinical trials is often unknown. This creates a risk to third-party payers, as the outcome of these drugs in real-world practice is uncertain at the time of introduction. A pay-for-performance risk-sharing agreement is a type of contract that shares part of this risk with the manufacturer by linking the performance of a drug to the manufacturer’s revenue. This dissertation consists of three essays to examine the performance of two types of pharmaceutical pay-for-performance risk-sharing agreements.

In my first essay I examine the performance of a pay-for-performance risk-sharing agreement in which patients are assessed at some evaluation time to determine their response to the drug. The manufacturer rebates to the payer a proportion of the sales from all patients excluding the sales from those responding at the evaluation time. I model disease progression using a continuous time Markov chain with uncertain transition rates. I address the following questions regarding the performance of this agreement: What is the optimum evaluation time and under what conditions will the manufacturer make a profit? What is the distribution of the manufacturer’s profit resulting from different sources of uncertainty?

In the second essay I extend the model developed in the first essay to calculate the net monetary benefits of the payer and identify the conditions under which both parties have incentives to introduce the new drug.

The third essay focuses on the analysis of a risk-sharing agreement in which patients are prescribed a drug only if their probability of response lies within a range of success probabilities. The payer determines this range such that the use of the drug is cost-effective. I generalize from the existing literature by allowing the rebate to be different from the price of the drug and incorporating two types of administrative costs. I seek to answer two important policy questions: First, under what conditions does the payer benefit from the agreement? Second, under what conditions does the agreement become welfare-improving?