
Optimal Pricing and Treatment Policies in Health Care
Abstract
Health care decision makers are faced with many types of uncertainty. There may be uncertainty in the benefit a new treatment provides, uncertainty in the demand for the new treatment, and uncertainty in the drug approval process. In three essays, I study the impact of various types of uncertainty within health care systems.
In the first essay, I study the welfare properties of six practical pharmaceutical pricing and access policies. Using a game theoretic approach, I find that when demand can be influenced by costly marketing effort, there are meaningful differences to the desirability of various pricing and listing agreements, compared to the results in existing literature. I find that all non-value-based agreements result in at least one type of inefficiency. I find that a value-based policy with risk-sharing is preferred by the manufacturer and from a societal perspective, while the payer's preference depends on the manufacturer's negotiating power and the risk-sharing rebate rate.
In the second essay, I study the impact of health care fragmentation on treatment policies. I consider the scenario where a patient's health care is covered by multiple payers across their lifespan. I formulate a multi-decision-maker Markov decision process to capture the payers' repeated intervention decisions and I partition the optimization problem using a threshold patient age that defines when the patient transitions from one payer to the next. I find that a fragmented health care system always results in a treatment policy that is a subset of the treatment policy in a centralized system and that a simple transfer payment between payers can coordinate the system.
In the third essay, I study the scenario where payers in a fragmented health care system incur different costs for the same treatment. Using a similar multi-decision-maker Markov decision process as in my second essay, I find that when payers incur different costs of treatment, that over- and under-treatment will occur in a fragmented system compared to a centralized system. I find that when pricing is endogenous, the payers will prefer a setting where coordination is not possible and the manufacturer will prefer coordination.