UC Berkeley

This thesis studies the impacts of a consumer cost-sharing reform on several health economic outcomes. It is motivated in part by a need to better understand the effects of patient-cost sharing programs, which have become increasingly popular over time. Patient cost-sharing programs have become a particular subject of discussion as both public and private agencies are interested in effective ways of curbing the growth in health care spending.

The specific cost-sharing program analyzed is reference pricing, which was introduced by the California Public Employees' Retirement System (CalPERS) in 2011. Reference pricing changed health insurance benefits for CalPERS PPO members such that it was now relatively more expensive for them to use higher-cost facilities (versus lower-cost facilities) when receiving certain medical procedures in the post-reference pricing period. This thesis uses detailed medical claims data to compare health outcomes and cost savings between an unaffected control group and CalPERS members.

In the first chapter, Difference-in-Difference estimators are used to analyze the change in the share of patients using lower-cost health care facilities (versus higher-cost health care facilities) and the associated cost savings from this switching. Results indicate that patients are responsive to the reference pricing program and move to lower cost facilities in response to changes in the financial generosity of health insurance coverage due to reference price implementation. Furthering the analysis, instrumental variables methods are used to estimate the local average treatment effects in order to understand the treatment effects for the subpopulation of compliers. The focus is on measuring the changes in total cost-savings and health outcomes for this group. Additionally, compliers are characterized to understand how their observable characteristics tie into observed cost-savings from their movement to lower-cost health care facilities.

The second chapter estimates the quantile treatment effects of the reference pricing health insurance reform. Specifically, using the estimator of Firpo, 2007, estimations of how the quantiles of the unconditional distribution of medical spending change after the introduction of reference pricing are performed. Heterogeneous impacts along the distribution of medical spending are found for the quantile treatment effect estimates. This suggests that there is not simply a location shift in the distribution of medical spending resulting from the introduction of reference pricing. Rather, some parts of the distribution are more affected than others, particularly, the right-tail of spending. These heterogeneous impacts are also larger for certain procedures and smaller for others.

In the third chapter, I study how physicians' practice patterns respond to changes in patients' financial incentives, owing to the introduction of reference pricing. Motivating this analysis, I focus on issues of physician agency and how physicians' practice locations are a conduit by which cost-savings may be achieved. The results indicate that those physicians who have flexibility in where they can perform medical services are responsive to the change in the patients' cost-sharing and subsequently respond by increasingly treating CalPERS patients in lower-cost health care facilities (versus higher-cost facilities). This is measured relative to a control group who did not experience cost-sharing changes. These findings suggest that physician practice patterns are responsive to changes in patients' financial incentives.