Essays on the Effects of Health Insurer and Health Care Provider Organization on Patient Treatments

The United States health care system has garnered significant policy attention in recent years. The importance of the health care system is in part a fiscal concern, but also because of the role it plays in peoples' lives. In many cases, patients' lives greatly depend on the quality of the health care they receive. The organization of the health care system has undergone many dramatic changes in recent years and it is important to understand how these changes impacted the way the health care system functions. The effects of the various changes on health care spending and, in particular, on patient treatment outcomes remain poorly understood. In this dissertation, I investigate how three recent trends in the organization of the U.S. health care system have impacted the treatments that patients receive: the decrease in health insurer competition, the increase in hospital-physician integration, and the use of managed care in public insurance programs.

In Chapter 1, I study whether health insurer competition increases the use of costly treatments. Health insurers have an incentive to influence health care providers' treatment decisions to reduce their own reimbursement costs. If a single insurer is effective at inducing a provider to use less costly treatments, the provider may do so for patients of other insurers. While an insurer can reduce their own costs, spillover of their cost reduction can allow rival insurers to also benefit. I develop a theoretical framework demonstrating how competition can deter insurers from limiting costly treatments due to provider-level spillover concerns. I empirically test the relationship between Health Maintenance Organization (HMO) competition and Cesarean section (C-section) use. C-sections are a costly treatment that HMOs have an incentive and the potential ability to influence. I argue that HMOs' return to limiting C-sections should be lower at hospitals that contract with more HMOs due to potential spillover. I find that patients are more likely to receive C-sections at hospitals with lower HMO concentration - where spillover poses a greater deterrent to HMO cost reduction. The magnitude of this effect increases with the level of HMO competition in the market containing the hospital. Taken together, these results provide evidence that HMO competition can increase the use of C-sections at hospitals that contract with multiple HMOs.

In Chapter 2, I study whether hospital-physician integration affects health care utilization by altering the treatments patient receive. I estimate the effect of hospital-physician integration on the use of C-sections in childbirth using a sample of privately insured patients from California over 2005-2012. Childbirth is a convenient treatment setting to study utilization because it presents a binary choice between a high intensity, high cost procedure (C-section) and a comparatively low intensity, low cost alternative (vaginal birth). I am able to decompose the effect of hospital-physician integration on C-section use by the form of integration. I exploit heterogeneity in the various forms hospital-physician integration to investigate whether particular characteristics of integration affect C-sections use. This allows me to consider mechanisms suggested by previous literature through which integration potentially affects C-section use. I find that C-sections are 2% less likely at integrated hospitals than at hospitals with no physician affiliations. The negative effect of hospital-physician integration on C-section use is consistent across forms of integration where hospitals contract with physician practices and forms where hospitals own physician practices.

In Chapter 3, I study whether shifting Medicaid beneficiaries from Fee-for-Service (FFS) to Medicaid managed care (MMC) affects the treatments that they receive. I perform a case study of the effect of two California counties switching their Medicaid beneficiaries from FFS to MMC in October of 2009 on the use of C-sections in childbirth for their beneficiaries. I use hospital discharge data from California over 2006-2012 to estimate the “intent-to-treat” effect of counties switching from FFS to MMC on the likelihood that their beneficiaries receive C-sections. I find that switching from FFS to MMC is associated with an 11.9% increase in C-sections.