Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department


Major Professor

Padmaja Ayyagari, Ph.D.

Committee Member

Gabriel Picone, Ph.D.

Committee Member

Haiyan Liu, Ph.D.

Committee Member

Etienne Pracht, Ph.D.


Health, Hospital Pricing, Medicaid Rebate Program, Outpatient, Market Share, HHI


In 1992, Congress created the 340B Drug Pricing Program that requires drug manufacturers to provide outpatient drugs to participating hospitals with substantial discounts. Although the intent of the program is to allow covered entities to increase access to care for more vulnerable patients, hospitals are not required by law to pass on the discounts. Therefore, a concern is that hospitals might over-prescribe. This dissertation includes three chapters to study the effects of the 340B program on hospitals’ behavior changes:

Chapter 1 uses state aggregate hospital service spending data from the Centers for Medicare and Medicaid Services (CMS) to study the nation-wide impact of state 340B hospital participation on state hospital service spending. Controlling for state fixed effects, time fixed effects and state specific time trends, I find, on average, a 1 percentage point increase in state 340B hospital share leads to a 12.8% increase in state hospital service spending per capita. With only hospital spending data, analysis in this chapter cannot distinguish between a scenario where hospitals increase their spending to improve quality of care, consistent with the intent of the 340B program, and a scenario where hospitals are simply increasing spending without improving quality to maximize profit.

Chapter 2 complements the analysis in Chapter 1 by exploring the causal impact of the 340B program on hospitals’ medication cost, patient mix and quality of care. Working with 15 million ambulatory visits to Florida hospitals from 2005 to 2015, I use a series of difference-in-difference (DID) and synthetic control methods (SCM) based on the 2010 340B eligibility expansion, I find an average increase of $111.35 in medication cost per visit due to the 2010 expansion. Quantile regressions reveal that hospitals with the highest proportion of charity care and uninsured patients keep medication cost low and on the most expensive visits, they significantly reduce medication cost for patients. The remaining newly eligible hospitals significantly raise medication cost after the expansion. The increase becomes larger the more expensive the treatment is. Finally, I find some indications that newly eligible hospitals increased Medicaid patient mix and improved quality of care, but the evidence is not strong enough to be conclusive.

Chapter 3 further extends the analysis by examining the impact of market power on 340B hospitals’ behavior changes. Using the CMS nation-wide state aggregate data, I find the positive relationship between the state’s 340B hospital share and state aggregate hospital service spending is stronger when hospitals’ market share is higher. Working with the Florida data, using a series difference-in-difference-indifference (DDD) regressions, complemented by DID and SCM estimations, I find the 340B hospitals with low market shares seem to fulfill the mission of the program by keeping medication cost low, treating more low-income patients covered by Medicaid and Medicaid managed care and provide more charity to the communities. Compared to them, hospitals with high market shares significantly raise additional medication cost, treat fewer low-income patients but substantially more commercially insured patients. There are some signs of post-expansion quality improvement among all the newly eligible hospitals, measured by the post-operative adverse reaction rates, but heterogeneity exists in hospitals’ length of stay and nonroutine discharge rates. Hospitals with high market shares seem to treat more patients in their own outpatient facilities with a shorter length of stay. While the ones with low market shares experience increased length of stay, possibly due to worse health conditions among the additional Medicaid and Medicaid managed care patients they treat.

As a summary, this dissertation finds the average 340B hospital raise their medication cost upon participation in the program, but heterogeneity exists that some of them seem to fulfill the mission of the program. There are signs of quality improvement in the data, but future research could adopt more quality measures to study the cost-effectiveness on the price increase, as well as the welfare influence on the cost reduction.