Essays on health economics
Date
2019
DOI
Authors
Fonseca Martins, Bruno Duarte
Version
OA Version
Citation
Abstract
This dissertation is composed of three essays studying agents' behavior in the health care sector - insurers, patients, and providers - facing different regulatory settings.
The first essay studies how vertical integration between pharmacies and
insurers affects Medicare Part D premiums. I propose a model of insurer-pharmacy bargaining, which suggests that exposure to a vertically integrated firm should increase other insurers' premiums while lowering the integrated insurer's. I test this using plan-level data from 2006 to 2017, exploiting CVS's acquisition of Target pharmacies. Exposure to this integration increases non-CVS premiums as expected; however, CVS premiums are unchanged. I estimate a demand model of plan choice to show that consumers value a large network and having CVS pharmacies in their plan's network.
The second essay (joint with Randall P. Ellis and Wenjia Zhu), estimates within-year price elasticities of demand for detailed health care services. We use an instrumental variable strategy, in which individual monthly cost shares are instrumented by employer-year-plan-month average cost shares. We show that using backward myopic prices gives more plausible results than using forward myopia. Using 171 million person-months from 73 employers between 2008-2014, we estimate an overall demand elasticity by backward myopic consumers of -0.44, with high demand elasticities for pharmaceuticals, specialists visits, MRIs and mental health/substance abuse, and lower for prevention and emergency departments.
The third essay (joint with Luis Filipe), evaluates how doctors in an emergency department react to the number of patients waiting for treatment. Our outcomes reflect the time spent with the patient, the intensity of treatment and discharge destination. Using visit-level data in one Lisbon-area hospital, we use a fixed effects model to exploit variation in the queue size, while addressing endogeneity using the number of arrivals to the hospital as an instrumental variable. Results show that doctors discharge patients more rapidly as queues increase, and this effect is stronger for patients that do not have life-threatening conditions. We also find that the intensity of diagnosis/treatment procedures decrease when patients face longer queues, driven by the extensive margin. Finally, doctors are less likely to admit patients to inpatient care.