Econometric Analyses Of Health Care Markets
| dc.contributor.author | LeCates, Joseph | en_US |
| dc.contributor.chair | Nicholson, Sean | en_US |
| dc.contributor.committeeMember | Kenkel, Donald S. | en_US |
| dc.contributor.committeeMember | Lucarelli, Claudio | en_US |
| dc.date.accessioned | 2013-02-22T14:16:10Z | |
| dc.date.available | 2017-09-26T06:01:00Z | |
| dc.date.issued | 2012-05-27 | en_US |
| dc.description.abstract | This dissertation uses novel econometric methods to correct for biases stemming from imperfect information in health care markets. Because of the nature of health and the endogeneity of treatment choices, modeling the demand for health and outcomes in the health care market can be difficult, and special econometric approaches are needed to identify causal effects. The first essay extends empirical models of consumer behavior to allow individuals to learn across bundled goods, and applies the framework to physician treatment choice. Such perception spillovers can affect the adoption rate and steady-state market shares of goods, particularly for market entrants which are strong complements to available products. I develop and estimate a Bayesian model of learning which allows the quality signal arising from one particular bundle to be correlated with physician beliefs about the quality of all other potential bundles in the context of oncology. Using cancer surveillance data with patient and physician characteristics, I find that physicians do learn across bundles, though the magnitudes of between-regimen correlations in quality vary substantially. The second essay investigates how the atypically heaped smoking cessation rate computed from retrospective survey questions biases coefficient estimates in discrete time hazard models. If this characteristic of the data is ignored, estimates of the effect of price on smoking cessation decisions can be attenuated up to sixty-five percent. This paper uses Monte Carlo simulation to compare an adjusted likelihood approach and several ad hoc approaches to correct the bias and finds that the adjusted likelihood approach and one of the ad hoc approaches perform well. The methods are then applied to the Tobacco Use Supplement to the Current Population Survey and the results are consistent with the simulation analysis. Returning to the treatment of cancer, the final essay explores how average costs and outcomes change in the United States as physicians alter chemotherapy treatment patterns over time. Considering chemotherapy for metastatic colorectal cancer and metastatic nonsmall cell lung cancer, and adjuvant therapy for breast cancer, I find that between 2003 and 2006 quality-adjusted life years of survival increased by 11.2 percent on average, but costs increased by 143.8 percent. | en_US |
| dc.identifier.other | bibid: 8251314 | |
| dc.identifier.uri | https://hdl.handle.net/1813/31445 | |
| dc.language.iso | en_US | en_US |
| dc.subject | Consumer Learning | en_US |
| dc.subject | Physician Behavior | en_US |
| dc.subject | Smoking Cessation | en_US |
| dc.subject | Retrospective Data | en_US |
| dc.subject | Chemotherapy | en_US |
| dc.subject | Cost Effectiveness | en_US |
| dc.title | Econometric Analyses Of Health Care Markets | en_US |
| dc.type | dissertation or thesis | en_US |
| thesis.degree.discipline | Economics | |
| thesis.degree.grantor | Cornell University | en_US |
| thesis.degree.level | Doctor of Philosophy | |
| thesis.degree.name | Ph. D., Economics |
Files
Original bundle
1 - 1 of 1