Document Type


Date of Degree

Fall 2009

Degree Name

MS (Master of Science)

Degree In


First Advisor

John M. Brooks

First Committee Member

William Doucette

Second Committee Member

Elizabeth Chrischilles

Third Committee Member

Yang Xie


Background: With healthcare costs increasing faster than inflation in the United States. it is expected that fewer people will be able to afford health care in the future without changes to the healthcare system. It has been argued that much of the healthcare utilization presently observed is unnecessary and that a large portion of this wasteful spending can be attributed to the characteristics of local area physician supply and specialty mix. These arguments are based on observed positive correlations between local area health utilization and the local area supply of physicians and physician specialists. Commentators suggest that greater physician supply leads to the use of more physicians by patients resulting in diffuse, uncoordinated, and wasteful care. There have been calls for modifying the physician training system in the United States based on these correlations. However, we have found no studies demonstrating direct relationships between local area physician supply and the use of physicians by individual patients. Studies in physician-induced demand theory have examined the relationships between physician supply and healthcare utilization, whereas studies of physician referrals have only investigated the effects of local supply on referral rates not the number of physicians used by patients. Method: We isolated a set of patients with a consistent diagnosis and course of treatment (stage III colorectal cancer Medicare patients receiving surgery and chemotherapy but not radiation therapy) and investigated whether local area physician supply leads to the use of more physicians by these patients and more diffuse care across physicians during their first treatment course. Result: Our results show that not only the specialists' characteristics but also the market level variable as local physician supply affect the care dispersion in terms of numbers and specialty mix of physicians treating colorectal cancer patients with the same condition. Discussion: According to these findings, it is suggested that local physician supply has the causal relationship between the patterns of care dispersion. Supply of primary care physicians and oncology specialists tend to spur the physician utilization across four specialty groups which are more likely to have discretionary utilization.


care dispersion, local physician supply, Unique physician use


vi, 46 pages


Includes bibliographical references (pages 43-46).


Copyright 2009 An-Chen Fu