Defense Date


Document Type


Degree Name

Doctor of Philosophy


Health Related Sciences

First Advisor

Dianne Jewell


Background and Purpose A debate over the practice of physician self-referral has been ongoing in health care since the 1980s. At issue is the practice of physicians who refer patients to facilities in which they share a financial interest, a phenomenon known as referral for profit. Physician investment or ownership interest in ancillary (e.g., physical therapy) services has been shown to have an impact on utilization rates, costs, access to care, and quality of care. What has not been identified in previous research is the influence of physicians‘ selective referral on competing clinics, particularly the hospital-based outpatient centers that share their health care markets. The purpose of this research was to examine the relationship between the emergence of orthopaedic physician owned physical therapy services (POPTS) and changes in physical therapy referrals made to two groups of not-for-profit, hospital-based outpatient physical therapy (OPPT) centers in one health care market. Methods This study examined the referrals made by orthopaedic physicians to two large hospital systems in the Orlando, Florida, outpatient physical therapy market between 1999 and 2007. This study was conducted using existing proprietary databases maintained by the Orlando Regional Healthcare System (ORHS) and Florida Hospital System (FHS), as well as phone surveys conducted by the primary investigator. Information regarding the orthopaedic physicians‘ ownership status and the patients‘ payer types was combined into analyses to determine if physician status was related to the number of physical therapy patients from each payer type referred, or the number of total referrals made to the hospital-based outpatient physical therapy facilities. Comparisons were made between physicians who became owners of physical therapy services during the study period and physicians who never became owners of physical therapy services. Mixed Linear Models (MLM) were used to test for the effects of physician ownership and the combined influence of physician ownership and payer type on referrals for OPPT. Point estimates and 95% confidence intervals were calculated for the mean differences between Group 1 and Group 3 physicians for changes in OPPT referrals over time. The analyses were conducted first with samples of physicians who met a minimum criterion of ten referrals within the first year of data included in the data sets. Use of this criterion resulted in a small pool of physicians who qualified for inclusion in the testing. Data were compared between years 1 and 5 and then between years 1 and 2 versus 4 and 5. The criterion for physicians‘ inclusion was revised for post hoc analysis in an attempt to increase the sample size. All of the statistical tests were repeated in post hoc with the larger samples of physicians who met the minimum criterion of an average of ten referrals per year for years 1 and 2 rather than just the referral count for year 1. Results Overall, there was no statistically significant change in the total referrals as a result of a change in physicians‘ ownership status. Tests for the influence of payer type, physician group, and ownership status on referrals also revealed no significant differences between the two physician groups. Point estimates of the differences between Group 1 and Group 3 for changes in mean referrals supported the hypothesized relationships between physicians‘ ownership status and total referrals, referrals of commercially insured patients, and referrals of underinsured patients; however, the 95% confidence intervals for the point estimates were consistent with the non-significant MLM results. The hypothesized relationship between POPTS and referrals of Medicare patients was not supported in any of the analyses. In post hoc testing of the combined influence of payer type, physician group, and ownership status on referrals, a three-way interaction between physician group, payer type, and status was found (p=0.034, α<0.05). Including a larger sample size in the post hoc analyses led to outcomes that were different than those seen in the initial statistical tests. Conclusion This research outlines a novel approach to analyzing the influence of physician ownership and payer type on referral behaviors. The findings suggest that physicians‘ ownership of physical therapy services was not a predictor of their referrals to hospital-based OPPT services. Specifically, there was no significant effect of physician ownership of OPPT services on the total volume of referrals made to two hospital-systems‘ OPPT clinics. There also was no significant relationship between physician ownership, payer type, and referrals made to the hospital-based clinics. The theory predicting that POPTS physicians would work to eliminate market competition by reducing referral volumes and retaining patients with more lucrative reimbursement for their own practices was not supported. However, post hoc analysis with a larger sample size provided some evidence that a larger sample may have revealed the hypothesized relationships between physician ownership, payer type, and referrals for OPPT. Future research utilizing larger samples and data tracking physicians‘ OPPT referrals from their origins to their final dispositions are needed to clarify the relationships between physicians‘ ownership of OPPT services and the referrals they make for those services.


© The Author

Is Part Of

VCU University Archives

Is Part Of

VCU Theses and Dissertations

Date of Submission

July 2011


View graphic versionView graphic versionView graphic version