Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Introduction and Background: The use of costly (though also potentially beneficial) medical imaging has risen dramatically in the last two decades. One important question that has been raised in this context is the role played by the growing rate of physician ownership of imaging equipment. Physicians who own imaging equipment, such as MRI machines, may obtain higher returns from recommending imaging tests for their patients than physicians who do not, and thus growth in the extent of physician ownership may contribute to growth in imaging use. This could have implications for the efficiency of the health care system and social welfare. We use Medicare data to track changes in the number of physicians who are performing MRI studies in their offices, and relate the prevalence of office-based MRI to the overall rates of MRI use observed in the Medicare population.

Data and Methods: We focus on 20% samples of physician claims from the Medicare program for 1998-2005. Using these claims, we identify physicians who bill for the performance of MRI exams in their offices. We track the overall number of physicians billing for MRI by Hospital Referral Region (HRR) by year, and also separately track physicians performing in-office MRI by specialty. We augment these measures with claims-based measures of the number of diagnostic imaging centers and the number of hospitals performing MRI, as well as survey-based data on the number of MRI scanners, by HRR and year. We use the claims to track the number of MRI procedures performed for Medicare beneficiaries, by HRR by year. We then use regression analyses to examine the relationship between the availability of MRI, particularly MRI in physician offices, and MRI use rates. Our regressions include fixed effects for HRRs and years, so that we identify relationships from changes within area over time, adjusting for generalized time trends. We also control for a range of other demographic and related characteristics of areas. Initial analyses focus on MRI overall; modeling in progress focuses on specific types of MRI as associated with specific types of physicians with in-office MRI.

Results: Results obtained thus far do not show strong relationships between growth in the number of physicians with in-office MRI and the rate at which MRI procedures are performed for Medicare patients. Growth in MRI availability overall is associated with strong increases in the use of MRI, but results for physician-owned MRI specifically do not show a clear relationship. Some initial results with orthopedist owned imaging and MRI for orthopedic studies do produce some evidence of a positive relationship. Work is ongoing to refine and extend these results.

Discussion and Conclusion: Contrary to expectations, results available thus far suggest that there is not a strong relationship between overall levels of in-office MRI and the rate of MRI use in the Medicare population.