Study Suggests MRI Financial Bias

To MRI or not to MRI—that is the question. In a study conducted at Duke University Medical Center, Matthew P. Lungren, M.D., and his colleagues found that images of patients who had been referred by a physician group with a financial interest in the MR imaging equipment were significantly more likely to be negative than were those among patients referred by a physician group with no financial interest in the equipment.

Crystal Phend, senior staff writer for MedPage Today wrote that 33% of MRI scans turned up negative in patients referred by physicians with an ownership interest in the imaging equipment as compared with 25% of the scans that were positive. In their online article in Radiology, the investigators reported that the characteristics of both the patients and physicians appeared similar between those who tested positive and negative—except for the self-referral difference.

Phend reported the researchers conclusion as: “Our findings suggest that there may be a lower threshold for referral to MR imaging of the knee in the financially incentivized ordering group.” The study group suggested that the possibility of intentional overuse related to a profit motive should be considered by third-party payers.

They wrote in their study, “Despite the limitations, we believe that the inherent conflict of interest present when physicians both order and then perform and collect fees on the basis of the acquisition of MR imaging examinations is an important factor to consider whenever the matter of health care cost is raised.”

According to Phend, Lungren’s group studied two similar referring groups, one with a financial interest in the equipment used and one without. The retrospective study included all 700 consecutive diagnostic MRI scans of the knee interpreted by a single radiology practice from January to April 2009 referred by two orthopedics groups in the same community.

Among the scans, 117 of 350 referred by financially-interested orthopedic physicians versus 88 of 350 referred by others turned out negative. Among positive scans, the mean number of these abnormalities per scan was similar between the financially vested and independent groups supporting the similarity of the patient cohorts. Bilateral scans were more than twice as commonly ordered by orthopedists with an ownership interest in the MRI machines.

Phend, in her report, noted that patient age, physician years of practice, and proportion with specialized or fellowship training did not differ between the groups. The researchers could not entirely rule out selection or referral bias as they lacked data on insurance, socioeconomic status and ethnicity.