Economic Motivation May Affect How Often Some Physicians Order Imaging Studies

"It's looking like a significant part of the increase in imaging utilization is due to self- and same specialty referral," says G. Scott Gazelle, MD, MPH, PhD, director of the Institute for Technology Assessment, who led the study being published in the November issue of Radiology. "We need to have some mechanism in place to control this sort of inappropriate utilization of imaging."

Potential conflicts of interest in imaging studies have been of concern for several decades, leading to the passage in the 1990s of laws restricting some forms of self-referral -- physicians' ordering tests that they will be reimbursed for interpreting. Some earlier studies found that self-referral was associated with higher utilization of imaging studies. Other studies that documented a greater number of imaging studies being performed by non-radiologists were not able to distinguish whether that increase was due to differences in referral patterns or to increased numbers of non-radiologists offering such procedures.

The current study was designed to retrospectively examine how frequently physicians ordered imaging studies for outpatients treated for specific medical conditions and if that varied depending on whether the studies were performed by radiologists or by members of the referring physician's own specialty, who could be colleagues in the same practice.

The researchers analyzed data from a nationwide, employer-based health plan with around 4 million members: employees at all levels, their dependents and retirees. They identified pairings of six common medical conditions -- cardiopulmonary disease, heart disease, arm or leg fracture, knee pain or injury, abdominal malignancy, or stroke -- with a diagnostic imaging procedure, and compiled information on each physician who had seen at least six patients in one or more of those condition/imaging pairings in the years 1999 to 2003.

The investigators analyzed every imaging referral and categorized the physicians according to their referral patterns. For purposes of this study, they compared only those physicians who referred all patients to themselves or members of their own specialty for imaging studies with those physicians who always referred to radiologists. The results showed that, depending on the condition/imaging pairings, those making self- or same-specialty referrals were from 12 percent to more than 200 percent more likely to order an imaging procedure than were physicians referring to radiologists.

"Some of those who self-refer will say that their patients are sicker, but we found that controlling for the effect of patients' age and additional health conditions made the likelihood of imaging among self- or same-specialty referrers even stronger -- more than 300 percent in some conditions. Others may claim to be offering greater convenience for patients, but when we looked specifically at CT and MR studies, less than 20 percent were carried out on the same day as the referring physician visit, which means the procedures should not be considered part of an office visit but rather separate visits specifically for the purpose of imaging," says Gazelle.

Gazelle notes that the range of variation in imaging utilization among same-specialty referrers was much greater than among radiologist referrers. "It's not that everyone is overutilizing imaging; but some appear to be, and we need to take a look at those with the highest rates of utilization. We also need to consider whether the self-referral laws, which currently exempt most procedures offered in physicians' own offices, need to be toughened," he says.

Gazelle is a professor of Radiology at Harvard Medical School and a professor in the Department of Health Policy and Management at Harvard School of Public Health. The study's co-authors are Elkan Halpern, PhD, Heather Ryan, MSc, and Angela Tramontano, MPH, all of the MGH Institute for Technology Assessment and Department of Radiology. The study was funded, in part, by the American College of Radiology.