Objectives and Rationale.
—For 10 common clinical presentations, we assessed differences in physicians' utilization of and charges for diagnostic imaging, depending on whether they performed imaging examinations in their offices (self-referral) or referred their patients to radiologists (radiologist-referral).
—Using previously developed methodologies, we generated episodes of medical care from an insurance claims database. Within each episode, we determined whether diagnostic imaging had been performed, and if so, whether by a self-referring physician or a radiologist. For each of the 10 clinical presentations, we compared the mean imaging frequency, mean imaging charges per episode of care, and mean imaging charges for diagnostic imaging attributable to self- and radiologist-referral.
—Depending on the clinical presentation, self-referral resulted in 1.7 to 7.7 times more frequent performance of imaging examinations than radiologistreferral (P<.01, all presentations). Within all physician specialties, self-referral uniformly led to significantly greater utilization of diagnostic imaging than radiologistreferral. Mean imaging charges per episode of medical care (calculated as the product of the frequency of utilization and mean imaging charges) were 1.6 to 6.2 times greater for self-referral than for radiologist-referral (P<.01, all presentations). When imaging examinations were performed—including those performed in both physicians' offices and hospital outpatient departments—mean imaging charges were significantly greater for radiologists than for self-referring physicians in seven of the clinical presentations (P<.01). This result is related to the high technical charges of hospital outpatient departments; in office practice, radiologists' mean charges for imaging examinations were significantly less than those of self-referring physicians for seven clinical presentations (P<.01).
—Nonradiologist physicians who operate diagnostic imaging equipment in their offices perform imaging examinations more frequently, resulting in higher imaging charges per episode of medical care. These results extend our previous research on this subject by their focus on a broader range of clinical presentations; a mostly elderly, retired population; and the inclusion of highertechnology imaging examinations.(JAMA. 1992;268:2050-2054)