Physicians who do not refer their patients to radiologists for medical imaging use imaging examinations more frequently than do physicians who refer their patients to radiologists, and the charges are usually higher when the imaging is done by the self-referring physician. From our results it is not possible to determine which group of physicians uses imaging more appropriately.

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 higher-technology imaging examinations.

To achieve this benefit, it would be necessary to treat 100% of patients, assuming they were all symptomatic to begin with.To correct the analysis, the authors could reduce the per¬ centage who benefit from symptom relief from 28% to 7% of all patients or add the costs of treating 100% of patients to benefit 28%. They also should add the costs of réévaluation to stop medication in nonresponders or acknowledge that 21% of all patients would be treated indefinitely without benefit. Whichever approach is taken, the actual cost-effec¬ tiveness of screening will be less than the results published in the article.