We studied the relationship of the appropriateness of the use of coronary angiography, carotid endarterectomy, and upper gastrointestinal tract endoscopy to their rates of use in 23 adjacent counties in one state. We measured appropriateness by means of a detailed review of the medical records of Medicare beneficiaries who had the procedures performed in 1981, using preset criteria derived by an expert panel. Use rates per 10 000 Medicare enrollees in a county varied from 13 to 158 for coronary angiography, 5 to 41 for carotid endarterectomy, and 42 to 164 for upper gastrointestinal tract endoscopy. Inappropriate use varied by county from 8% to 75% for coronary angiography, from 0% to 67% for carotid endarterectomy, and from 0% to 25% for endoscopy. For coronary angiography, inappropriate use accounted for 28% of the variance in the county rate. For the other two procedures, no significant correlations were found between inappropriateness of use and rate of use. We conclude that little of the variation in the rates of use of these procedures can be explained by inappropriate use.