We reviewed the charts of patients returning within 72 hours to our emergency department to determine whether monitoring revisits is a useful quality assurance indicator. Patient visits for June and December 1987 were selected to eliminate a potential seasonal difference. Of the 13,261 visits during these two months, 455 (3.4%) were revisits within 72 hours. Charts were available on 444 patients, of whom 407 (91.7%) represented cases in which the return and the initial visits were clearly related. Charts were reviewed for deficiencies in medical management, appropriate prescribed follow-up, patient education, and patient compliance. Suspected medical management problems were discussed by the three senior authors, and a consensus decision was made. Return visits were considered avoidable if a deficiency was noted in at least one of the areas listed above. There were 297 unscheduled related return visits, 96 (32.3%) of which were avoidable. Of these avoidable visits, 38 (39.6%) had medical management deficiencies, 14 (14.6%) had inappropriate prescribed follow-up, 20 (20.8%) had not been given proper education, and 35 (36.5%) were due to patient noncompliance. Of the 110 scheduled return visits, there was one (0.9%) deficiency in medical management and none in the other categories. Of the unscheduled return visits, 146 (49.2%) returned within 24 hours; 89 (30.0%) between 24 and 48 hours; and 62 (20.8%) between 48 and 72 hours. Of the avoidable visits, 85% returned within 48 hours, as did 92% of those with medical management deficiencies.(ABSTRACT TRUNCATED AT 250 WORDS)