Background: Digitalis products are among the agents most frequently prescribed to the elderly, yet previous studies have not provided age-, race- , and sex-specific rates of utilization of digitalis by this population. Estimates of the rate of hospitalization with an adverse reaction from digitalis therapy have varied considerably between systems relying on passive reports and those using active surveillance. Methods: Medicare data from 1985 through 1991 and data from the 1987 National Medical Expenditure Survey were used to determine population-based estimates of the use of digitalis in elderly beneficiaries by age group, sex, and race. Hospitalization rates with an adverse event caused by digitalis therapy were calculated for those persons estimated to be using digitalis. Medicare data were used to identify the frequency of selected comorbidities among persons with an adverse event caused by digitalis therapy as well as the frequency of clinical manifestations associated with digitalis intoxication. Results: Over 3 million Medicare beneficiaries were estimated to be using digitalis in 1987. A total of 202 011 hospitalizations with a coded adverse event caused by digitalis therapy were reported during the 7-year study period. Of persons estimated to be using digitalis, 8.53 per 1000 were hospitalized annually with an adverse event caused by digitalis therapy. Women, individuals with increasing age, and persons of black race, especially those with impaired renal function, were significantly (P<.05) more likely to experience hospitalization with an adverse event caused by digitalis therapy. Conclusion: This information may help identify categories of elderly patients who require more frequent monitoring to prevent adverse effects of digitalis therapy. Changes in the format of the hospital bill to include more diagnoses along with increased mandatory reporting of adverse drug events will improve the sensitivity of Medicare data for surveillance of adverse drug events.