ASSURANCE AGAINST DRUG-INDUCED DEATH IN PATIENTS WITH HEART-FAILURE

Several recent studies of medications for heart failure have observed increased mortality in the treatment group. Historically, adequate data on mortality have not been provided to regulatory agencies, physicians, and patients. Simple confidence interval calculations suggest that from 5000 to more than 13,000 patients, depending on the mortality rate in the control group, would have to be studied in randomized controlled trials to exclude a risk of two to three drug-induced deaths per 100 patients treated. A sample size of less than 1000 patients would be needed to be confident that the incidence of drug-induced death is less than 10 per 100 treated if an agent truly had no effect on mortality. Guidelines for studies to provide estimates of the effect of a medication on mortality should be developed based in part on the relative values given to the benefits of an agent versus an increased risk of mortality.