The automatic implantable defibrillator has been shown to decrease the mortality of patients who have survived cardiac arrest due to ventricular tachycardia or fibrillation and are at high risk for recurrence. We performed a cost-effectiveness analysis of this seemingly expensive new technology with data obtained from the 1984 Medicare data base, the medical literature, Medicare carriers, individual pharmacies and hospitals, and expert opinion. Analyzing combinations of principal and secondary discharge diagnoses across 18 diagnosis-related groups, we estimated the cost of hospitalization for a comparison group of patients. Hospitalization costs for the defibrillator group were obtained from reported empirical data. Rehospitalization rates and other health-care use estimates were solicited from an expert panel of physicians, and mortality rates for both groups were obtained from the literature. Using a decision-analytic model, we estimated that the net cost effectiveness of the defibrillator, when used in the high-risk patient, is approximately $17,100 per life-year saved, with sensitivity analyses suggesting that the true value lies between $15,000 and $25,000. This estimate is well within the range that is currently accepted by the US medical care system for other life-saving interventions. We also estimated the cost effectiveness of the defibrillator in a 1991 scenario to be $7,400 per life-year saved, when the device would have greater longevity, would be programmable, and would not require a thoracotomy. Sensitivity analyses suggest that the true value lies between a value that is cost saving (less expensive than pharmacologic therapy) and $19,600 per life-year saved.