From November 1982 through April 1989, 111 patients with refractory sustained ventricular tachycardia/fibrillation had the automatic cardioverter-defibrillator implanted at our institution, the first community hospital involved in implantation of such a device. We have reviewed our long-term clinical experience to assess the feasibility, learning curve, and efficacy of device implantation in a facility with cardiac electrophysiology expertise but without open-heart surgery facilities. All patients were considered inoperable or at high risk for other concomitant surgery. Eighty-six patients (77%) underwent uneventful implantation. Nine patients (8%) died prior to hospital discharge. Operative mortality declined from 10.9% to 5.4% during the first half (55 patients; November 1982 through September 1986) and second half (56 patients; October 1986 through April 1989) of the experience. Other postoperative complications occurred in 16 patients (14%), 12 of whom experienced complications during the first half of the experience. At 22 +/- 20 (mean +/- SD) months' follow-up, 78 (76%) of 102 patients discharged were alive, and 24 patients (24%) had died. Fifty patients (49%) had experienced at least one automatic cardioverter-defibrillator discharge associated with hypotensive symptoms. The actuarial incidence of sudden death at 1, 2, and 3 years was 1.2%, 5.5%, and 6.2%, respectively. We concluded that the automatic implantable cardioverter-defibrillator is an effective therapy for refractory ventricular tachycardia/fibrillation and that device implantation at community hospitals with an experienced cardiac electrophysiology team is both feasible and practical.