In this study 377 patients with hypertension from four general practices were referred to a nurse for care. A management protocol was agreed by all participating doctors and programmed into a microcomputer. Three practices representing approximately 80% of the total patients had a computer-assisted care programme and there is no evidence that this improved their care. Three-quarters of the patients had reached their target diastolic blood pressure by the end of the study compared with 50% at the beginning. The proportion of patients with normal weights for their heights increased, and almost one-fifth of those patients who admitted to smoking at the onset subsequently stated that they had stopped. Patients and doctors were satisfied with the system.