OBJECTIVE--To determine the potential impact of emergency intervention strategies to prevent deaths from coronary heart disease outside hospital. DESIGN--Analysis of routine medical and legal records of all persons dying of coronary heart disease in a defined population. SETTING--Glasgow City, north of the river Clyde, 1984. SUBJECTS--420 people under 65 years for whom the underlying cause of death on the death certificate was coronary heart disease (ICD 410-414, 9th Revision). RESULTS--Of the 296 deaths outside hospital, 73% occurred at home. The deaths of 40% of those who died outside hospital were not witnessed and these people could not have received prompt cardiopulmonary resuscitation. Only 16% of the witnesses of a death attempted cardiopulmonary resuscitation before the arrival of a doctor or an ambulance crew. Over half (53%) of the cases in which cardiopulmonary resuscitation could have been attempted by a witness, but was not attempted, death occurred in the presence of the spouse or other close relative. Death occurred in the presence of a duty doctor or the ambulance crew in a maximum of 5% of deaths outside hospital. Ninety one per cent of people were dead before a call for help was made. CONCLUSION--Unless a greater proportion of patients receive cardiopulmonary resuscitation before emergency staff arrive at the scene the provision of emergency care staff with defibrillators is unlikely to have a significant impact on deaths outside hospital caused by coronary heart disease.