The early administration of thrombolytic agents significantly reduces mortality following a myocardial infarct and ideally they could be given by general practitioners when the patient is first seen. However, the diagnosis of myocardial infarction in the early stages can be very difficult especially if an electrocardiogram is not available. This may limit the use of thrombolytic drugs by general practitioners. We assessed the accuracy of diagnosis in general practice by asking general practitioners referring patients with chest pain, the likelihood that the event was due to a myocardial infarction and if they would use thrombolysis if it were available. Diagnostic accuracy and appropriate use of thrombolysis was analysed retrospectively, comparing the general practitioner with the admitting hospital doctor. One hundred consecutive patients were studied. The general practitioners accurately diagnosed myocardial infarctions in approximately 45% of cases and would have given thrombolysis inappropriately on 67% of occasions mainly because the final diagnosis in most of these patients was unstable angina rather than infarction. The hospital doctors administered streptokinase inappropriately to 33% of the patients and four had complications during treatment. Of those patients receiving thrombolysis, the average time delay from the general practitioner referring the patient to hospital to the patient being treated was 107 minutes. This study confirms that the diagnosis of myocardial infarction in the early stages is difficult and that thrombolytic therapy may be given inappropriately (mainly to patients with unstable angina). We conclude that until the accuracy of diagnosis of myocardial infarction can be improved in general practice it would seem inappropriate for thrombolysis to be given in the community at the moment.