This study prospectively determined the feasibility and accuracy of prehospital thrombolytic therapy candidate selection by base station emergency physicians. During a 6-month period, paramedics acquired and transmitted prehospital 12-lead electrocardiograms (ECGs) and then applied a thrombolytic therapy contraindication checklist. Emergency physicians interpreted prehospital ECGs and prospectively selected candidates for thrombolytic therapy. A safety committee of cardiologists reviewed prehospital ECGs, checklists and hospital records to determine accuracy independently. Six hundred-eighty stable adult prehospital patients with a chief complaint of nontraumatic chest pain were initially evaluated. Two hundred forty-one patients were excluded because of (1) unsuccessful electrocardiographic transmission (149), (2) transport to nonparticipating facilities (72), and (3) unavailable medical records (20). No prehospital thrombolytic therapy was administered in this study. Of 439 cases, 91 (21%) had the final diagnosis of acute myocardial infarction, 38 (8.7%) had diagnostic prehospital ECGs, and 12 (2.7%) were selected by emergency physicians as candidates for thrombolytic therapy. Seventy percent of patients with myocardial infarction had checklist exclusions for thrombolytic therapy. Prehospital evaluation increased mean scene time (paramedic arrival on scene to scene departure) by 4 minutes. The median time from chest pain onset to paramedic arrival in patients with myocardial infarction was 60 minutes. The estimated average time saved if prehospital thrombolytic therapy had been available was 101 +/- 81 minutes. The safety committee concluded that acceptable accuracy of emergency physician prehospital electrocardiographic interpretation, checklist and case selection was achieved. It is concluded that emergency physicians can accurately identify candidates for prehospital thrombolytic therapy.