EMS/ORIGINAL CONTRIBUTION emergency medical services, evaluation emergency medical service systems prehospital care Prospective Validation of a New Model for Evaluating Emergency Medical Services Systems by In-Field Observation of Specific Time Intervals in Prehospital Care From the Arizona Emergency Medicine Research Center, College of Medicine, University of Arizona, Tucson. Receivedfor publication June 29, I992. Revision received September 22, I992. Accepted for publication September 28, 1992. This project was supported by a grantfrom the Arizona Department of Health Services Office of Emergency Medical Services. Presented at the 5ocietyfor Academic Emergency Medicine Annual Meeting in Toronto, Ontario, Canada, May I992. Daniel W Spaite, MD, FACEP Terence D Valenzuela, MD, FACEP Harvey W Meislin, MD, FACEP Elizabeth A Criss, RN Paul Hinsberg Study objective: To develop and validate a new time interval model for evaluating operational and patient care issues in emergency medical service (EMS) systems. Design/setting/type of participant: Prospective analysis of 300 EMS responses among 20 advanced life support agencies throughout an entire state by direct, in-field observation. Results: Mean times (minutes) were response, 6.8; patient access, 1.O; initial assessment, 3.3; scene treatment, 4.4; patient removal, 5.5; transport, 11.7; delivery, 3.5; and recovery, 22.9. The largest component of the on-scene interval was patient removal. Scene treatment accounted for only 31.0% of the on-scene interval, whereas accessing and removing patients took nearly half of the on-scene interval (45.8%). Operational problems (eg, communi-cations, equipment, uncooperative patient) increased patient removal (6.4 versus 4.5; P= .004), recovery (25.4 versus 20.2; P= .03), and out-of-service (43.0 versus 30.1; P= .007) intervals. Rural agencies had longer response (9.9 versus 6,4; P= .014), transport (21.9 versus 10.3; P< .0005), and recovery (29.8 versus 22.1; P= .049)intervals than nonrural. The total on- scene interval was longer if an IV line was attempted at the scene (17.2 versus 12.2; P< .0001). This reflected an increase in scene treatment (9.2 versus 2.8; P< .0001), while patient access and patient removal remained unchanged. However, the time spent attempting IV lines at the scene accounted for only a small part of scene treatment (1.3 minutes; 14.1%) and an even smaller portion of the overall on-scene interval (7.6%). Most of the increase in scene treatment was accounted for by other activities than the IV line attempts. Conclusion: A new model reported and studied prospectively is useful as an evaluative research tool for EMS systems and is broadly applicable to many settings in a demographically diverse state. This model can provide accurate information to system researchers, medical directors, and administrators for altering and improving EMS systems. APRIL 1993 22:4 ANNALS OF EMERGENCY MEDICINE 6 3 8 / 9