Prehospital Electrocardiographic Computer Identification of ST-segment Elevation Myocardial Infarction

Mary Colleen Bhalla MDa*, Francis Mencl MD, MSa, Mikki Amber Gist MD, MHAa, Scott Wilber MD, MPHa & Jon Zalewskiaa From the Department of Emergency Medicine, Summa Health System (MCB, MAG), Akron, Ohio; and Summa Akron City Hospital (FM, SW, JZ), Akron, Ohio.The authors report no conflicts of interest.Address correspondence and reprint requests to: Mary C. Bhalla, MD, Summa Health System, Emergency Medicine, 525 East Market Street, Akron, OH 44304. E-mail: mcmcquown@yahoo.comAbstractBackground. Identifying ST-segment elevation myocardial infarctions (STEMIs) in the field can decrease door-to-balloon times. Paramedics may use a computer algorithm to help them interpret prehospital electrocariograms (ECGs). It is unknown how accurately the computer can identify STEMIs.Objectives. To Determine the sensitivity and specificity of prehospital ECGs in identifying patients with STEMI.Methods. Retrospective cross-sectional study of 200 prehospital ECGs acquired using Lifepak 12 monitors and transmitted by one of more than 20 emergency medical services (EMS) agencies to the emergency department (ED) of a Summa Akron City Hospital, a level 1 trauma center between January 1, 2007, and February 18, 2010. The ED sees more than 73,000 adult patients and treats 120 STEMIs annually. The laboratory performs 3,400 catheterizations annually. The first 100 patients with a diagnosis of STEMI and cardiac catheterization laboratory activation from the ED were analyzed. For comparison, a control group of 100 other ECGs from patients without a STEMI were randomly selected from our Medtronic database using a random-number generator. For patients with STEMI, an accurate computer interpretation was “acute MI suspected.” Other interpretations were counted as misses. Specificity and sensitivity were calculated with confidence intervals (CIs). The sample size was determined a priori for a 95% CI of ±10%.Results. Zero control patients were incorrectly labeled “acute MI suspected.” The specificity was 100% (100/100; 95% CI 0.96–1.0), whereas the sensitivity was 58% (58/100; 95% CI 0.48–0.67). This would have resulted in 42 missed cardiac catheterization laboratory activations, but zero inappropriate activations. The most common incorrect interpretation of STEMI ECGs by the computer was “data quality prohibits interpretation,” followed by “abnormal ECG unconfirmed.”Conclusions. Prehospital computer interpretation is not sensitive for STEMI identification and should not be used as a single method for prehospital activation of the cardiac catheterizing laboratory. Because of its high specificity, it may serve as an adjunct to interpretation.