Sustained Performance of a "Physicianless" System of Automated Prehospital STEMI Diagnosis and Catheterization Laboratory Activation.

BACKGROUND: Treatment times for primary percutaneous coronary intervention frequently exceed the recommended maximum delay. Automated "physicianless" systems of prehospital cardiac catheterization laboratory (CCL) activation show promise, but have been met with resistance over concerns regarding the potential for false positive and inappropriate activations (IAs).

METHODS: From 2010 to 2015, first responders performed electrocardiograms (ECGs) in the field for all patients with a complaint of chest pain or dyspnea. An automated machine diagnosis of "acute myocardial infarction" resulted in immediate CCL activation and direct transfer without transmission or human reinterpretation of the ECG prior to patient arrival. Any activation resulting from a nondiagnostic ECG (no ST-elevation) was deemed an IA, whereas activations resulting from ECG's compatible with ST-elevation myocardial infarction but without angiographic evidence of a coronary event were deemed false positive. In 2012, the referral algorithm was modified to exclude supraventricular tachycardia and left bundle branch block.

RESULTS: There were 155 activations in the early cohort (2010-2012; prior to algorithm modification) and 313 in the late cohort (2012-2015). Algorithm modification resulted in a 42% relative decrease in the rate of IAs (12% vs 7%; P < 0.01) without a significant effect on treatment delay.

CONCLUSIONS: A combination of prehospital automated ST-elevation myocardial infarction diagnosis and "physicianless" CCL activation is safe and effective in improving treatment delay and these results are sustainable over time. The performance of the referral algorithm in terms of IA and false positive is at least on par with systems that ensure real-time human oversight.