Eli Segal1, Dave Ross2, Marie-Hélène Proulx3, Xiaoqing Xue4, Charlene Vacon5. 1. Urgences-santé, 6700 Jarry Street East, Montreal, Quebec H1P 0A4, Canada; Department of Prehospital Medicine, Sacré-Coeur Hospital, 5400 Boul Gouin O, Montréal, Quebec H4J 1C5, Canada; Emergency Department, Jewish General Hospital, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Quebec H3T 1E2, Canada; McGill University, 845 Sherbrooke St W, Montreal, Quebec H3A 0G4, Canada. Electronic address: eli.segal@mcgill.ca. 2. Department of Prehospital Medicine, Sacré-Coeur Hospital, 5400 Boul Gouin O, Montréal, Quebec H4J 1C5, Canada; Prehospital Medicine Regional Authority, CISSS de la Montérégie-Centre, 3120 Boulevard Taschereau, 7e étage, Greenfield Park, Quebec J4V 2H1, Canada; Advanced Care Paramedic Training Program, Université de Montréal, 2900 Edouard Montpetit Blvd, Montreal, Quebec H3T 1J4, Canada. 3. Intensive Care Unit, Centre hospitalier de l'Université de Montréal, 1000, rue Saint-Denis, Montréal, Québec H2X 0C1, Canada. 4. Emergency Department, Jewish General Hospital, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Quebec H3T 1E2, Canada. 5. Regional Paramedic Program for Eastern Ontario, Ottawa Hospital, 2475 Don Reid Dr, Ottawa, Ontario K1H 1E2, Canada.
Abstract
BACKGROUND: Prehospital ECGs (phECGs) are the main screening tool used by paramedics to identify ST elevation myocardial infarction (STEMI). In the absence of telemetry or personnel trained in ECG interpretation, paramedics must rely on computerized interpretation of phECGs, which suffer from an elevated false-positive (FP) rate, impairing reliable early activation of reperfusion centers by Emergency Medical Services. OBJECTIVE: Develop a clinical prediction rule to reduce the frequency of FPs for STEMI in prehospital patients. METHODS: This was a retrospective analysis of prehospital patients with a computer interpretation of '***ACUTE MI***' on phECG. We used logistic regression analysis to identify the independent variables for derivation of the rule. Once derived, we validated the rule on a distinct cohort of consecutive phECGs. RESULTS: Among the 654 cases in the derivation cohort, 46.2% were FP STEMIs. Four elements emerged as independent FP predictors: HR ≥ 120, no ongoing chest pain, no interpretable ST-segments in a lead, and presence of baseline wander or pacemaker spikes. In the derivation cohort this rule decreased FPs to 15.1% of the total cohort, while labelling 13.8% of STEMI cases as false-negatives (FNs). In the validation cohort (386 phECGs, 41.7% FPs), the rule decreased FPs down to 8.0%, while 25.9% were FN. CONCLUSION: Use of computer interpretation alone leads to a high STEMI FP rate. A clinical prediction rule based upon four elements available to paramedics can substantially lower the proportion of FPs. This clinical prediction rule should be incorporated into the decision for prehospital activation of the cardiac catheterization laboratory.
BACKGROUND: Prehospital ECGs (phECGs) are the main screening tool used by paramedics to identify ST elevation myocardial infarction (STEMI). In the absence of telemetry or personnel trained in ECG interpretation, paramedics must rely on computerized interpretation of phECGs, which suffer from an elevated false-positive (FP) rate, impairing reliable early activation of reperfusion centers by Emergency Medical Services. OBJECTIVE: Develop a clinical prediction rule to reduce the frequency of FPs for STEMI in prehospital patients. METHODS: This was a retrospective analysis of prehospital patients with a computer interpretation of '***ACUTE MI***' on phECG. We used logistic regression analysis to identify the independent variables for derivation of the rule. Once derived, we validated the rule on a distinct cohort of consecutive phECGs. RESULTS: Among the 654 cases in the derivation cohort, 46.2% were FP STEMIs. Four elements emerged as independent FP predictors: HR ≥ 120, no ongoing chest pain, no interpretable ST-segments in a lead, and presence of baseline wander or pacemaker spikes. In the derivation cohort this rule decreased FPs to 15.1% of the total cohort, while labelling 13.8% of STEMI cases as false-negatives (FNs). In the validation cohort (386 phECGs, 41.7% FPs), the rule decreased FPs down to 8.0%, while 25.9% were FN. CONCLUSION: Use of computer interpretation alone leads to a high STEMI FP rate. A clinical prediction rule based upon four elements available to paramedics can substantially lower the proportion of FPs. This clinical prediction rule should be incorporated into the decision for prehospital activation of the cardiac catheterization laboratory.