Doctor, Interrupted: Preemptive Workflow and Accuracy of Rapid Electrocardiogram Screening for ST-Elevation Myocardial Infarction by Emergency Medicine Providers

摘要

Background: Interruptions are thought to contribute tomedical errors. However, interruptions are also importantto patient care in the emergency department. Prior researchhas failed to reliably demonstrate a relationship betweeninterruptions and medical errors.Objectives: Explore associations of interrupted, preemptiveworkflow on accuracy of interpreting interrupting clinicalstimuli. We hypothesized accuracy would decrease duringpreemptive workflow compared to sequential workflow.Methods: A 2x2 factorial crossover simulation trial wascreated. Resident and attending physicians from a singleacademic emergency department were invited to participate.Participants first completed a preemptive module, viewingpatient presentations interrupted by clinical stimuli requiringinterpretation every minute. Participants then completed asequential module where presentations and clinical stimuliwere completed sequentially without interruption. Theprimary outcome was accuracy of interpreting clinical stimuli,specifically electrocardiograms (ECG’s) for ST elevationmyocardial infarction (STEMI), during preemptive andsequential modules. Generalized estimating equation logisticregression evaluated factors, defined a priori, that influencedodds of correct ECG interpretation.Results: 35 participants completed the study. Overall,there was no significant difference in accuracy of ECGinterpretation for STEMI in the preemptive compared with thesequential module (Mean 0.89, 0.91, Paired T test p=0.21).Attending physician status (OR 2.56, CI 1.66-3.94, p<0.01)and inferior STEMI (OR 0.08, CI 0.04-0.14, p<0.01) wereassociated with increased and decreased odds of correctinterpretation, respectively. Self reported confidence wasassociated with increased odds of correct interpretation inthe preemptive module, but not in the sequential module.(Interaction p=0.02)Conclusions: Preemptive interrupted workflow was notassociated with accuracy of ECG interpretation for STEMI.However odds of correct interpretation during preemptivesimulations were significantly decreased in ECG’s participantsreported low confidence in interpretation. Providers may beable to self identify “high risk” tasks prone to error in aninterrupted environment.