Abstract 201: Evaluating Time from Electrocardiogram (ECG) Acquisition to ST-Elevation Myocardial Infarction (STEMI) Activation for Inpatients: Deployment of a Preliminary Automated ECG Reading System

Abstract

Background: Inpatients with STEMI experience lengthier symptom-to-intervention times, decreased likelihood of invasive management, and an estimated 3-10-fold increased mortality risk compared to outpatients. Launched in October 2018, the Chest Pain - MI Registry™ now includes inpatient STEMIs, further motivating improvement of care for this vulnerable population. We investigated whether deployment of an automated ECG reading system across the inpatient enterprise at a large academic hospital is associated with improvement in time from ECG acquisition to STEMI activation. Methodology: From May 2018-December 2018, a pre-existing automatic ECG interpretation software on all inpatient ECG machines (Philips PageWriter TC70) was activated at our institution. The interpretation level on the software was changed from “none” to “measurements and interpretations.” This change affected all inpatient ECGs. A protocol for response to acute STEMI auto-interpretations was developed, which instructed all ECG technicians to hand ECGs ordered “stat” directly to a member of the patient’s clinical team within 5 minutes of acquisition. A run chart ( Figure ) was analyzed per the IHI “Four Rules for Interpreting Run Charts.” Baseline and intervention samples were assessed by Wilcoxon rank-sum test. Results: From May 2016-May 2018, median time from ECG acquisition to STEMI activation for inpatients at our institution was 23 minutes (N=14) [range 5, 90] with mean time of 35 minutes (+/- 30 per SD). During the subsequent seven-month study period, median time from ECG acquisition to STEMI activation was 21 minutes (N=5) [range 11, 64] with mean time of 34 minutes (+/- 26 per SD, p = 0.36). Of note, there was no reported increased workload or consultation to cardiology due to erroneous preliminary automated ECG reading. Conclusions: Following deployment of an automated inpatient ECG reading system for seven months, there was no significant decrease in the time from ECG acquisition to STEMI activation. Due to the relative infrequency of inpatient STEMIs and the resulting small number of cases during the study period, our results may be under-powered (power of 6.4%) to detect an improvement in STEMI activation times. This data suggests more aggressive and efficient strategies are required for timely diagnosis of acute inpatient STEMI.