A Real-Time Early Warning System for Monitoring Inpatient Mortality Risk: Prospective Study Using Electronic Medical Record Data

• DOI: 10.2196/13719
• Publication Date: 2019-05-25T08:32:15Z
• Authors: Chengyin Ye; Oliver Wang; Modi Liu; Le Zheng; Minjie Xia; Shiying Hao; Bo Jin; Hua Jin; Chunqing Zhu; Chao Jung Huang; Peng Gao; Gray Ellrodt; Denny Brennan; Frank Stearns; Karl G Sylvester; Eric Widen; Doff B McElhinney; Xuefeng Ling
• Keywords: Triage; Early warning system; Medical record

The rapid deterioration observed in the condition of some hospitalized patients can be attributed to either disease progression or imperfect triage and level care assignment after their admission. An early warning system (EWS) identify at high risk subsequent intrahospital death an effective tool for ensuring patient safety quality reducing avoidable harm costs.The aim this study was prospectively validate a real-time EWS designed predict inpatient mortality during hospital episodes.Data were collected from system-wide electronic medical record (EMR) two acute Berkshire Health System hospitals, comprising 54,246 admissions January 1, 2015, September 30, 2017, which 2.30% (1248/54,246) resulted deaths. Multiple machine learning methods (linear nonlinear) explored compared. tree-based random forest method selected develop predictive application assessment. After constructing model, we validated algorithms as mortality.The algorithm scored patients' daily long-term probability admission stratified them into distinct groups. In prospective validation, attained c-statistic 0.884, where 99 encounters captured highest group, 69% (68/99) whom died episodes. It accurately predicted possibility top 13.3% (34/255) least 40.8 hours before death. Important clinical utilization features, together with coded diagnoses, vital signs, laboratory test results recognized impactful predictors final EWS.In study, demonstrated capability newly-designed monitor alert clinicians about in-hospital real time, thereby providing opportunities timely interventions. This is able assist decision making enable more actionable individualized better health outcomes target facilities.