Evaluating a Widely Implemented Proprietary Deterioration Index Model among Hospitalized Patients with COVID-19

• DOI: 10.1513/annalsats.202006-698oc
• Publication Date: 2020-12-24T21:53:57Z
• Authors: Karandeep Singh; Thomas S. Valley; Shengpu Tang; Benjamin Y. Li; Fahad Kamran; Michael W. Sjoding; Jenna Wiens; Erkin Otles; John P. Donnelly; Melissa Y. Wei; Jonathon P. McBride; Jie Cao; Carleen Penoza; John Z. Ayanian; Brahmajee K. Nall...
• Keywords: Interquartile range; Pandemic

Rationale: The Epic Deterioration Index (EDI) is a proprietary prediction model implemented in over 100 U.S. hospitals that was widely used to support medical decision-making during the coronavirus disease (COVID-19) pandemic. EDI has not been independently evaluated, and other models have shown be biased against vulnerable populations. Objectives: To evaluate hospitalized patients with COVID-19 overall disproportionately affected subgroups. Methods: We studied adult admitted units than intensive care unit at large academic center from March 9 through May 20, 2020. EDI, calculated 15-minute intervals, predict composite outcome of unit-level care, mechanical ventilation, or in-hospital death. In subset for least 48 hours, we also evaluated ability identify low risk experiencing this their remaining hospitalization. Results: Among 392 hospitalizations meeting inclusion criteria, 103 (26%) met outcome. median age cohort 64 (interquartile range, 53-75) 168 (43%) Black 169 women. area under receiver-operating characteristic curve 0.79 (95% confidence interval, 0.74-0.84). predictions did differ by race sex. When exploring clinically relevant thresholds found who exceeded an 68.8 made up 14% study had 74% probability hospitalization sensitivity 39% lead time 24 hours when threshold first exceeded. 286 experienced outcome, 14 (13%) never 37.9, negative predictive value 90% above 91%. Conclusions: identifies small subsets high-risk low-risk good discrimination, although its clinical use as early warning system limited sensitivity. These findings highlight importance independent evaluation before widespread operational among COVID-19.