Budget constrained machine learning for early prediction of adverse outcomes for COVID-19 patients.

The combination of machine learning (ML) and electronic health records (EHR) data may be able to improve outcomes of hospitalized COVID-19 patients through improved risk stratification and patient outcome prediction. However, in resource constrained environments the clinical utility of such data-driven predictive tools may be limited by the cost or unavailability of certain laboratory tests. We leveraged EHR data to develop an ML-based tool for predicting adverse outcomes that optimizes clinical utility under a given cost structure. We further gained insights into the decision-making process of the ML models through an explainable AI tool. This cohort study was performed using deidentified EHR data from COVID-19 patients from ProMedica Health System in northwest Ohio and southeastern Michigan. We tested the performance of various ML approaches for predicting either increasing ventilatory support or mortality. We performed post hoc analysis to obtain optimal feature sets under various budget constraints. We demonstrate that it is possible to achieve a significant reduction in cost at the expense of a small reduction in predictive performance. For example, when predicting ventilation, it is possible to achieve a 43% reduction in cost with only a 3% reduction in performance. Similarly, when predicting mortality, it is possible to achieve a 50% reduction in cost with only a 1% reduction in performance. This study presents a quick, accurate, and cost-effective method to evaluate risk of deterioration for patients with SARS-CoV-2 infection at the time of clinical evaluation.

Critically Ill Adults With Coronavirus Disease 2019 in New Orleans and Care With an Evidence-Based Protocol.

BACKGROUND: Characteristics of critically ill adults with coronavirus disease 2019 (COVID-19) in an academic safety net hospital and the effect of evidence-based practices in these patients are unknown. RESEARCH QUESTION: What are the outcomes of critically ill adults with COVID-19 admitted to a network of hospitals in New Orleans, Louisiana, and what is an evidence-based protocol for care associated with improved outcomes? STUDY DESIGN AND METHODS: In this multi-center, retrospective, observational cohort study of ICUs in four hospitals in New Orleans, Louisiana, we collected data on adults admitted to an ICU and tested for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) between March 9, 2020 and April 14, 2020. The exposure of interest was admission to an ICU that implemented an evidence-based protocol for COVID-19 care. The primary outcome was ventilator-free days. RESULTS: The initial 147 patients admitted to any ICU and tested positive for SARS-CoV-2 constituted the cohort for this study. In the entire network, exposure to an evidence-based protocol was associated with more ventilator-free days (25 days; 0-28) compared with non-protocolized ICUs (0 days; 0-23, P = .005), including in adjusted analyses (P = .02). Twenty patients (37%) admitted to protocolized ICUs died compared with 51 (56%; P = .02) in non-protocolized ICUs. Among 82 patients admitted to the academic safety net hospital's ICUs, the median number of ventilator-free days was 22 (interquartile range, 0-27) and mortality rate was 39%. INTERPRETATION: Care of critically ill COVID-19 patients with an evidence-based protocol is associated with increased time alive and free of invasive mechanical ventilation. In-hospital survival occurred in most critically ill adults with COVID-19 admitted to an academic safety net hospital's ICUs despite a high rate of comorbidities.