Development and validation of early warning score systems for COVID‐19 patients

Abstract

COVID‐19 is a major, urgent, and ongoing threat to global health. Globally more than 24 million have been infected and the disease has claimed more than a million lives as of November 2020. Predicting which patients will need respiratory support is important to guiding individual patient treatment and also to ensuring sufficient resources are available. The ability of six common Early Warning Scores (EWS) to identify respiratory deterioration defined as the need for advanced respiratory support (high‐flow nasal oxygen, continuous positive airways pressure, non‐invasive ventilation, intubation) within a prediction window of 24 h is evaluated. It is shown that these scores perform sub‐optimally at this specific task. Therefore, an alternative EWS based on the Gradient Boosting Trees (GBT) algorithm is developed that is able to predict deterioration within the next 24 h with high AUROC 94% and an accuracy, sensitivity, and specificity of 70%, 96%, 70%, respectively. The GBT model outperformed the best EWS (LDTEWS:NEWS), increasing the AUROC by 14%. Our GBT model makes the prediction based on the current and baseline measures of routinely available vital signs and blood tests.

Threshold of increase in oxygen demand to predict mechanical ventilation use in novel coronavirus disease 2019: A retrospective cohort study incorporating restricted cubic spline regression

“…Second, the current study analyzed all data related to oxygen therapy at any given time point, which would have captured rapidly changing respiratory status of COVID-19 [ 26 ]. While preexisting scores, such as Respiratory Rate Oxygenation and National Early Warning Score, utilized clinical parameters only at defined time points, including on admission and/or a few days after admission [ 9 , 10 ], each patient in this study had detailed data with nearly 30 different time points. Third, several clinically valuable covariates were also obtained directly from the hospital information system and analyzed along with the changes in oxygen demand.…”

“…Since candidates for MV are usually on oxygen therapy and changes in respiratory status are frequently assessed within a day, estimation with such a long-term interval using a score is not practical. Moreover, although machine learning incorporating vital signs, laboratory data, and images could accurately calculate the risks for MV [ 7 , 10 ] it would be difficult for most health care facilities to adopt the complicated program without trained experts.…”

Development and validation of a dynamic 48-hour in-hospital mortality risk stratification for COVID-19 in a UK teaching hospital: a retrospective cohort study

“… 16 17 70 Others have used additional postadmission data in time-varying Cox models (for mortality and escalation), 18 19 which require knowledge of clinical information from the future and so should not be used for prediction. 21 Machine-learning models for mortality 20 and respiratory support 71 have also been reported to have good discrimination, although missing data are either discarded or imputed, which ignores the problem of informative missingness and may limit their applicability on an individual patient level. They also do not account for competing risks, for example, hospital discharge.…”

Development and Validation of a Dynamic 48-Hour In-Hospital Mortality Risk Stratification for COVID-19 in a UK Teaching Hospital: A Retrospective Cohort Study

“…16 17 70 Others have used additional postadmission data in time-varying Cox models (for mortality and escalation), 18 19 which require knowledge of clinical information from the future and so should Open access not be used for prediction. 21 Machine-learning models for mortality 20 and respiratory support 71 have also been reported to have good discrimination, although missing data are either discarded or imputed, which ignores the problem of informative missingness and may limit their applicability on an individual patient level. They also do not account for competing risks, for example, hospital discharge.…”