Optimal Triage for COVID-19 Patients Under Limited Intensive Care Unit Capacity: Development of a Parsimonious Machine Learning Prediction Model and Threshold Optimization Using Discrete-Event Simulation

Abstract

Background: The coronavirus disease 2019 (COVID-19) pandemic has caused an unprecedented burden on healthcare systems. To effectively triage COVID-19 patients within situations of limited data availability and to explore optimal thresholds to minimize mortality rates while maintaining the healthcare system capacity.Methods: A nationwide sample of 5601 patients confirmed for COVID-19 up until April 2020 was retrospectively reviewed. XGBoost and logistic regression analysis were used to develop prediction models for the patients’ maximum clinical severity during hospitalization, classified according to the WHO Ordinal Scale for Clinical Improvement (OSCI). The recursive feature elimination technique was used to evaluate the extent of the model performance’s maintenance when clinical and laboratory variables are eliminated. Using populations based on hypothetical patient influx scenarios, discrete-event simulation was performed to find the optimal threshold within limited resource environments that minimizes mortality rates.Results: The cross-validated area under the receiver operating characteristics (AUROC) of the baseline XGBoost model that utilized all 37 variables was 0.965 for OSCI ≥ 6. Compared to the baseline model’s performance, the AUROC of the reduced model that utilized 17 variables was maintained at 0.963 with statistical insignificance. Optimal thresholds were found to minimize mortality rates in a hypothetical patient influx scenario. The benefit of utilizing an optimal triage threshold was clear, reducing mortality up to 18.1% compared to the conventional Youden Index.Conclusions: Our adaptive triage model and its threshold optimization capability reveal that COVID-19 management can be integrated using both medical and healthcare management sectors to guarantee maximum treatment efficacy.