Prediction of mortality risk and duration of hospitalization of COVID-19 patients with chronic comorbidities based on machine learning algorithms

This retrospective study investigated the prediction of mortality risk and length of stay (LoS) in COVID-19 patients with chronic comorbidities using machine learning (ML) algorithms. Data from 1291 patients (900 alive, 391 dead) at Afzalipour Hospital in Kerman, Iran (March 2020-January 2021) were analyzed. Gradient boosting showed the best performance (84.15% accuracy) for predicting mortality risk, while multilayer perceptron (MLP) with a rectified linear unit function (MSE = 38.96) was best for predicting LoS. Hyperlipidemia, diabetes, asthma, and cancer were significant predictors of mortality, while shortness of breath was most important for LoS prediction. The study concludes that ML algorithms can effectively predict mortality risk and LoS, aiding in resource allocation and clinical decision-making.

Parastoo Amiri, Mahdieh Montazeri, Fahimeh Ghasemian, Fatemeh Asadi, Saeed Niksaz, Farhad Sarafzadeh, Reza Khajouei