Machine learning-based mortality prediction model for heat-related illness

Abstract

In this study, we aimed to develop and validate a machine learning-based mortality prediction model for hospitalized heat-related illness patients. After 2393 hospitalized patients were extracted from a multicentered heat-related illness registry in Japan, subjects were divided into the training set for development (n = 1516, data from 2014, 2017–2019) and the test set (n = 877, data from 2020) for validation. Twenty-four variables including characteristics of patients, vital signs, and laboratory test data at hospital arrival were trained as predictor features for machine learning. The outcome was death during hospital stay. In validation, the developed machine learning models (logistic regression, support vector machine, random forest, XGBoost) demonstrated favorable performance for outcome prediction with significantly increased values of the area under the precision-recall curve (AUPR) of 0.415 [95% confidence interval (CI) 0.336–0.494], 0.395 [CI 0.318–0.472], 0.426 [CI 0.346–0.506], and 0.528 [CI 0.442–0.614], respectively, compared to that of the conventional acute physiology and chronic health evaluation (APACHE)-II score of 0.287 [CI 0.222–0.351] as a reference standard. The area under the receiver operating characteristic curve (AUROC) values were also high over 0.92 in all models, although there were no statistical differences compared to APACHE-II. This is the first demonstration of the potential of machine learning-based mortality prediction models for heat-related illnesses.

A prehospital risk assessment tool predicts clinical outcomes in hospitalized patients with heat-related illness: a Japanese nationwide prospective observational study

“…Ninan et al reported that a higher J-ERATO score at hospital presentation was an independent predictor of underlying multiple organ dysfunction syndromes, defined as a SOFA score > 2 points ( P < 0.029) 13 . On the one hand, we previously reported the potential of machine learning-based mortality prediction models for heat-related illnesses (AUC, 0.92), using 24 variables at hospital arrival 22 . However, at present, such a model is not clinically feasible in most cases where heat-related illnesses are encountered, as machine learning algorithms require certain computer equipment for calculation.…”

Heat stroke prediction: a perspective from the internet of things and machine learning approach

“…First, Hirano et al [46] used some machine learning algorithm like logistic regression, support vector machine, random forest and XGBoost for mortality prediction for heat-related illness. "Heatstroke study" database in Japan was used in this system for machine learning prediction.…”

Machine learning analysis and future risk prediction of weather-sensitive cardiovascular disease mortality during summer in Tokyo, Japan

“…Moreover, the Japanese government reported that the number of patients hospitalised owing to cardiovascular diseases will reach twice the 2005 level in 2035, and cardiac and cerebrovascular diseases are estimated to increase 2.15-and 2.05-fold in 2055, respectively 18 . ML techniques can be used to target problems like summer heatstroke 19,20 , although no studies have applied these techniques to the mortality or morbidity risk owing to cardiovascular diseases related to summer weather.…”