Heat Strain Decision Aid (HSDA) acceptably predicts core body temperature during self-paced load carriage within multiple environmental conditions

Abstract

The accuracy of the Heat Strain Decision Aid (HSDA) was assessed for predicting core body temperature (Tc) associated with US Army Ranger Training Brigade (RTB) self-paced road marches during Spring, Summer, and Winter classes. Physiological data was collected from 65 Ranger students (Spring: n = 15, Summer: n = 20, Winter: n = 30) along with an assessment of clothing and equipment worn, and continuous measurements were taken of the environment. This observed data was used as inputs into HSDA and comparisons were made between observations and predictions. Five statistical assessments methods were used to assess the validity of HSDA to predict Tc; Bias, mean absolute error (MAE), root mean square deviation (RMSD), limits of agreement (LoA) and a non-parametric comparison similar to a Bland-Altman analysis. Calculated Bias, MAE, and RMSD between predicted and actual Tc showed a calculated Bias of -0.02, MAE of 0.40, and RMSD of 0.45 °C for the three classes combined. These analyses showed HSDA predictions were able to meet many of the accuracy criterions used to determine acceptability. Additionally, this work highlights areas for potential improvement of the HSDA modeling method.