Degree Name

Masters of Science (Hons)


Department of Biomedical Science - Faculty of Science


Heat stress is a major consideration for a wide range of occupations including mining, construction and defence. Traditionally, heat stress equations have been used to identify scenarios likely to cause heat illness, based on the environmental conditions, and estimates of workload. However, as individuals respond differently to the same heat stress, it is often considered more appropriate to directly measure physiological responses (strain) so that behavioural changes may be used to reduce thermal strain. This project aimed to provide a theoretical basis for the development of a personal heat strain monitor, by evaluating a suitable surrogate index of core temperature for use in a working environment. This index was an insulated skin temperature (Tskin-insul), located laterally to the T2-T4 spine. Using linear regression modelling, we have found this Tskin-insul to closely approximate changes in oesophageal temperature (Tes). In this research, we applied these regression model to pre-existing data from our laboratory, resulting in 83% of the variance in Tes being explained on the basis of Tskin-insul, with a standard error of the estimate of 0.15°C on individual trials. With multiple-linear regression analyses using physiological and psychophysical measures as potential predictor variables, the standard error of the estimate fell to 0.10°C. Due to variation in the intercept and slope values, no single prediction equation was able to be derived for all situations, and this technique was considered impractical for use in a commercial monitor. By combining data collected at the same air temperature, four sets of equations were developed for each of three ambient temperature ranges. The predictor variables in these equations were: (i) Tskin-insul only (simple linear regression modelling); (ii) Tskin-insul and heart rate; (iii) Tskin-insul, heart rate, mass and sum of six skinfolds; and (iv) perceived exertion, Tskin-insul, mass and sum of six skinfolds. It was possible to predict Tes to an acceptable accuracy (standard error of estimate 0.2°C. The equations developed on data collected at 40°C had a similarly high error, with the most accurate equation, which utilised only Tskin-insul as a predictor variable, having a standard error of the estimate of 0.05°C above the maximum recommended level. The prediction equations utilising additional variables recorded standard error of the estimate values exceeding twice the recommended maximum level. These prediction equations were examined in relation to their efficacy for use with industrial heat strain guidelines. The equations developed using Tskin-insul, heart rate, mass and sum of six skinfolds were the most accurate equations at measuring a 0.8°C and 1.0°C change in Tes, at air temperatures of 30°C and 40°C. The prediction of an Tes measurement of 38.0°C and 38.5°C revealed that equations using Tskin-insul as the only predictor variable were most accurate. The equations for use at an air temperature of 36-40°C were considered sufficiently accurate for use in a personal monitoring system, with those developed for 30-36°C deemed to be unsuitable in the present form. The prediction equations were trialed in two heat strain indices: the Physiological Strain Index and the Cumulative Heat Strain Index. The predicted Tes measurement was inserted into these indices. It was found that the prediction equations using Tskininsul as the only predictor variable in the Physiological Strain Index, created a result that could be considered desirable in the air temperature range of 30-40°C. It was determined that, due to the level of inaccuracy of the equations, it was not possible to use surrogate measures of Tes in the Cumulative Heat Strain Index. Future research should primarily focus on reducing Tskin-insul measurement error, with the measurement then being trialed across a broad range of conditions to examine its efficacy for use in a personal monitoring system.

02Whole.pdf (7563 kB)