A-32 Free Communication/Poster – Military Physiology: MAY 30, 2007 7:30 AM - 12:30 PM ROOM: Hall E
The accuracy of sweat prediction models may be affected by the potential for body armor (BA) and spacer-vest (SV) technologies to alter heat strain.
PURPOSE: This study examined the effects of BA and SV technologies on the accuracy of an empirical sweat prediction model (HSDA2004C).
METHODS: Sweat losses were measured in 13 male Soldiers during 3 trials of intermittent treadmill walking (3.5 mph, 2-3% grade) over 4h in a hot, dry environment (35°C, 30% rh, 1m/s wind speed). Clothing configurations for the 3 trials were 1) battledress uniform (BDU), 2) BDU + BA, and 3) BDU + BA + SV. Model inputs included anthropometrics, exercise intensity, clothing and environmental conditions. Exercise intensity was equated in all trials (∼530 W) to eliminate the independent effects of added weight in BA and BA + SV trials. Actual sweat rates and differences in sweat rates (predicted-actual) were assessed by one-way repeated measures ANOVA with Tukey's HSD post-hoc test. The practical importance of the differences between actual and predicted sweating rates was examined by comparing the mean and 95% confidence intervals (CI) for group differences against an evidentiary threshold of ± 0.125 L/h.
RESULTS: The mean measured sweating rates were lower (P<0.05) in BDU (0.670 L/h) than both BA (0.773 L/h) and SV (0.797 L/h) trials. HSDA2004C over-predicted sweating rates for BDU (0.095 L/h, 95%CI = 0.028 - 0.161 L/h), BA (0.201 L/h, 95%CI = 0.092 - 0.310 L/h), and SV (0.131 L/h, 95%CI = 0.029 - 0.233 L/h). All predictions were unacceptable using classical statistical criteria (P<0.05), but none of the 95% CI's for group differences fell entirely outside the practical threshold of importance (± 0.125 L/h). Prediction error of the BA trial was statistically greater (P<0.05) than BDU and SV but differences among trials were small (0.036 - 0.106 L/h).
CONCLUSION: These results indicate that HSDA2004C significantly over-predicts sweating rates for all clothing configurations (P<0.05), but the magnitude of the error (< 0.125 L/h) is of little practical consequence. Differences among BDU, BA, and SV prediction errors were also small. These results suggest that the clothing properties of BA and SV do not substantially alter the ability of existing models to accurately predict Soldier sweat losses.