Validation of Core Temperature Estimation Algorithm
MASSACHUSETTS INST OF TECH LEXINGTON LEXINGTON United States
Pagination or Media Count:
Real-time knowledge of core body temperature is a key input to determine thermal work-strain and risk of heat injury. An algorithm for estimating core temperature based on heart rate has been developed by others in order to avoid standard but more invasive measurements, such as ingestible capsules and rectal or esophageal probes. This report provides an independent assessment of the algorithm, based on both parametric variations and field data that were not used in algorithm development. Assessment through parametric variation shows qualitatively expected behavior. Field data were taken from a study of 33 young, male, military personnel who tested six prototype uniforms over the course of 12 days in Okinawa, Japan. The field data selected for assessment consist of nearly 48,000 measurements of heart rate and core temperature. Core temperature was measured by an ingestible capsule. The observed core temperature range was 36.139.5C. Bland-Altman analysis yielded a bias of 0.01C and a 95 Limit of Agreement LoA of 0.58C. The LoA is similar to the 0.5C LoA resulting from variations in ingestible capsule temperatures, and is also consistent with the 0.63C LoA found by the algorithm developer on other field data. The accuracy of the Physiological Strain Index PSI, a standard estimate of thermal-work strain on a nominal 010 scale, was also assessed with the estimated core temperature as an input. A bias of 0.01 and LoA of 1.2 was found. The PSI LoA is comparable to the LoA resulting from the variability in computing PSI with two different ingestible capsule temperatures. Overall, within the applicability of the field data that were evaluated, the core temperature estimation algorithm is sufficiently accurate and precise to provide a field-expedient indication of core temperature and thermal-work strain.
- Numerical Mathematics
- Anatomy and Physiology