Abstract
Gastrointestinal temperature (Tgi) is measured in occupational settings to monitor heat stress, but Tgi telemetric pill sensors are often ingested without sufficient time to exit the stomach, increasing the likelihood of errors and thereby necessitating data cleaning. It is unknown if data cleaning method affects the interpretation of heat stress. PURPOSE: The purpose of this study was to compare 2 recently published methods of cleaning Tgi data. METHODS: Tgi was measured once every 5 min during work from 6:00 am – 2:30 pm in 19 people (13 on Day 1 and 17 on Day 2). Data were cleaned using the methods from Yoder et al. (A) and from Petropoulos et al. (B). Method A involved visual inspection of each data set and removal of the first 2 h of data, data points that were below 36.5 °C, and consecutive time points that changed by more than 0.5 °C, with usable data beginning when temperature was stable (change 2.0 °C over 15 min. Min, max, and mean Tgi between the 2 cleaning methods were compared using paired samples t-tests. Number of points removed were compared with a Wilcoxon Signed-Rank Test. RESULTS: On Day 1 and 2, max Tgi did not differ [(Day 1: A and B = 37.7 ± 0.3 °C; p = 0.17) (Day 2: A = 37.8 ± 0.3 °C; B = 37.8 ± 0.2 °C; p = 0.12)]. Min Tgi was lower for B on Day 1 (36.6 ± 0.6 °C; A = 36.8 ± 0.2 °C; p = 0.04) and Day 2 (B = 36.8 ± 0.5 °C; A = 37.0 ± 0.2 °C; p = 0.006). Mean Tgi was lower for B on Day 1 (37.2 ± 0.3 °C; A = 37.3 ± 0.2 °C; p < 0.001) and Day 2 (B = 37.3 ± 0.2 °C; A = 37.4 ± 0.2 °C; p = 0.03). For Day 1, 6 had fewer points removed with Method B, 4 had fewer points removed with Method A, and 3 showed no difference in the number of points removed (p = 0.28). For Day 2, 8 had fewer points removed with Method B, 3 had fewer points removed with Method A, and 6 showed no difference (p = 0.03). CONCLUSION: Both cleaning methods produced similar max Tgi values, but different mean and min Tgi. Method A resulted in higher Tgi values, likely due to its stricter removal criteria, while method B preserved more data points. This suggests cleaning method can influence the interpretation of heat stress levels.
Gastrointestinal temperature (Tgi) is measured in occupational settings to monitor heat stress, but Tgi telemetric pill sensors are often ingested without sufficient time to exit the stomach, increasing the likelihood of errors and thereby necessitating data cleaning. It is unknown if data cleaning method affects the interpretation of heat stress. PURPOSE: The purpose of this study was to compare 2 recently published methods of cleaning Tgi data. METHODS: Tgi was measured once every 5 min during work from 6:00 am – 2:30 pm in 19 people (13 on Day 1 and 17 on Day 2). Data were cleaned using the methods from Yoder et al. (A) and from Petropoulos et al. (B). Method A involved visual inspection of each data set and removal of the first 2 h of data (24 points), data points that were below 36.5 °C, and consecutive time points that changed by more than 0.5 °C, with usable data beginning when temperature was stable (change 2.0 °C over 15 min. Min, max, and mean Tgi between the 2 cleaning methods were compared using paired samples t-tests. Number of points removed were compared with a Wilcoxon Signed-Rank Test. RESULTS: On Day 1 and 2, max Tgi did not differ [(Day 1: A and B = 37.7 ± 0.3 °C; p = 0.17) (Day 2: A = 37.8 ± 0.3 °C; B = 37.8 ± 0.2 °C; p = 0.12)]. Min Tgi was lower for B on Day 1 (36.6 ± 0.6 °C; A = 36.8 ± 0.2 °C; p = 0.04) and Day 2 (B = 36.8 ± 0.5 °C; A = 37.0 ± 0.2 °C; p = 0.006). Mean Tgi was lower for B on Day 1 (37.2 ± 0.3 °C; A = 37.3 ± 0.2 °C; p < 0.001) and Day 2 (B = 37.3 ± 0.2 °C; A = 37.4 ± 0.2 °C; p = 0.03). For Day 1, 6 had fewer points removed with Method B, 4 had fewer points removed with Method A, and 3 showed no differencewas no difference in the number of points removed (p = 0.28). For Day 2, 8 had fewer points removed with Method B, 3 had fewer points removed with Method A, and 6 showed no difference (p = 0.03). CONCLUSION: Both cleaning metods produced similar max Tgi values, but different mean and min Tgi. Method A resulted in higher Tgi values, likely due to its stricter removal criteria, while method B preserved more data points. This suggests cleaning method can influence the interpretation of heat stress levels.
Recommended Citation
Yoder, Hillary A.; Mulholland, Anne M.; Wingo, Jonathan E.; Krisher, Lyndsay K.; and Li, Yaqiang
(2025)
"Comparison of Data Cleaning Procedures for Gastrointestinal Temperature in an Occupational Setting,"
International Journal of Exercise Science: Conference Proceedings: Vol. 14:
Iss.
5, Article 32.
Available at:
https://digitalcommons.wku.edu/ijesab/vol14/iss5/32
Included in
Health and Physical Education Commons, Medical Education Commons, Sports Sciences Commons