Quantifying misclassification in maternal heat exposure: Implications for heat-health alert thresholds for pregnant women
1
Wits Planetary Health Research, University of Witwatersrand, Parktown, South Africa
Popul. Med. 2026;8(Supplement Supplement 1):A413
ABSTRACT
BACKGROUND:
Rising ambient temperatures are a global public health concern, particularly in settings with limited access to cooling. Heat exposure during pregnancy is associated with adverse maternal and foetal outcomes. Heat-health alert systems are intended to reduce risks but mostly rely on weather station data. This may underestimate risk and misclassify heat exposure, with potential implications for the timing of protective actions.
METHODS:
Maternal heat exposure was assessed among 63 women enrolled in the Bio-HEAT study in Johannesburg, South Africa. Personal-level exposure was measured using iButtons, while population level exposure was obtained from a local weather station. Daily mean (Tmean), maximum (Tmax) temperatures and heat index values were calculated for each method. We assessed agreement using mean difference (bias), Pearson correlation and Bland-Altman analysis. Misclassification was evaluated using National Weather Service Heat Index thresholds to examine whether personal measurement exceeded alert thresholds earlier than weather station data.
RESULTS:
A total of 4,191 participant-days were analysed. Mean and maximum iButton temperatures exceeded station values by >5°C (Tmean: 5.67°C [SD 4.76; range −4.03–27.77 ]); (Tmax: 5.32°C [SD 7.56; range −10.8–39.58]). Agreement was moderate for mean temperature (r = 0.43) and weak for maximum temperature (r = 0.21). Bland–Altman analyses showed proportional bias with increasing variability at higher temperatures. Categorical agreement using Heat Index thresholds was poor. Weather station data classified 93.9% of participant-days as “Caution” (≥26.7°C), whereas iButton data classified 62.4% of the same days as “Extreme Danger” (≥32.2°C). Using an alert threshold of ≥32.2°C, 47.2% were misclassified, largely due to personal measurements reaching alert thresholds earlier than station data.
CONCLUSIONS:
Personal and ambient temperature measurements differed substantially resulting in high misclassification of maternal heat exposure. These findings highlight limitations of ambient temperature-based heat-health alerts and the importance of considering vulnerable populations, including pregnant women in low resource settings, when setting thresholds.
Rising ambient temperatures are a global public health concern, particularly in settings with limited access to cooling. Heat exposure during pregnancy is associated with adverse maternal and foetal outcomes. Heat-health alert systems are intended to reduce risks but mostly rely on weather station data. This may underestimate risk and misclassify heat exposure, with potential implications for the timing of protective actions.
METHODS:
Maternal heat exposure was assessed among 63 women enrolled in the Bio-HEAT study in Johannesburg, South Africa. Personal-level exposure was measured using iButtons, while population level exposure was obtained from a local weather station. Daily mean (Tmean), maximum (Tmax) temperatures and heat index values were calculated for each method. We assessed agreement using mean difference (bias), Pearson correlation and Bland-Altman analysis. Misclassification was evaluated using National Weather Service Heat Index thresholds to examine whether personal measurement exceeded alert thresholds earlier than weather station data.
RESULTS:
A total of 4,191 participant-days were analysed. Mean and maximum iButton temperatures exceeded station values by >5°C (Tmean: 5.67°C [SD 4.76; range −4.03–27.77 ]); (Tmax: 5.32°C [SD 7.56; range −10.8–39.58]). Agreement was moderate for mean temperature (r = 0.43) and weak for maximum temperature (r = 0.21). Bland–Altman analyses showed proportional bias with increasing variability at higher temperatures. Categorical agreement using Heat Index thresholds was poor. Weather station data classified 93.9% of participant-days as “Caution” (≥26.7°C), whereas iButton data classified 62.4% of the same days as “Extreme Danger” (≥32.2°C). Using an alert threshold of ≥32.2°C, 47.2% were misclassified, largely due to personal measurements reaching alert thresholds earlier than station data.
CONCLUSIONS:
Personal and ambient temperature measurements differed substantially resulting in high misclassification of maternal heat exposure. These findings highlight limitations of ambient temperature-based heat-health alerts and the importance of considering vulnerable populations, including pregnant women in low resource settings, when setting thresholds.
| eISSN: | 2654-1459 |
We process personal data collected when visiting the website. The function of obtaining information about users and their behavior is carried out by voluntarily entered information in forms and saving cookies in end devices. Data, including cookies, are used to provide services, improve the user experience and to analyze the traffic in accordance with the Privacy policy. Data are also collected and processed by Google Analytics tool (more).
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
