Heat Exposure Among Adult Women in Rural Tamil Nadu, India

Abstract

Exposure to heat is associated with a substantial burden of disease and is an emerging issue in the context of climate change. Heat exposure is of particular concern in India, one of the world's hotter countries and soon to be its most populous, where a large fraction of the population works outdoors, lives in dwellings that are thermally inefficient, and is unable to access cooling technologies. Despite these concerns, relatively little is known about personal heat exposure in India, particularly in rural areas. Here we leverage temperature data collected as part of a randomized controlled trial of cookstove replacement to describe personal temperature exposures of older adult women in rural Tamil Nadu. We also compare personal exposure measurements to the nearest ambient monitoring stations, as well as to commonly used modelled temperature data products. We find that temperatures differ across individuals in the same location on the same day, sometimes by more than 5 degrees C within the same hour, and that some individuals experience sharp increases in heat exposure in the early morning or evening, potentially a result of cooking with solid fuels. In comparisons with the personal exposure measurements, we find stronger correlations with the modelled products (R2 of ~0.74) than with ambient monitors (R2 of ~0.6). We did not find evidence of systematic biases, which indicates that adjusting for discrepancies between personal and ambient exposure estimates is not straightforward. This study indicates a need for improved heat exposure assessment in epidemiological and burden of disease studies in India.

Modelling personal temperature exposure using household and outdoor temperature and questionnaire data: implications for epidemiological studies

“…14,16,[28][29][30][31] Of the limited studies on personal temperature measurements in LMICs, most were conducted in warm, rural regions of India (generally >18°C in the cool season) and involved relatively short-term (24-hour) assessment in limited sample sizes, with only two studies having repeated measurements across seasons. [18][19][20] These studies found a moderate-to-strong agreement between personal or household temperatures and outdoor temperature in general, but a somewhat weaker association in the cool season. [18][19][20] Compared to previous personal measurement studies, in the CKB-Air we captured 6-10 times greater amount of data (in person-hours) across a broader outdoor temperature range (mean daily temperature range: -3.9°C to 42.6°C) and across the personal, household, and outdoor environments.…”

Modelling personal temperature exposure using household and outdoor temperature and questionnaire data: implications for epidemiological studies

“…[14,16,[28][29][30][31] Of the limited studies on personal temperature measurements in LMICs, most were conducted in warm, rural regions of India (generally >18°C in the cool season) and involved relatively short-term (24-hour) assessment in limited sample sizes, with only two studies having repeated measurements across seasons. [18][19][20] These studies found a moderate-to-strong agreement between personal or household temperatures and outdoor temperature in general, but a somewhat weaker association in the cool season. [18][19][20] Compared to previous personal measurement studies, in the CKB-Air we captured 6-10 times greater amount of data (in person-hours) across a broader outdoor temperature range (mean daily temperature range: -3.9°C to 42.6°C) and across the personal, household, and outdoor environments.…”

“…[18][19][20] These studies found a moderate-to-strong agreement between personal or household temperatures and outdoor temperature in general, but a somewhat weaker association in the cool season. [18][19][20] Compared to previous personal measurement studies, in the CKB-Air we captured 6-10 times greater amount of data (in person-hours) across a broader outdoor temperature range (mean daily temperature range: -3.9°C to 42.6°C) and across the personal, household, and outdoor environments. In summer, we found a strong agreement between personal and household or outdoor temperatures.…”

“…[14,16] Only a few studies in LMICs reported distinctive personal temperature exposure patterns, but they were restricted to relatively small sample sizes (n≤50 except one) and/ or geographic regions. [17][18][19][20] Furthermore, while personal air pollution exposure prediction models using questionnaire and household measurements data have been developed, [21] to our knowledge, no studies using the same approach have been published for personal temperature exposure modelling. Importantly, to our knowledge, there is no systematic investigation on the possibly distinctive relationships of personal versus household or outdoor temperatures with health outcomes in the same study sample.…”

Invited Perspective: Beating the Heat

“…For example, cookstove intervention studies often monitor temperature exposures while also collecting biospecimens and cardiorespiratory data. 8 However, studies employing the technologies used by Meade et al. to evaluate physiological temperature (e.g., ingestible monitors, rectal thermocouple-based monitors, skin-based thermistors) are rare and may not be feasible in the field from the perspective of either acceptability or practicality.…”