Assessment of Heat Exposure and Health Outcomes in Rural Populations of Western Kenya by Using Wearable Devices: Observational Case Study

Abstract

Background: Climate change increasingly impacts health, particularly of rural populations in sub-Saharan Africa due to their limited resources for adaptation. Understanding these impacts remains a challenge, as continuous monitoring of vital signs in such populations is limited. Wearable devices (wearables) present a viable approach to studying these impacts on human health in real time.

Objective: The aim of this study was to assess the feasibility and effectiveness of consumer-grade wearables in measuring the health impacts of weather exposure on physiological responses (including activity, heart rate, body shell temperature, and sleep) of rural populations in western Kenya and to identify the health impacts associated with the weather exposures.

Methods: We conducted an observational case study in western Kenya by utilizing wearables over a 3-week period to continuously monitor various health metrics such as step count, sleep patterns, heart rate, and body shell temperature. Additionally, a local weather station provided detailed data on environmental conditions such as rainfall and heat, with measurements taken every 15 minutes.

Results: Our cohort comprised 83 participants (42 women and 41 men), with an average age of 33 years. We observed a positive correlation between step count and maximum wet bulb globe temperature (estimate 0.06, SE 0.02; P=.008). Although there was a negative correlation between minimum nighttime temperatures and heat index with sleep duration, these were not statistically significant. No significant correlations were found in other applied models. A cautionary heat index level was recorded on 194 (95.1%) of 204 days. Heavy rainfall (>20 mm/day) occurred on 16 (7.8%) out of 204 days. Despite 10 (21%) out of 47 devices failing, data completeness was high for sleep and step count (mean 82.6%, SD 21.3% and mean 86.1%, SD 18.9%, respectively), but low for heart rate (mean 7%, SD 14%), with adult women showing significantly higher data completeness for heart rate than men (2-sided t test: P=.003; Mann-Whitney U test: P=.001). Body shell temperature data achieved 36.2% (SD 24.5%) completeness.

Conclusions: Our study provides a nuanced understanding of the health impacts of weather exposures in rural Kenya. Our study's application of wearables reveals a significant correlation between physical activity levels and high temperature stress, contrasting with other studies suggesting decreased activity in hotter conditions. This discrepancy invites further investigation into the unique socioenvironmental dynamics at play, particularly in sub-Saharan African contexts. Moreover, the nonsignificant trends observed in sleep disruption due to heat expose the need for localized climate change mitigation strategies, considering the vital role of sleep in health. These findings emphasize the need for context-specific research to inform policy and practice in regions susceptible to the adverse health effects of climate change.

Keywords: Africa; Kenya; climate; climate change; environment; environmental; exposure; fitness trackers; health; heat; outcome; outcomes; rural; sub-Saharan Africa; temperature; tracker; trackers; wearable; wearables; weather.

Figure 1

Schematic overview of the observational case study. Reference values are according to Parsons [12] and the National Weather Service [13]. Wearables depiction provided by Withings and e-TakesCare. (A) Methodology involves 83 participants from rural Siaya Health and Demographic Surveillance System, Kenya, equipped with wearables for a 3-week data collection period (total study duration: 9 weeks), coupled with local climate monitoring via a state-of-the-art weather station. (B) Collected data include metrics such as sleep duration, step count, pulse rate, and body temperature, as well as environmental data. (C) Results explore the reliability, completeness, and validity of data, with implications for climate change and health research. HDSS: Health and Demographic Surveillance System; HR: heart rate; WBGT: wet bulb globe temperature.

Figure 2

Participant stratification in this study.

Figure 3

Average daily wet bulb globe temperature in °C (dark blue line) and average daily heat index in °C (orange line) with daily ranges (daily wet bulb globe temperature range in blue ribbon; daily heat index range in orange ribbon). WBGT: wet bulb globe temperature.

Figure 4

(A) Mean daily step count of adult participants per day and season, wherein dry season is highlighted in orange and rainy season is highlighted in blue. (B) Mean daily sleep duration per adult participant per day and season. Trend lines are added using locally weighted scatterplot smoothing; seasons are classified according to Odhiambo et al [17].