Feasibility of Integrating Wearable Devices and Ecological Momentary Assessment for Real-Time Environmental Exposure Estimation: Proof-of-Concept Study

Clinical Snapshot

Key Findings

• P Value: Not reported for feasibility outcomes

• Effect Size: Mean 565 GPS coordinates per participant, 1.4 EMA surveys per day, 64% complete sleep data days

• Primary Outcome: 71% of participants (5/7) met predefined feasibility targets

• Nnt Or Sensitivity: Feasibility threshold: ≥3 GPS measurements/day, ≥1 survey/day, ≥50% complete sleep data days

• Confidence Interval: Not reported

Clinical Application

Technically feasible but requires significant participant engagement and technology literacy; compliance barriers identified Relevant for Australian environmental health research, particularly in urban centres; could complement existing air quality monitoring networks; no specific regulatory considerations for research use Young adults in urban environments with access to wearable technology; limited generalisability to broader populations

Abstract

BACKGROUND: Environmental exposures such as heat and air pollution are critical determinants of health, yet traditional assessment methods rely on stationary monitors or residential address proxies that fail to capture the exposures that individuals experience throughout the day. OBJECTIVE: This pilot study aimed to assess the feasibility of integrating ecological momentary assessment (EMA), wearable devices, and continuous GPS tracking to capture real-time environmental exposures and to explore associations with concurrent health outcomes. METHODS: In total, 7 young adults (aged approximately 16 to 24 years; 5/7, 71% female) wore Fitbit Charge 6 watches from July 2025 to August 2025 (mean 28.1, SD 1.1 days), recording sleep quality and duration, resting heart rate, breathing rate, heart rate variability, and physical activity. GPS location measured at up to 5-minute intervals (mean 19.7, SD 25.8 measurements per day) was linked to ambient temperature, humidity, and air pollution data (particulate matter <2.5 um or <10 um in diameter, nitrogen dioxide, sulfur dioxide, ozone, and carbon monoxide) derived from monitoring stations, satellites, and climate models using data-integration algorithms accessed via an application programming interface. EMA surveys administered 3 times per day captured participants' emotional states and location (inside or outside). Feasibility targets were ≥3 GPS measurements per day, ≥1 survey completed per day, and complete sleep data on ≥50% of days. We examined exploratory bivariate correlations between environmental exposures, physiological measures, and self-reported mood, adjusting for multiple comparisons using false discovery rate correction. RESULTS: Of the 7 participants, 5 (71%) met predefined feasibility targets. Mean compliance included 565 (SD 457) GPS coordinates per participant, 1.4 (SD 0.2) EMA surveys per day, and complete Fitbit sleep data on 64% (SD 27%) of days. Surveys identified barriers to compliance, including perceived complexity of the system and forgetting to put the Fitbit watch back on after removing it. Exploratory correlations (6/7, 86% of participants with complete Fitbit data) revealed associations between nitrogen dioxide and heat exposure and reduced heart rate variability (a marker of parasympathetic tone), and between air pollutants (sulfur dioxide) and increased negative emotions. Heat exposure showed a paradoxical pattern of lower self-reported sadness but reduced heart rate variability with higher levels of heat exposure. Given the small sample size, these correlations should be considered preliminary and hypothesis generating rather than definitive findings. CONCLUSIONS: This study demonstrates that the multimodal integration of wearable devices, GPS tracking, and EMA is feasible for capturing real-time environmental exposures and concurrent health outcomes in young adults. This approach addresses critical exposure misclassification issues in environmental health research that relies on residential addresses as proxies. Preliminary patterns suggest complex relationships between environmental exposures and both physiological and emotional outcomes, warranting further investigation in larger, more diverse samples. This approach could inform future personalized environmental health interventions.

References

1. 1.Ramjan, S., Blum, M., Tseng, R. Y., Davey, K., Shereen, A. D., & Nomura, Y. (2026). Feasibility of integrating wearable devices and ecological momentary assessment for real-time environmental exposure estimation: Proof-of-concept study. JMIR Formative Research. https://doi.org/10.2196/86615