Back

TracMyAir: Smartphone-enabled spatiotemporal estimates for inhaled doses of particulate matter and ozone to personalize health outcomes

Lahens, N. F.; Isakov, V.; Chivily, C.; El Jamal, N.; Mrcela, A.; FitzGerald, G. A.; Skarke, C.

2026-02-16 public and global health
10.64898/2026.02.13.26346275 medRxiv
Show abstract

Accurate quantification of individual exposure to air pollutants remains a major challenge in environmental health, as fixed-site monitoring fails to account for mobility, indoor environments, and physiological variability. We deployed TracMyAir, a smartphone-based digital health platform designed to generate time-resolved, personalized exposure and inhaled dose estimates for PM2.5 and ozone under real-world conditions. In an exploratory study of 18 adults contributing more than 1,500 participant-hours, the platform integrated smartphone geolocation, regulatory (AirNow) and community-based (PurpleAir) air quality data, building infiltration modeling, microenvironment classification, and wearable-derived physical activity metrics to compute eight tiers of hourly exposure estimates, culminating in individualized inhaled dose. Hourly dose estimates derived from smartphone-and smartwatch-based step counts were concordant (Spearman correlation p=0.97-0.98), while heart rate-based estimates yielded greater variability and higher mean values (p=0.82-0.92). Exposure explained 51-73% of variance in inhaled dose of PM2.5 and 68-84% of ozone, suggesting that physiological-based modeling approaches improve hyperlocal estimates of personal pollutant burden. Substantial inter-and intra-individual variability reflect dynamic microenvironmental transitions and activity patterns. Modeled doses based on regulatory and community sensor networks were strongly correlated (R=0.84), with community sensors located closer to participants on average, supporting the feasibility of integrating dense, low-cost monitoring networks. No consistent association was observed between outdoor pollutant levels and neighborhood socioeconomic status in this cohort. These findings demonstrate the feasibility of a scalable, smartphone-centered digital health approach for hyperlocal exposure and inhaled dose modeling. By leveraging ubiquitous consumer devices and existing air quality networks, TracMyAir enables personalized environmental exposure assessment with potential applications in epidemiology, population health, and precision environmental medicine.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.