Improving estimates of air pollution exposure through ubiquitous sensing technologies

Abstract

Traditional methods of exposure assessment in epidemiological studies often fail to integrate important information on activity patterns, which may lead to bias, loss of statistical power, or both in health effects estimates. Novel sensing technologies integrated with mobile phones offer potential to reduce exposure measurement error. We sought to demonstrate the usability and relevance of the CalFit smartphone technology to track person-level time, geographic location, and physical activity patterns for improved air pollution exposure assessment. We deployed CalFit-equipped smartphones in a free-living population of 36 subjects in Barcelona, Spain. Information obtained on physical activity and geographic location was linked to space-time air pollution mapping. We found that information from CalFit could substantially alter exposure estimates. For instance, on average travel activities accounted for 6% of people's time and 24% of their daily inhaled NO2. Due to the large number of mobile phone users, this technology potentially provides an unobtrusive means of enhancing epidemiologic exposure data at low cost.

Figure 1

Selected 31 days of CalFit data (circles represent GPS tracking of participants, color intensity represent energy intensity); background shading is NO₂ concentration.

Figure 2

Details of one day of activity for one participant.

Figure 3

(A) Time spent in activity spaces; (B) NO2 concentration in activity spaces, derived from fully adjusted model (temporal and microenvironmental factors); (C) Percent contribution from different activity spaces to total time-weighted average NO2 concentration, using concentrations from fully adjusted model (temporal and microenvironmental factors); (D) Percent contribution to total NO2 daily inhaled dose from different activities, using concentrations from fully adjusted model (temporal and microenvironmental factors).