University of California Berkeley
Validating the CalFit Smartphone Sensor in Two Epidemiological Investigations
Our research focuses on assessing the validity, usability, and value of a novel cell phone-based personal exposure measurement system known as "CalFit" in two existing epidemiologic studies. The CalFit system runs on commercially available Android phones as currently configured and allows us to collect data on physical activity with accelerometry, geographic location through global positioning (GPS), and air pollution. It also collects data on self- reported environments, including mood and behaviors, with ecological momentary assessment (EMA) through interactive text messages.
The CalFit system had already undergone numerous controlled validation studies, making it ready for evaluation in large epidemiological studies. The combined measurement of physical activity, location and mobility, exposure to air pollution, and EMA represents a major advance in technologies for characterizing the personal environment. We are evaluating the instrument in two ongoing epidemiological investigations:
- The Healthy PLACES study in Southern California, which is examining the impact of the built environment on physical activity and obesity in a Smart Growth community and five adjacent communities
- The Transportation, Air Pollution and physical ActivitieS (TAPAS) project in Barcelona, Spain, focused on determinants of active commuting by bicycling or walking
Both studies have enrolled subjects and are actively collecting accelerometry, GPS, or travel survey data. For Aim 1, we are validating the CalFit GPS and accelerometer by comparing the integrated sensor against current gold standard instruments (the Actigraph accelerometer and the Globalsat GPS). In Aim 2, we are demonstrating the added value of the EMA texting system for understanding self-reported behavior, mood, and social and physical contexts as antecedents, concomitants, and consequences of environmental exposures.
Another focus of our work has been combining commercially available sensors with circuit boards developed by our team to measure nitrogen dioxide and ozone in the field. We merge these data with GPS and accelerometry information to infer likely inhalations of air pollution. To demonstrate value added, we compare the inhalation estimates derived against those from assigning resolved spatial estimates of exposure to the home location.
We also deploy passive sensors for the two pollutants that yield an integrated 1-week sample. These are compared with the real-time assessments to demonstrate the added value of having the locational and physical activity information merged with the real-time air pollution data, which allows us to assess indoor and outdoor exposures, the likely inhalation, and the geographic context of exposures.
Cell phones are currently the most ubiquitous computing platform globally, and our research also demonstrates the viability of using a cell phone as an exposure characterization tool that will have widespread utility for numerous epidemiological applications. Historically, the collection of individualized exposure data has been intrusive to study participants and expensive, largely due to the need to wear multiple devices. The CalFit system integrates numerous data related to individual behaviors and environmental exposures using one cell phone.
Because cell phones are already used extensively, our sensor offers substantial advantages for large epidemiological studies. Our research is, therefore, highly relevant to the protection of public health because the new sensors allow researchers to unobtrusively measure many environmental exposures in a way that is convenient to study subjects and relatively inexpensive to researchers.