In the last decade, our understanding of neighborhoods'role in obesity prevention has progressed. However, much of the research has focused on the relationships between the built environment around a residential address and physical activity (PA). Focusing on the environment around static address points has several drawbacks. Health behaviors occur in multiple locations and along routes to destinations. Focusing on one location, such as residential address, underestimates the exposure to multiple environments. Further, health behaviors are usually assessed in an aggregate manner, for example, total PA across the day. Relating one location to total behavior results in underestimated effect sizes, especially if the behavior does not primarily occur at that location. The emergence of lightweight, low cost, and accurate GPS devices has enabled researchers to objectively track the location of an individual. While GIS data have been used in conjunction with GPS data, to date only a static approach has been employed;i.e. using the GIS to determine time in certain locations, but not exploring route buffers or daily exposures to multiple land uses across time and space. In two study samples of almost 500 participants with GPS and accelerometer data, we will develop novel, dynamic GIS based variables of daily exposure to environments that support PA (e.g. parks, connected streets, access to public transportation, and multiple destinations) and healthy eating (e.g. grocery stores). With the GPS traces, we can create """"""""expozones"""""""" around multiple locations, investigate temporal associations, and estimate environmental exposures along routes. We can weight exposures by mode and speed and investigate GIS based land uses """"""""in view"""""""" of the GPS traces. New techniques to summarize the environmental exposures will be developed to avoid regression to the mean from varying exposures across the day. One data set includes prospective data from an ecological intervention trial in 16 retirement communities. This will allow us to assess changes in """"""""home range"""""""" (how widely people travel) based on the built environment. With only 16 communities, we can also feasibly observe and code the local land uses to assess the validity of the GIS databases. With more refined behavioral measures, inaccuracies in the GIS data become more problematic. The other dataset includes greater variability in residential addresses, increasing the generalizability of our findings and improving the comparison with traditional static location based GIS buffers against our dynamic exposure measures across the day. Further this second data set includes measures of dietary intake that will allow us to assess the impact of full-day exposure to food environments on healthy eating, which is likely to be a major improvement. Our multidisciplinary team of experts will develop and validate the new measures against behaviors and health outcomes using a spilt halves method. This measurement development work will support two R01 applications: 1) a multi-country GPS study 2) application of new methods to existing GPS samples (N=2850).
We are only just starting to understand how where we live, work and spend our time affects how healthy we are. This study will use GPS data that provides traces of people throughout the day, showing where they are spending their time and how they get there. Understanding the types of places where people are active and get their food will help us create environments that help people be active and eat better.
|Jankowska, Marta M; Natarajan, Loki; Godbole, Suneeta et al. (2017) Kernel Density Estimation as a Measure of Environmental Exposure Related to Insulin Resistance in Breast Cancer Survivors. Cancer Epidemiol Biomarkers Prev 26:1078-1084|
|Meseck, Kristin; Jankowska, Marta M; Schipperijn, Jasper et al. (2016) Is missing geographic positioning system data in accelerometry studies a problem, and is imputation the solution? Geospat Health 11:403|
|Jankowska, Marta M; Schipperijn, Jasper; Kerr, Jacqueline (2015) A framework for using GPS data in physical activity and sedentary behavior studies. Exerc Sport Sci Rev 43:48-56|