Background: Physical inactivity and obesity are major public health issues. Recent studies have provided evidence that attributes of the built environment influence physical activity among adults and that factors such as greater urban sprawl are related to overweight and obesity. Few studies have developed objective individual-level measures of the built environment, a geographic scale that may be more relevant to certain types of physical activity, such as walking. In addition, further research is needed to assess the associations of both objective and perceived environmental factors with physical activity. In this 2-year exploratory study funded by the National Cancer Institute, we are addressing these research gaps. Purpose: The purpose of this poster presentation is to provide a brief overview of progress to date on a major component of this study, which is to develop objective measures of the built environment for approximately 30,000 women in the Nurses’ Health Study (NHS) using Geographic Information Systems (GIS) techniques and to examine associations with physical activity and weight-related outcomes. In particular, we will briefly summarize pilot work focused on development and testing of built environment variables. Methods: A sample of 300 NHS participants from six counties in Massachusetts, Pennsylvania, and California were selected for the pilot GIS work. Geocoded home addresses, U.S. Census population data, an InfoUSA facilities database, and street network files were loaded into ArcGIS 9.3. GIS methods were used to derive variables in three domains: 1) street connectivity, 2) land use mix, and 3) population density. For each domain at least two variables were created using different operational definitions. We also created variables using 400, 800, and 1200 m network buffers. We merged the built environment data with NHS survey data. Statistical analyses included calculating mean values for environmental variables, both overall and at the county level, and running correlations between environmental variables and physical activity outcomes. Next steps: A next step in the project is to create environmental variables for the full sample of NHS participants living in the three states (n≈30,000) and merge these data with NHS survey data. In addition, we are conducting a small validation study with the InfoUSA data. During November, we are also implementing a supplemental survey with a sub-sample (n≈3,800) of NHS participants to assess perceptions of the neighborhood environment and to collect detailed information on physical activity. In another component of the study, we are testing the use of available tools such as Google Map/Earth, Google Street View, and Microsoft Visual Oblique to develop micro-scale measures of the built environment, such as the presence of sidewalks availability and their condition.


physical activity, walking, built environment, urban design



Date of this Version