What’s this about?
The Built Environment & Health (BEH) project is an interdisciplinary program of research at Columbia University. Led by epidemiologist Andrew Rundle, BEH uses spatial data to examine the implications of the built environment, including land use, public transit, and housing, for physical activity, diet, obesity, and other aspects of health. With a focus on New York City, BEH research will inform public policy to promote health in the city and metropolitan area. BEH is affiliated with the Robert Wood Johnson Foundation Health & Society Scholars program at Columbia.
The Urban Built Environment and Obesity in New York City
A Multilevel Analysis
In the BEH group’s first publication, Andrew Rundle and colleagues examine the relationship between built environment characteristics and obesity in New York City, replicating measures widely used in previous studies of other cities. Urban planning and public health research suggests that pedestrian-oriented environments, characterized by high street connectivity, mixed land use and high population density, encourage travel by walking and bicycling. By reducing reliance on privately-owned automobiles, public transit is also expected to promote pedestrian activity. As we might expect based on this perspective, recent research finds an association between environmental characteristics such as population density, availability of nearby destinations, and intersection density and both self-reported and objective measures of physical activity, walking, and biking. In a handful of recent studies, these built environment characteristics are also associated with rates of obesity or Body Mass Index (BMI).
Questions remain, however, about whether these conclusions apply to extremely densely-settled cities such as New York. Most national studies in this literature employ measures of the built environment at the county or metropolitan level, which fail to capture the variability within cities. Further, most studies that have focused on individual cities are sited in lower-density places such as Austin, Texas; Atlanta, Georgia; and Portland, Oregon.
To examine whether urban form is associated with body size within New York City, we conducted a multi-level analysis using data from the New York Cancer Project. The sample includes 13,102 adult volunteers from the city’s five boroughs who were recruited to the study between January 2000 and December 2002. The dependent variable was objectively measured BMI. Our built environment measures were land use mix, bus and subway stop density, population density and intersection density, each constructed for the census tract in which the study subject lived. Covariates included age, gender, race, education, and census-tract level poverty and race/ethnicity.
Mixed land use (Beta = -0.55, P<0.01), bus stop density (Beta = -0.01, P<0.01), subway stop density (Beta = -0.06, P<0.01), and population density (Beta = -0.25, P<0.001), but not intersection density (Beta = -0.002) were significantly inversely associated with BMI after adjustment for individual- and neighborhood-level sociodemographic characteristics. If we compare the 90th to the 10th percentile of each built environment variable, the predicted adjusted difference in BMI with increased mixed land use was -0.41 units, bus stop density was -0.33 units, subway stop density was -0.34 units, and population density was -0.86 units.
Although the variation in BMI across census tracts represents only a modest portion of the total variation in BMI, individuals living in tracts that were more “pedestrian-friendly” had significantly lower BMI compared with other New Yorkers. These findings are consistent with current literature relating the built environment to travel mode, physical activity, and obesity; their distinctive contribution lies in showing that these relationships hold within a high-density urban environment. From a public health perspective, if each study subject’s BMI were reduced by a half unit, 10 percent of the overweight subjects in our sample would move into the normal BMI range, and 10 percent of obese subjects would shift into the overweight category.
Forthcoming, American Journal of Health Promotion. By Andrew Rundle, Ana V. Diez Roux, Lance Freeman, Douglas Miller, Kathryn M. Neckerman, Christopher Weiss.
