Metropolitan status and self-reported health in the BRFSS: Does where you live matter?

Numerous sources suggest that cities are healthier than other areas, promoting the idea of an urban health advantage due to characteristics including access to commodities, a health-promoting built environment, or increased social cohesion. Others disagree, citing air pollution, crime, and socioeconomic inequalities in urban areas. However, little data exists to support either position. Given the increasing urbanization of the domestic and international population, there is a need to quantify the presence or absence of an urban health advantage.

Complex survey data from the 2005 iteration of the Behavioral Risk Factor Surveillance System was used to assess the relationship between metropolitan status and self-rated health. Numerous analyses have ascertained that self-reported health serves as an efficacious measure of overall health and mortality. The study design controlled for age, sex, smoking status, educational attainment, employment status and income. Two statistical packages (SAS software, Version [9] and Stata Statistical Software: Release 9) were utilized for data cleaning and analysis.

Linearized logistical regression models for survey data were utilized to analyze the data. The data were weighted by the primary sampling unit (psu), strata, and sampling weight. For each data set, the data was viewed by psu and any strata with only one entry in the psu were dropped to facilitate statistical analysis. Subpopulation analyses were conducted by gender and by racial/ethnic subgroups. The analysis on the total data set controlled for race/ethnicity and gender, whereas the analysis on the gender subpopulations controlled for race/ethnicity and each the racial/ethnic subpopulation analysis controlled for gender.