Measures of neighborhood walkability and their association with diabetes and depressive symptoms in black women

Abstract

Using data from the prospective Black Women's Health Study (BWHS; n=18,525), the impact of neighborhood walkability on incident diabetes and depressive symptomology was assessed. Neighborhood walkability was assessed for BWHS participants residing in Chicago, IL, Los Angeles, CA or New York, New York in 1995, 1997 or 1999 by factor-analyzing 14 components (e.g., sidewalk coverage, number/type of intersections, population/housing density) into a single continuous measure ("neighborhood walkability"), divided into four categories (least walkable, 2nd least walkable, 2nd most walkable, most walkable). Compared to women living in a most walkable neighborhood in 1995, women living in a least walkable neighborhood had a modestly higher hazard of incident diabetes over 16 years of follow-up (IRR=1.06; 95% CI=0.90-1.24) and a higher risk of depressive symptomology, using 1999 and 2005 Center for Epidemiologic Studies (CES-D) scores (CES-D≥16: RR=1.02, 95% CI=0.94-1.11; CES-D≥25: RR=1.18, 95% CI=1.02-1.37). Associations with incident diabetes were strongest among women who resided in a higher SES neighborhood, among the healthiest women (lowest body mass index, most frequent vigorous exercise), and when follow-up was lagged four or 10 years, while associations with depressive symptomology were strongest among women who resided in Los Angeles, who engaged in no vigorous activity at baseline, or who averaged less than one per hour travelling in a car or bus. Neighborhood walkability and neighborhood socioeconomic status (SES) were strongly inversely related; disentangling these two aspects is a methodological challenge to assessing the health impact of the built environment. In a subset of 3,000 addresses, neighborhood walkability in 1999 was highly positively correlated (r=0.70) with WalkScore® (a free, publicly-available measure relying upon walking distances to specified amenities) in 2012-13. Neighborhood walkability and WalkScore® were both valid measures of neighborhood walkability, best gauged using total length of bus routes, number of intersections, and total sidewalk length. WalkScore® has the potential to serve as a "standard" neighborhood walkability measure, allowing easier comparison of the health impact of the built environment across many studies.