Neighborhood Built Environment and Sleep Health: A Longitudinal Study in Low-Income and Predominantly African-American Neighborhoods

Abstract

In the present study, we examined the associations between physical characteristics of neighborhoods and sleep health outcomes and assessed the mediating role of physical activity in these associations. A longitudinal study (the Pittsburgh Hill/Homewood Research on Eating, Shopping, and Health (PHRESH) Zzz Study; n = 1,051) was conducted in 2 low-income, predominately African-American neighborhoods in Pittsburgh, Pennsylvania, with repeated measures of neighborhood characteristics and sleep health outcomes from 2013 to 2018. Built environment measures of walkability, urban design, and neighborhood disorder were captured from systematic field observations. Sleep health outcomes included insufficient sleep, sleep duration, wakefulness after sleep onset, and sleep efficiency measured from 7-day actigraphy data. G-computations based on structural nested mean models were used to examine the total effects of each built environment feature, and causal mediation analyses were used to evaluate direct and indirect effects operating through physical activity. Urban design features were associated with decreased wakefulness after sleep onset (risk difference (RD) = -1.26, 95% confidence interval (CI): -4.31, -0.33). Neighborhood disorder (RD = -0.46, 95% CI: -0.86, -0.07) and crime rate (RD = -0.54, 95% CI: -0.93, -0.08) were negatively associated with sleep efficiency. Neighborhood walkability was not associated with sleep outcomes. We did not find a strong and consistent mediating role of physical activity. Interventions to improve sleep should target modifiable factors, including urban design and neighborhood disorder.

Keywords: actigraphy; built environment; mediation analysis; neighborhood; physical activity; sleep; urban design; walkability.

Figure 1

Directed acyclic graph for a mediation analysis of neighborhood built environment and sleep health, PHRESH Zzz Study, 2013–2018. A_i,j represents neighborhood characteristics (e.g., urban design, walkability, neighborhood disorder) for the ith subject at the jth visit; M_i,j represents physical activity for the ith subject at the jth visit; and Y_i,j represents sleep health outcomes (e.g., sleep duration, insufficient sleep, wakefulness after sleep onset, and sleep efficiency) for the ith subject at the jth visit. C^1,2,3_i,j represents exposure-outcome confounders (C₁), exposure-mediator confounders (C₂), and mediator-outcome confounders (C₃). PHRESH, Pittsburgh Hill/Homewood Research on Eating, Shopping, and Health.

Figure 2

Exposure-mediator and mediator-outcome associations, PHRESH Zzz Study, 2013–2018. Each mixed-effect regression model adjusted for age, sex, income, education, employment, marital status, family structure, and duration of residence (years) in the current neighborhood. CI, confidence interval; OR, odds ratio; PHRESH, Pittsburgh Hill/Homewood Research on Eating, Shopping, and Health; WASO, wakefulness after sleep onset.

Figure 3

Mediation analysis results for A) insufficient sleep, B) sleep duration, C) wakefulness after sleep onset, and D) sleep efficiency outcomes, PHRESH Zzz Study, 2013–2018. Each plot shows results for the 4 built environment characteristics assessed (urban design, walkability, neighborhood disorder, and crime). “Total” depicts the total effect. Bars show 95% confidence intervals. NDE, natural direct effect; NID, natural indirect effect; OR, odds ratio; PHRESH, Pittsburgh Hill/Homewood Research on Eating, Shopping, and Health.