Is the Association Between Neighborhood Characteristics and Sleep Quality Mediated by Psychological Distress?
An Analysis of Perceived and Objective Measures of 2 Pittsburgh Neighborhoods
Published in: Sleep Health, v. 2, no. 4, Dec. 2016, p. 277-282
Posted on RAND.org on November 30, 2016
Living in disadvantaged neighborhood environments is associated with poorer physical and mental health outcomes and higher overall mortality. However, the specific mechanisms underlying associations of neighborhood characteristics with health are not fully understood. Sleep quality represents an important potential mediator of these associations. Objectives: The objectives were to investigate associations of perceived and objective neighborhood characteristics with sleep and the extent to which associations are explained by psychological distress.
The sample includes randomly selected households from 2 racially/ethnically and socioeconomically similar Pittsburgh communities.
Participants included 873 African American adults (77% female) with a median per capita household income of $13,300. Data were collected from in-person household surveys (sociodemographics, psychological distress, perceived neighborhood characteristics), daily sleep diaries, objective neighborhood street segment audits, and city crime data. We analyzed perceived and objective neighborhood characteristics and their association with sleep quality, and the degree to which psychological distress explained observed associations.
Perceived neighborhood characteristics, including perceived safety (β = 0.13), neighborhood satisfaction (β = 0.14), social cohesion (β = 0.08), and perceived infrastructure (β = 0.07), were significantly associated with sleep quality (P values < .05), but objective neighborhood characteristics were not. Once psychological distress was considered, associations with perceived social cohesion and neighborhood infrastructure were fully attenuated. Associations of perceived safety and neighborhood satisfaction with sleep were attenuated by 20%–30% but remained significant.