The association between lower neighborhood socioeconomic status and sleep health mediated by physical activity and TV viewing: Findings from a large U.S. cohort

Abstract

Despite adverse neighborhood effects on sleep, the underlying behavioral factors linking this association remain uncertain. This study aimed to (1) evaluate whether physical activity (PA) and sedentary behavior (SB) mediate associations between neighborhood socioeconomic status (nSES) and sleep, and (2) explore whether PA- and SB-mediated associations vary by racial and/or ethnic groups. Data came from the NIH-AARP Diet and Health Study (n=233,335). Self-reported sleep outcomes included: short sleep (< 7 h [h] versus 7–8 h), long sleep (≥ 9 h versus 7–8 h), and long napping (≥ 1 h versus <1 h). A standardized nSES was derived from census variables. Primary cross-sectional mediation analyses for self-reported moderate-to-vigorous PA (MVPA) and SB (television [TV] viewing) were conducted using linear regression with bootstrap-generated 95% bias-correct confidence intervals (BC CIs), adjusted for age, sex, racial and/or ethnic groups (White, Black, Hispanic, and/or other groups), education, and marital status. Secondary mediation analyses varied by racial and/or ethnic groups. Overall, lower nSES was associated with short sleep, long sleep, and long napping, mediated through MVPA (Odds ratio [OR] 1.002, 95%BC CI 1.001, 1.002; OR 1.008, 95%BC CI 1.007, 1.010; and OR 1.011, 95%BC CI 1.009, 1.012, respectively, all p <0.05) and TV viewing (OR 1.003, 95%BC CI 1.003, 1.004, OR 1.010, 95%BC CI 1.008, 1.012; and OR 1.033, 95%BC CI 1.030, 1.035, respectively, all p <0.05). Among White adults, lower nSES was associated with short sleep, long sleep, and long napping, mediated through MVPA and TV viewing. Among Black adults, lower nSES was related to long sleep, mediated through TV viewing, and long napping through MVPA and TV viewing. Among Hispanic adults, lower nSES was linked to long sleep, mediated through MVPA. Among other adults, lower nSES was associated with long napping, mediated through TV viewing. Living in lower nSES areas was related to sleep outcomes, partly mediated through MVPA and TV viewing. Exploratory mediation analyses revealed suggestive evidence for subgroup differences in the nSES-sleep associations. Efforts to reduce lower nSES to encourage PA and reduce sedentary time are warranted to improve sleep health.