Aim: The current study had two aims: (1) to investigate whether single-item measures of subjective evaluation of neighborhood (i.e., perceived neighborhood safety and quality) predict long-term risk of mortality and (2) to test whether these associations depend on race and gender.
Methods: The data came from the Americans' Changing Lives Study (ACL), 1986-2011, a nationally representative longitudinal cohort of 3361 Black and White adults in the USA. The main predictors of interest were perceived neighborhood safety and perceived neighborhood quality, as measured in 1986 using single items and treated as dichotomous variables. Mortality due to all internal and external causes was the main outcome. Confounders included baseline age, socioeconomic status (education, income), health behaviors (smoking, drinking, and exercise), and health (chronic medical conditions, self-rated health, and depressive symptoms). Race and gender were focal effect modifiers. Cox proportional hazard models were ran in the pooled sample and stratified by race and gender.
Results: In the pooled sample, low perceived neighborhood safety and quality predicted increased risk of mortality due to all causes as well as internal causes, net of all covariates. Significant interaction was found between race and perceived neighborhood safety on all-cause mortality, indicating a stronger association for Whites compared to Blacks. Race did not interact with perceived neighborhood quality on mortality. Gender also did not interact with perceived neighborhood safety or quality on mortality. Perceived neighborhood safety and quality were not associated with mortality due to external causes.
Conclusion: Findings suggest that single items are appropriate for the measurement of perceived neighborhood safety and quality. Our results also suggest that perceived neighborhood safety better predicts increased risk of mortality over the course of 25 years among Whites than Blacks.
Keywords: Ethnic groups; Life course; Mortality; Neighborhood; Race; Safety.