Exploring How the Interplay Between Neighborhood Perceptions and Social Cohesion Relates to Self-Assessed Health: A Four-Way Decomposition with a 10-Year Follow Up

Abstract The present study investigated the extent to which the relationship between perceptions of neighborhood unsafety and disorder and self-assessed health (SAH) is explained by mediation, interaction, or a combination of both through perceived neighborhood social cohesion. We analyzed data on 5650 respondents over a 10-year follow-up using the GLOBE study, a prospective cohort study in the Netherlands. Log-linear regression analyses were used to estimate the total effects of feeling unsafe and perceived neighborhood disorder with poor SAH. A four-way decomposition approach was used to decompose the total effects into four components: controlled direct effect (neither mediation nor interaction), pure indirect effect (mediation only), reference interaction effect (interaction only), and mediated interaction effect (both mediation and interaction). The results indicate that feeling unsafe had a positive estimated total effect on poor SAH (RR = 1.05; 95% CI, 1.01, 1.10). For perceived neighborhood disorder, the estimated total effect was smaller (RR = 1.01; 95% CI, 1.00, 1.04). Decomposition analysis indicated that the majority of the estimated effect of perceived unsafety was attributed to the controlled direct effect (ERR = 0.09; 95% CI, 0.05, 0.13), with no evidence of mediation or interaction through social cohesion. None of the estimated decomposition components were significant for neighborhood disorder. These findings suggest that perceived neighborhood disorder and unsafety and perceived neighborhood social cohesion appear to influence health through independent pathways, rather than through their interplay.