Assessing the impact of social cohesion on suicide-related emergency calls: a spatial modeling approach

Given the relevance of community-level approaches to suicide prevention, this study aims to examine how neighborhood social cohesion relates to spatial variation in suicide-related emergency calls and whether its inclusion enhances model performance compared to sociodemographic factors alone.

We analyzed geocoded data on suicide-related emergency calls (N = 6,271) aggregated across 552 census block groups in Valencia (Spain) from 2021 to 2023, provided by the Valencian Regional Government. Sociodemographic indicators were obtained from official municipal statistics (2020), and neighborhood social cohesion was measured using survey data collected across census block groups between 2021 and 2022. Bayesian hierarchical Poisson models were implemented to estimate area-level associations, including models with (1) sociodemographic covariates only, (2) addition of social cohesion, (3) inclusion of an unstructured spatial random effect, and (4) a full spatial model with structured and unstructured random effects.

The model including sociodemographic factors, social cohesion, and an unstructured random effect provided the best overall fit. Including social cohesion markedly improved model fit and revealed a robust negative association with suicide-related emergency calls: areas with lower cohesion showed a higher relative risk. Additionally, indicators of social disorganization were positively associated with call rates, confirming their joint contribution to neighborhood-level differences in suicide-related calls.

The results support the association between higher community social cohesion and lower rates of suicide-related emergency calls. Introducing neighborhood-level social cohesion improves the explanatory power of the spatial model, highlighting its role in neighborhood inequalities and its potential to guide targeted preventive interventions.