![[ Visit Client Website ]](images/banner.gif)
Previous studies have investigated spatial patterning and associations of area characteristics with suicide rates in Western and Asian countries, but few have been conducted in the United States. This study aims to: 1) identify spatial and spatio-temporal clusters of high suicide rates in Ohio; and 2) assess the relationship between demographic and socioeconomic contextual factors and suicide clusters. Understanding the spatial distribution of suicide can inform the planning, implementation, and evaluation of suicide prevention efforts.
Methods: This ecological study included all people who died by suicide in Ohio between January 1, 2004 to December 31, 2013. Suicide decedents were identified from death certificate data obtained from the Ohio Department of Health. Deaths by suicide were identified based on the International Classification of Diseases, 10th revision (ICD-10) external cause of death codes (X60-X84, Y87.0, *U03). The unit of analysis in defining a suicide cluster was the census tract (N= 2, 952). To identify suicide clusters, each suicide case was assigned to a census tract based on the residential address before death recorded in the death certificates.
Census data were used to obtain measures of census level neighborhood characteristics. Because these variables were highly correlated, principal component analysis with Varimax rotation was used to create latent variables. Using the 16 variables obtained from the 2008-2012 American Community Survey and the 2010 Area Resource File resulted in three area-level latent variables: economic deprivation, density of providers, and social fragmentation. Logistic regression was then used to examine the association between the three latent variables and being in a high-risk suicide cluster.
To identify the spatial distribution of suicide and test for clustering, spatial and spatio-temporal scan statistics were used to detect high-risk clusters of suicide at the census tract level.
Results: Nine statistically significant (p <0.05) high-risk spatial clusters and two space-time clusters were identified. The risk of suicide was up to 2.1 times higher in high-risk clusters than in areas outside these clusters (relative risk ranged from 1.22 to 2.14, p<0.01). In the multivariate model, factors strongly associated with area suicide rates were socio-economic deprivation (Odds ratio [OR] = 3.4, 94% CI = 2.6-4.5) and density of providers (OR = 0.35, 95% CI= 0.27-0.46).
Conclusions and Implications: Statistically significant high-risk spatial clusters were identified; most were in major cities and rural counties located in the Appalachian regions of Ohio. Heightened community wide suicide risk is associated with socioeconomic deprivation and lower levels of provider density relative to other communities. Efforts to reduce poverty and improve access to health and mental health services on the community level represent potentially important strategies to prevent suicide.