Identifying suicidal behavior: A latent class analysis of local yrbs data

Background/purpose: Risk factors associated with suicidal behavior (suicide ideation, suicide planning, suicide attempt, and injury from suicide attempt) have been well documented. However, patterns of co-occurrence among these factors and other health behaviors have not been fully identified. Geographical variables representing sub-county areas have been included in the analysis to make results from the YRBS more useful to local policy makers. Objective: To assess health behavior patterns of suicidal adolescents and identify if those patterns change between health zones (place of adolescents' residence). Methods: Latent class analysis was used to investigate patterns of health behaviors presented by suicidal adolescents using data from the 2011 Duval County High School Youth Risk Behavioral Survey. This data is unique in that it contains information about the adolescents' place of residence. The sample included 3,336 students from 21 Duval County public high schools. The analysis is designed to identify how health behaviors among suicidal adolescents change specific to where they live. The health behaviors examined include variables related to: Victim of violent behaviors, substance abuse and demographic variables. Results: Currently, the prevalence of each suicide related questions (ideation to attempt with injury) are 14.9%, 14.7%, 12.7% and 4.9%, respectively. Preliminary analysis shows that distinct classes related to suicidal behaviors were identified with gender and health zones. Conclusions: Adolescent suicidal behavior can be identified through latent class analysis. Identification of characteristics can focus limited funding and define policy to address morbidity and mortality.