1169 — Linking VA and CDC data to improve understanding of circumstances contributing to US Veteran suicides and deaths of undetermined intent
Lead/Presenter: Claire Hoffmire,
Rocky Mountain MIRECC for Suicide Prevention
All Authors: Hoffmire CA (Rocky Mountain MIRECC for Suicide Prevention), Gaeddert LA (Rocky Mountain MIRECC for Suicide Prevention) Schneider AL (Rocky Mountain MIRECC for Suicide Prevention) Kittel-Moseley JA (Rocky Mountain MIRECC for Suicide Prevention) Monteith LL (Rocky Mountain MIRECC for Suicide Prevention)
US Veterans experience elevated suicide rates, yet research predominantly focuses on the 30% using Department of Veterans Affairs (VA) healthcare. Furthermore, while suicide rates in men are higher than in women, excess risk conferred by Veteran status is higher among women. Finally, little is known about relationships between social determinants of health (SDOH) and suicide, which may differ by gender. Linking VA and non-VA data sources and conducting comparative analyses by gender and VA healthcare use can improve understanding of suicide drivers to inform enhanced, gender-sensitive Veteran suicide prevention. Accordingly, we linked VA and Centers for Disease Control and Prevention (CDC) data to describe SDOH, mental health (MH) and physical health (PH) circumstances surrounding US Veteran suicides and violent deaths of undetermined intent.
A multi-stage deterministic linkage of VA-DoD Mortality Data Repository (MDR) death records and CDC National Violent Death Reporting System (NVDRS), which contains limited personal identifying information but no direct identifiers, was conducted across 42 participating states (2012-2018). Exact (all 7 linkage variables matched), probable (all but one variable matched), and possible (all but 2-3 variables matched) matches were identified. Demographic characteristics and NVDRS-documented circumstances prior to death (MH and PH problems, SDOH) were compared across six study groups defined by gender and VA healthcare use (never; past [preceding the year prior to death], current [within year prior to death]). Frequencies with 95% confidence intervals were computed and chi-square tests compared findings across all groups and across VA groups, by gender.
Across all states and years, 23,997 matches (89.39%; 26,852 MDR records) were identified, which increased to 94.3% when high missingness or incomplete reporting NVDRS state-years were excluded. Age at death varied by VA group for men; past and current VA users were older. Race/ethnicity varied by VA group for men and women; current VA users were more likely to be Black or Hispanic compared to never users while past users were more likely to be Hispanic compared to never users. Across study groups, 89-94% of decedents had at least one known circumstance contributing to death. Gender differences were noted in the proportion experiencing specific circumstances, including a higher frequency of substance abuse, current MH problems and history of suicidal ideation and attempt among women and a higher frequency of crises, legal problems, and PH problems among men. Among men only, financial and job problems varied by VA group with frequency highest among never users.
Linking VA and NVDRS data is feasible and provides an important opportunity to improve understanding of circumstances of violent death among Veterans, including those for whom information has been lacking (e.g., those not engaged in VA healthcare, women). Legal and financial problems may be particularly important factors to address for men Veterans who do not use VA services whereas substance use may be particularly important to address among women Veterans.
Continued linkage of MDR and NVDRS data can enhance Veteran suicide surveillance efforts to ensure prevention efforts are better designed to address contributing factors for all Veterans, including those not using VA healthcare.