Modeling stable and dynamic vulnerabilities in suicide risk: A mechanistic test of fluid vulnerability theory in military personnel with suicidal ideation

Behav Res Ther. 2026 Apr 25;202:105052. doi: 10.1016/j.brat.2026.105052. Online ahead of print.

ABSTRACT

BACKGROUND: Suicide attempts among U.S. service members and Veterans with suicidal ideation remain poorly characterized despite their disproportionate contribution to suicide mortality. The present study tested whether the latent structure of suicide risk aligns with Fluid Vulnerability Theory’s (FVT) distinction between stable vulnerabilities and fluid state processes.

METHODS: Pooled baseline data from the Military Suicide Research Consortium Common Data Elements (2010-2023; N = 2246) were analyzed using multiple imputation and zero-inflated negative binomial (ZINB) regression. We compared three specifications: a fully specified model including all candidate predictors, an FVT-guided model differentiating trait and state processes, and a hybrid model incorporating demographic covariates.

RESULTS: Consistent with preregistered hypotheses, fully specified models showed poor convergence, indicating overparameterization. In contrast, both the FVT and hybrid models converged robustly and demonstrated interpretable dual-process structure. Across models, symptoms of posttraumatic stress disorder (PTSD; OR = 1.14-1.15) and traumatic brain injury (TBI; OR = 1.20-1.22) consistently predicted higher lifetime suicide-attempt counts. Inclusion of demographic covariates improved overall fit (ΔAIC = -41.1) but increased model complexity (ΔBIC = +56.0). The zero-inflated (structural-zero) components were stable, suggesting that suicide risk in this population is primarily driven by gradients of vulnerability rather than a distinct “non-attempter” class. Notably, stable vulnerabilities (e.g., hopelessness) showed stronger bivariate associations with ideation severity, supporting their role as ‘gatekeepers’ to the risk process.

CONCLUSIONS: Modeling suicide attempts using a coupled stable and fluid process framework yields superior convergence and theoretical coherence relative to saturated specifications. The findings provide quantitative support for FVT, linking chronic vulnerabilities and transient perturbations in a unified mechanistic framework that can inform precision assessment and intervention development.

PMID:42085747 | DOI:10.1016/j.brat.2026.105052