Moving suicide prevention upstream by understanding the effect of flourishing on suicidal ideation in midlife: an instrumental variable approach

Abstract

Prior research has examined the association between flourishing and suicidal ideation, but it is unknown whether this association is causal. Understanding the causality between flourishing and suicidal ideation is important for clinicians and policymakers to determine the value of innovative suicide prevention programs by improving flourishing in at-risk groups. Using a linked nationwide longitudinal sample of 1619 middle-aged adults (mean age 53, 53% female, 88% White) from the National Survey of Midlife Development in the United States (MIDUS), this retrospective cohort study aims to assess the causal relationship between flourishing and suicidal ideation among middle-aged adults in the US. Flourishing is a theory-informed 13-scale index covering three domains: emotional, psychological, and social well-being. Suicidal ideation was self-reported in a follow-up interview conducted after measuring flourishing. We estimated instrumental variable models to examine the potential causal relationship between flourishing and suicidal ideation. High-level flourishing (binary) was reported by 486 (30.0%) individuals, and was associated with an 18.6% reduction in any suicidal ideation (binary) (95% CI, - 29.3- - 8.0). Using alternative measures, a one standard deviation increase in flourishing (z-score) was associated with a 0.518 (95% CI, 0.069, 0.968) standard deviation decrease in suicidal ideation (z-score). Our results suggest that prevention programs that increase flourishing in midlife should result in meaningful reductions in suicide risk. Strengthening population-level collaboration between policymakers, clinical practitioners, and non-medical partners to promote flourishing can support our collective ability to reduce suicide risks across social, economic, and other structural circumstances.

Figure 1

Directed Acyclic Graph.The relationship between flourishing, X, and suicidal ideation, Y, can be estimated using our set of conditional instruments (ACEs, daily discrimination), Z, that are valid conditional on included covariates, W. Y (suicidal ideation) was measured after X (flourishing). W includes severe psychological distress, depression, anxiety, chronic pain, a set of five inflammatory markers, substance use, binge drinking, health status, personality factors, health insurance, and socioeconomic factors (age, sex, race/ethnicity, education, marital status, household income adjusted for household size). The function of Z, conditional on W, in this model is to remove the confounding effects of U (which is a set of unobserved confounders in the error term of the regression analysis) regarding the effect of X on Y. In other words, all these variables in W are included in closing off backdoor paths (non-causal paths) between X and Y, or in econometric terms, to remove any correlation between the set of instruments and the error term.

Figure 2

Distribution of Flourishing (Continuous Measures).