Provision of social support and mental health in U.S. military veterans

Social support has long been linked to better health and longevity, with poor support associated with higher mortality and worse well-being. Most research has emphasized received social support rather than the support individuals provide to others. Evidence suggests providing support may confer health benefits and even account for associations between receiving support and mortality. However, little is known about how providing support relates to a comprehensive range of mental health outcomes, potential interactions with structural and received support, and effects in higher-risk groups such as U.S. military veterans. This study aimed to: (1) estimate the prevalence of different types of provided support among U.S. veterans; (2) identify sociodemographic, military, health, and psychosocial correlates of providing support; and (3) examine independent and interactive associations of provided, structural, and received support with current internalizing psychiatric disorders (major depressive disorder [MDD], generalized anxiety disorder [GAD], posttraumatic stress disorder [PTSD]) and suicidal ideation (SI).

Foundational work has established that social support improves health and reduces mortality risk, with strong links to outcomes across cardiovascular disease, cancer, infectious disease, and cognitive functioning. In mental health, systematic review evidence indicates poorer social support predicts worse symptom severity and recovery in depression, and preliminary evidence suggests similar links in schizophrenia, bipolar disorder, and anxiety disorders. Research has largely focused on received support; fewer studies have examined provided support. A seminal study showed the mortality benefit of receiving support was nullified after adjusting for providing support, implicating provision as a key factor. Subsequent work suggests providing support is associated with longevity and positive health outcomes and that volunteering/civic engagement may reduce depression. Gaps remain: few studies have assessed provided support in relation to multiple adverse mental health outcomes; interactions with structural support (network size) and received support are understudied; and little evidence pertains to high-risk populations such as U.S. veterans.

Design and sample: Cross-sectional analysis of the 2019–2020 National Health and Resilience in Veterans Study (NHRVS), a nationally representative online survey of U.S. military veterans drawn from Ipsos KnowledgePanel®. Of 7860 invited, 4069 (51.8%) completed a 50-minute survey (survey period 11/18/2019–03/08/2020; median completion 11/21/2019). Poststratification weights aligned the sample to the U.S. veteran population. The VA Connecticut Healthcare System IRB approved the protocol; all participants provided informed consent. Measures: Sociodemographics included age, gender, race/ethnicity, education, marital/partnered status, income, employment/retired status, military characteristics (combat veteran status, enlistment vs. drafted, years in service). A single item assessed perceived positive effect of military service (1–7). Health variables included number of medical conditions (0–24). Adverse childhood experiences (ACE) were measured with the ACE questionnaire (0–10). Potentially traumatic events (PTEs) were assessed using the Life Events Checklist for DSM-5, summing direct and indirect exposures. Lifetime psychiatric diagnoses of MDD, alcohol use disorder (AUD), and drug use disorder (DUD) were assessed with a modified MINI for DSM-5; lifetime PTSD was assessed with the PCL-5 (≥33 indicating a positive screen). Personality was measured via the Ten-Item Personality Inventory (TIPI) covering the Big Five traits on 1–7 Likert scales. Social support constructs: Provided social support was assessed with a modified 5-item Medical Outcomes Study Social Support Scale (MOS-SSS), querying frequency of providing different types of support (1=none to 5=all the time); total score 5–25; Cronbach’s α=0.88. Structural support measured the count of close friends/relatives (0–90). Frequency of social engagement captured days per week visiting family and visiting friends (0–7). Received social support used a parallel 5-item MOS-SSS of perceived availability (1–5; score 5–25; α=0.89). Outcomes: Current internalizing symptom screens included MDD (PHQ-4 depression subscale; score ≥3) and GAD (PHQ-4 anxiety subscale; established cut score), PTSD (current PCL-5 screening), and past-year suicidal ideation (SI). [Note: SI analyses additionally adjusted for lifetime MDD, PTSD, AUD, and DUD.] Analytic approach: Descriptive statistics characterized prevalence of provided support types and sample features. Bivariate correlations and multivariable linear regression identified correlates of provided support (model R²=0.48). For psychiatric outcomes, weighted hierarchical logistic regression models were conducted separately for MDD, GAD, PTSD, and SI: Step 1 included standardized main effects of provided, structural, and received support; Step 2 added interaction terms (provided×received; provided×structural). Models adjusted for age, gender, race/ethnicity, education, marital/partnered status, occupational status, income, enlistment status, combat veteran status, years of service, number of medical conditions, ACEs, direct/indirect PTEs, and frequency of social engagement with friends and family. Multiple-comparison-corrected significance threshold p<0.0125 was applied for main effects; non-significant (NS) interactions noted.

• Sample characteristics: Mean age 62.2 years (SD 15.7; range 22–99); 90.2% male; 78.1% White non-Hispanic; 35.0% combat veterans. Mean provided support score 19.0 (SD 4.3; range 5–25).
• Prevalence of provided support types: Most veterans reported providing support most/all of the time across domains—loving support 71.5%, confiding support 70.7%, helping with personal problems 64.8% (others included leisure/relaxation and instrumental support).
• Correlates of providing support (multivariable): Positive associations with being married/partnered, perceiving positive effect of military service (β=0.08, B=0.23, p<0.001), more direct and indirect PTEs (direct: β=0.05, B=0.09, p<0.001; indirect: β=0.03, B=0.02, p=0.002), higher Big Five traits—extraversion (β=0.07, p<0.001), agreeableness (β=0.12, p<0.001), conscientiousness (β=0.12, p<0.001), emotional stability (β=0.03, p=0.027), openness (β=0.10, p<0.001)—greater structural support (β=0.03, p=0.014), received support (β=0.49, p<0.001), and days visiting family per week (β=0.03, p=0.002). Negative associations with male gender (β=-0.05, B=-0.83, p<0.001) and number of medical conditions (β=-0.03, B=-0.06, p=0.002).
• Psychiatric outcome prevalence: MDD 8.7% (n=292), GAD 7.8% (n=232), PTSD 6.6% (n=219), past-year SI 11.7% (n=387).
• Main effects (weighted, adjusted logistic regressions): Greater provided support was associated with lower odds of adverse outcomes, with each 1 SD increase corresponding to approximately 22–32% lower odds across internalizing disorders and SI. Structural support was inversely associated with MDD and SI. Received support was inversely associated with all outcomes.
• Interactions: Significant provided×received support interactions for MDD, GAD, and SI. Among veterans with high received support (+1 SD), those also high in provided support (+1 SD) had markedly lower probabilities than those low in provided support (−1 SD): MDD 0.9% vs. 13.0% (≈14-fold lower), GAD 1.4% vs. 17.1% (≈12-fold lower), SI 3.8% vs. 13.4% (≈3.5-fold lower).
• Post-hoc item-level findings: Being a confidant predicted lower MDD odds (OR 0.79, 95% CI 0.70–0.89); providing companionship for relaxation predicted lower GAD (OR 0.84, 95% CI 0.71–0.99); helping with daily chores predicted lower suicidal intent (OR 0.55, 95% CI 0.42–0.73); helping with personal problems predicted lower GAD (OR 0.79, 95% CI 0.67–0.92); providing loving support predicted lower PTSD (OR 0.80, 95% CI 0.69–0.92) and SI (OR 0.82, 95% CI 0.73–0.92).

This nationally representative study of U.S. veterans shows that providing social support is common and is strongly linked to better mental health. Personality traits (agreeableness, conscientiousness, extraversion, emotional stability, openness) and greater received and structural support were associated with more frequent provision of support, suggesting that social connectedness and trait dispositions may promote supportive behaviors. The independent inverse associations between provided support and internalizing psychopathology and suicidal ideation, coupled with significant interactions indicating that high provision enhances the protective effects of high received support, underscore the multifaceted nature of social support. Neurobiological models involving reward-related regions (ventral striatum, septal area) and reduced amygdala reactivity provide plausible mechanisms for the mental health benefits of providing support. Socioeconomic and relational resources (e.g., higher income, marriage/partnership) were also linked to greater provision, aligning with literature on social capital. Clinically, findings suggest that interventions and social prescriptions that encourage veterans to provide support to others—through peer programs, volunteering, or structured clinical tasks (e.g., behavioral activation)—may reduce risk for depression, anxiety, PTSD symptoms, and suicidal ideation, particularly when combined with strengthening received support.

This study provides the first population-based estimates of how frequently U.S. veterans provide social support, identifies key correlates, and demonstrates that greater provision is associated with substantially lower odds of internalizing disorders and suicidal ideation. The combined presence of high received and high provided support confers especially strong protection. Findings support developing and testing interventions and policies that bolster veterans’ engagement in providing support to others, both within and outside the VA system. Future research should employ longitudinal designs to clarify causal pathways; examine mechanisms linking provided, received, and structural support to mental health; test personalized approaches based on social connectedness profiles; include more diverse veteran and non-veteran samples; and evaluate the efficacy of programs that train and encourage at-risk individuals to provide support.

• Generalizability: The sample was primarily older, male, and White non-Hispanic veterans; results may not generalize to younger or more diverse veterans or non-veterans.
• Cross-sectional design: Temporal/causal relationships between providing/receiving/structural support and mental health outcomes cannot be inferred; support behaviors may be time- and context-dependent.
• Measurement constraints: Social support measures did not capture potentially non-supportive or harmful aspects of close relationships.
• Outcome assessment: Psychiatric outcomes were assessed with screening instruments rather than structured clinical interviews, which may affect diagnostic accuracy.