Abstract
•This research evaluated a large study population of US veterans—including a representative proportion of female veterans—with severe mental illness (SMI), and used a structured clinical interview to confirm diagnosis, which is more reliable than a clinical diagnosis.•Factors associated with suicidal behavior and death by suicide among men and women veterans with severe mental illness have similarities and differences from patterns in the general population.•Although suicide-specific mortality involved relatively few events over the six years these occurrences were an order of magnitude more frequent compared to the U.S. general population and also compared to the overall veteran population.•Female veterans with SMI had more suicide attempts than would be expected based on general population standards.•Findings from this study may have health-policy implications, including for targeted prevention strategies.
We compared male and female American veterans with schizophrenia or bipolar disorder regarding clinical characteristics associated with lifetime suicidal ideation and behavior. Subsequent mortality, including death by suicide, was also assessed.
Data from questionnaires and face-to-face evaluations were collected during 2011–2014 from 8,049 male and 1,290 female veterans with schizophrenia or bipolar disorder. In addition to comparing male-female characteristics, Cox regression models—adjusted for demographic information, medical-psychiatric comorbidities, and self-reported suicidal ideation and behavior—were used to examine gender differences in associations of putative risk factors with suicide-specific and all-cause mortality during up to six years of follow-up.
Women overall were younger, more likely to report a history of suicidal behavior, less likely to be substance abusers, and had lower overall mortality during follow-up. Among women only, psychiatric comorbidity was paradoxically associated with lower all-cause mortality (hazard ratio [HR]=0.53, 95% CI, 0.29–0.96, p = 0.037 for 1 disorder vs. none; HR=0.44, 95% CI, 0.25–0.77, p = 0.004 for ≥2 disorders vs. none). Suicide-specific mortality involved relatively few events, but crude rates were an order of magnitude higher than in the U.S. general and overall veteran populations.
Incomplete cause-of-death information and low statistical power for male-female comparisons regarding mortality.
Female veterans with SMI differed from females in the general population by having a higher risk of suicide attempts. They also had more lifetime suicide attempts than male veterans with same diagnoses. These differences should inform public policy and clinical planning.
