were reported on a Report of Casualty Form collected routinely and unrelated to the study. Cause of death was reviewed on every casualty report, and the investigators changed the casualty code on the form if it was found to be clearly incorrect. When possible, the cause of death of those who served in the Persian Gulf was verified through a registry of autopsies maintained by the Office of the Armed Forces Medical Examiner; however, no central repository of autopsy data was available to determine the accuracy of deaths of service members deployed elsewhere. The investigators reported that although death rates from unintentional injury were significantly higher in those deployed to the gulf region than in those deployed elsewhere (69.1 vs 41.2 per 100,000 person-years), suicide rates were significantly lower in those deployed to the gulf (3.78 vs 10.82 per 100,000 person-years). The standardized mortality ratios were 0.34 (95% CI 0.16-0.63) for suicides and 1.54 (95% CI 1.32-1.77) for unintentional deaths. This study was considered secondary because no supplemental source was used to determine causes of death in the control group.
Adams et al. (1998) used the Southeast Asia Combat Area Casualties Database to assess the risk of suicide in U.S. ground troops during their tours of duty in Vietnam between 1957 and 1973. They found a higher risk of suicide in men serving in the Army than in other branches of the military (OR 7.86, 95% CI 1.64-5.31). The study was included as a secondary study because no confirmatory sources were used to check the accuracy of the cause of death and only deaths during deployment in country were considered.
Kaplan et al. (2007) examined suicide rates in a nationally representative sample of male community residents in the United States who had completed National Health Interview Surveys in 1986-1994. Vital status and causes of death were determined for survey respondents from 1986 through 1997. Data were collected with face-to-face household interviews; response rates were 94-98%. The outcome variable in the study was death by suicide. The relative risk of suicide was estimated with the Cox proportional-hazards model and adjusted for demographics and socioeconomic, health, and functional status. Time to death was measured from the month of the interview to the month of suicide. Analyses that adjusted for self-reports of demographic characteristics, socioeconomic factors, and health status found that respondents who reported having ever served on active duty in the armed forces had twice the risk of suicide of those who reported being nonveterans (hazard ratio 2.04, 95% CI 1.10-3.80). The committee emphasizes that a major limitation of this study that the investigators did not include a measure of deployment, which is the exposure of interest for this report.
Several studies of suicide in veterans that confined their samples to VA populations (Fontana and Rosenheck 1995b; Lambert and Fowler 1997; Thompson et al. 2002) could not avoid the inherent bias of a self-selected population that elects to receive health care from VA, and so were not considered in this review.
Boscarino (2005) and Kramer et al. (1994) found that the risk of suicide or suicidal behavior was higher in Vietnam veterans with PTSD than in those without it. Boscarino (2005) used data from the VES to examine causes of death in 15,288 Vietnam Army veterans 30 years after the war and 15 years after the phase I telephone surveys and personal interviews. When the DIS was used, 10.6% of the theater veterans and 2.9% of the era veterans were found to have PTSD. Vital status and cause-specific mortality were determined with the VA BIRLS, the Social Security Administration Death Master File, and the National Death Index Plus. Veterans with PTSD were more likely than those without PTSD to have died since the 1985-1986 survey