Evaluating Evidence of Risk for Suicide Among Veterans

Abstract

There is evidence of increased rates of suicide among veterans when compared to those without history of military service. However, empirical studies of the associations between military service and risk for suicide have reported confl icting results. Results from studies of mortality among Vietnam veterans have suggested that rates of suicide may be elevated following service in a combat area. 1 However, not all studies have reported statistically signifi cant increases in suicide mortality among veterans from the Vietnam War. 2‐4 Assessments of postservice mortality among veterans of the Persian Gulf War similarly failed to identify a statistically signifi cant increase in risk for suicide. 5,6 A more recent study of suicide risk among more than 490,000 veterans of Iraq and Afghanistan did report a statistically signifi cant increase in suicide risk among veterans who had been on active duty (standardized mortality ratio 1.33) and among those who had been diagnosed with a mental disorder and had received services from the Veterans Health Administration (VHA) (standardized mortality ratio 1.77). 7 Analyses of existing survey data, when linked with information on the participant’s manner of death, can be used to identify risk for suicide when veteran status is included in the original data fi le. A recent analysis of more than 300,000 male participants in the National Health Interview Survey series (1986‐ 1997) reported a twofold increase in risk for suicide among veterans when compared to those without self-reported history of military service. 8 Somewhat paradoxically, another prospective study of mortality among more than 500,000 male participants in a large cancer survey using similar methodology failed to identify an increased risk for suicide among veterans. 9 Contradictory fi ndings exist in prior work partially because, while national-level mortality data provide an important surveillance tool for understanding changes in overall and groupspecifi c U.S. rates of suicide, veteran-specifi c data are limited. The National Center for Health Statistics (Centers for Disease Control and Prevention) does not currently upload veteran identifi ers as part of their routine death reporting protocol from states that collect this information. Rates of suicide for the entire U.S. veteran population are therefore not routinely attainable. The consequence of current gaps in data availability has been a general reliance on convenience data (such as state-level mortality records) or other existing data sources (including a veteran identifi er) that are linked with information from the National Death Index. In either case, veteran status is often obtained through self- or proxy-report and the sampling procedures employed by existing studies are unlikely to ensure representation of high-risk subpopulations. Questions surrounding the reliability or validity of much existing data remain largely unanswered. Despite challenges associated with ascertainment of veteran status and the relative lack of data for the identifi cation of clinical and behavioral characteristics associated with increased risk for suicide, our understanding of suicide risk among some veteran groups is improving. By linking administrative data obtained from the VHA with national level cause of death data obtained from the National Death Index, researchers from the VA-funded Serious Mental Illness Treatment Research and Evaluation Center (SMITREC) were able to identify 1,613 suicides among veterans who had received inpatient or outpatient VHA services during the 2000 and 2001 fi scal years. 10 When compared to suicide rates for the general U.S. population, the authors identifi ed a statistically signifi cant increase in the relative risk for suicide for both male (standardized mortality ratio 1.66) and female (standardized mortality ratio 1.87) veterans. 10 Comparisons of rates of suicide for veterans and similarly aged groups from the U.S. general population are useful, yet may lack suffi cient information to identify potential differences in clinical or behavioral characteristics among veterans and nonveterans or between veterans who do and do not receive VHA services, effectively limiting our ability to identify those characteristics associated with increased risk. The introduction and limited expansion of the National Violent Death Reporting System (NVDRS) has provided data that can be used to compare the prevalence and characteristics of suicide among veterans and nonveterans in a small number of states. In 2006, there were 1,594 veteran suicides identifi ed in the 16 states participating in NVDRS and 5,966 suicides among those who were not identifi ed as veterans of military service (approximately 21% of all suicides in these areas). However, results from comparisons of NVDRS data should be interpreted with caution. To date, NVDRS has only been implemented in a limited number of states and the