Article summary

Suicide Risk and Precipitating Circumstances Among Young, Middle-Aged, and Older Male Veterans Mark S. Kaplan, DrPH, Bentson H. McFarland, MD, PhD, Nathalie Huguet, PhD, and Marcia Valenstein, MD, MS

As the 11th leading cause of death among Americans of all ages, suicide remains a serious public health problem, and reducing suicide is a national imperative.1,2 The suicide of veterans has become a topic of intense public policy scrutiny in recent years. Until recently, most US studies indicated that both active-duty military personnel3---5 and veterans6 were at a lower risk for suicide than their demographically matched peers.

The increased risk of suicide recently ob- served among veterans of Operations Enduring Freedom and Iraqi Freedom has generated nationwide concern.7,8 The Department of Veterans Affairs (VA) has drawn attention to the rising suicide rate among young veterans and declared the prevention of suicide to be a major priority.9

Lawmakers have also expressed concern about the heightened rate of veterans’ suicides across the age spectrum.10 For example, the House Committee on Veterans Affairs held a hearing entitled “The Truth About Veterans’ Suicide” in May 2008,11 and more recently, the Senate Committee on Veterans Affairs con- ducted a hearing entitled “Mental Health Care and Suicide Prevention for Veterans” in March 2010.12 In his testimony before the House Committee in May 2008, the Secretary of Veterans Affairs13 presented data showing that male veterans in the community had higher rates of suicide than did other men and that veterans aged 30 to 64 years had the highest rates.

According to the VA Office of the Inspector General, 1000 veterans who receive VA care and as many as 5000 of all veterans die by suicide every year.9 In an analysis of suicide rates among male veterans and nonveterans in Washington state, which has large military bases and a substantial population of veterans, May- nard and Boyko14 found that the 2006 suicide rate was higher among veterans in all age categories. Data from the 2007 Oregon Violent Death Reporting System revealed that the age- adjusted suicide rate was 45.7/100 000 among

male veterans but 27.4/100 000 among non- veteran men.15 McCarthy et al.16 found that male (43/100 000) and female (10/100 000) VA patients had higher suicide rates than did non- veteran men (23/100 000) and women (5/100 000) in the general population, although these increases were likely due, in part, to health problems leading to VA use in addition to veteran status.

In a prospective follow-up study of 320 890 men who participated in the 1986 to 1994 National Health Interview Surveys, Kaplan et al.17 showed that veterans were twice as likely (hazard ratio [HR] = 2.13; 95% confidence in- terval [CI] =1.14, 3.99) to die of suicide as were male nonveterans in the general population. However, not all studies found that veterans were at increased risk for suicide. On the basis of a review of 13 studies of suicide risk among current and former military personnel, Kang and Bullman6 noted that veterans historically had a lower risk of suicide than did the general population. Similarly, Miller et al.18 found no

connection between military service and suicide in a large (but not nationally representative) longitudinal study of middle-aged and elderly men.

The Blue Ribbon Work Group on Suicide Prevention in the Veteran Population19 men- tioned the conflicting evidence regarding the risk of suicide among veterans across age groups. Some evidence suggested that younger vet- erans20 were more vulnerable to suicide than were their older counterparts. In the United Kingdom, Kapur et al.21 found age differences in the rates of suicide and in the prevalence of contact with mental health professionals before death among veterans. They showed that British veterans younger than 24 years were at greater risk for suicide (vs nonveterans) but that fewer had been in contact with mental health pro- fessionals.21 According to a recent analysis of data from Oregon, the rates of suicide for younger veterans have increased since 2005, whereas the rates for older veterans have de- clined.15 The foregoing studies might offer

Objectives. The purpose of this study was to evaluate the risk of suicide among

veteran men relative to nonveteran men by age and to examine the prevalence of

suicide circumstances among male veterans in different age groups (18–34, 35–

44, 45–64, and ‡ 65 years). Methods. Data from the National Violent Death Reporting System (2003–2008)

were used to calculate age-specific suicide rates for veterans (n = 8440) and

nonveterans (n = 21 668) and to calculate the age-stratified mortality ratio for

veterans. Multiple logistic regression was used to compare health status,

stressful life events preceding suicide, and means of death among young,

middle-aged, and older veterans.

Results. Veterans were at higher risk for suicide compared with nonveterans in

all age groups except the oldest. Mental health, substance abuse, and financial

and relationship problems were more common in younger than in older veteran

suicide decedents, whereas health problems were more prevalent in the older

veterans. Most male veterans used firearms for suicide, and nearly all elderly

veterans did so.

Conclusions. Our study highlighted heightened risk of suicide in male veterans

compared with nonveterans. Within the veteran population, suicide might be

influenced by different precipitating factors at various stages of life. (Am J Public

Health. 2012;102:S131–S137. doi:10.2105/AJPH.2011.300445)

RESEARCH AND PRACTICE

Supplement 1, 2012, Vol 102, No. S1 | American Journal of Public Health Kaplan et al. | Peer Reviewed | Research and Practice | S131

conflicting evidence regarding the role of age in suicide among veterans because of different sampling methodologies and study designs (e.g., veterans of different eras, only VA health care users, and different follow-up periods) as well as possible misclassification (e.g., of persons on active duty or in reserve forces). An under- standing of the precipitating circumstances asso- ciated with suicide can help clarify the situation and could lead to more efficacious clinical and community-based veteran suicide prevention in- terventions.

Our study was the first to examine suicide risk and precipitating circumstances among male veterans across different age groups in the general population. To address these aims, we used population-based data (1) to calculate the standardized mortality ratios (SMRs) of dying by suicide among male veterans by age group, and (2) to assess and compare suicide means as well as health status and stressful life events preceding suicide deaths among young, mid- dle-aged, and older veterans. Our goal was to examine the precipitating suicide circum- stances across age groups to determine varia- tions across the life span. Information about age-associated precipitating circumstances might have important implications for devel- oping national suicide prevention strategies.

METHODS

This study used restricted data for decedents aged 18 years and older from the 2003 to 2008 National Violent Death Reporting Sys- tem (NVDRS). The NVDRS is a state-based active surveillance system that provides a de- tailed account of violent deaths that occur in the participating states. As of 2008, 16 states (Alaska, Colorado, Georgia, Kentucky, Mary- land, Massachusetts, New Jersey, New Mexico, North Carolina, Oklahoma, Oregon, South Carolina, Rhode Island, Utah, Virginia, and Wisconsin) contributed data to the NVDRS.

The NVDRS includes all suicides, homicides, legal intervention deaths, unintentional firearm deaths, and undetermined deaths. The data were gathered from coroner or medical exam- iner records; police reports; death certificates; toxicology laboratories; crime laboratories; and Alcohol, Tobacco, Firearms and Explosives (ATF) firearm trace reports. The circumstances preceding suicide were derived from death

investigations assembled by the coroner/med- ical examiner and law enforcement reports, based on on-scene investigations; information from next of kin and witnesses; autopsy exam- inations; postmortem toxicological testing; and, in some cases, contact with health care pro- viders.22

Veteran status was ascertained with infor- mation obtained from national standardized death certificates. Veteran status was indicated on death certificates in the section “Ever served in US Armed Forces.”22 This death certificate item does not distinguish between military per- sonnel currently serving on active duty, in the guard or reserves, or veterans who have sepa- rated from the service. In the present analysis, the term veteran denoted decedents who served in the US armed forces.

Pooled NVDRS data yielded 8440 veteran and 21 668 nonveteran (defined as decedents without US military service) male suicide de- cedents. Information on veteran status was missing or unknown for 9% of the suicide decedents. Suicide cases were identified using the International Classification of Diseases-10th Revision23 codes X60-X84, and Y87.

Measures

Age was categorized into 4 commonly used groups: 18 to 34, 35 to 44, 45 to 64, and 65 years and older. Ages 18 to 24 and 25 to 34 years were collapsed into 1 group because there were very few significant differences in the circumstances preceding suicide between these 2 age groups. The variables of interest included mental health status, blood alcohol concentration (BAC), alcohol problems, suicidal behaviors, and life crises or events. The selec- tion of variables of interest was based on evidence from previous studies that these factors were linked with suicidal behavior among veterans.16,17,19---21,24---38

Mental health status. Mental health status was ascertained with 4 separate items (“perceived” depressed mood, diagnosed with a mental health problem, ever treated for a mental health problem, and current treatment of a mental health problem). First, family mem- bers or friends reported if the decedent or others perceived that he was depressed, sad, down, blue, or unhappy shortly before the suicide. Second, they provided information on whether the decedent had had a mental health

diagnosis from his health care provider at the time of death. Third, current and past mental health treatment was assessed from evidence gathered at the scene or from family members or friends. Suicidal behavior. Family members or friends

reported whether the suicide decedent dis- closed his intention to complete suicide or had a history of suicide attempts. Blood alcohol concentration. The suicide de-

cedent’s blood was tested for the presence of alcohol, coded in terms of weight by volume, and then classified as BAC less than 0.08 grams per deciliter or 0.08 grams per deciliter or greater. Alcohol and other drug problems. Family

members or friends reported whether the de- cedent perceived himself or was perceived by others to have an alcohol or substance abuse problem at the time of death. Life events. Family members or friends

reported whether the decedent experienced a crisis within 2 weeks of the suicide or if a crisis was imminent. In addition, they were asked if the decedent experienced financial, physical health, job, intimate partner, or crim- inal or legal problems. Suicide methods. The coroner/medical ex-

aminer determined the method used to com- plete suicide. Suicide methods were dichoto- mized into firearm versus all other methods (i.e., a sharp or blunt instrument, poisoning, hanging, fall, drowning, fire or burns, a motor vehicle, or other). Demographic characteristics. Data included

race (White or non-White), marital status (married or not married), region of residence (Northeast, Midwest, South, or West), and metropolitan status. Counties of death were assigned a rural---urban continuum code based on a classification system developed by the US Department of Agriculture Economic Research Service.39 The continuum contains 9 categories and characterizes metropolitan counties by pop- ulation size and nonmetropolitan counties by level of urbanization and adjacency to metro- politan areas. The categories range from “1” (counties in metropolitan areas of 1 million population or more) to “9” (completely rural counties or those with less than 2500 urban population, not adjacent to a metropolitan area). The 9 categories were then collapsed into 2 categories: metropolitan (codes 1 through 3) and nonmetropolitan (codes 4 through 9) status.

RESEARCH AND PRACTICE

S132 | Research and Practice | Peer Reviewed | Kaplan et al. American Journal of Public Health | Supplement 1, 2012, Vol 102, No. S1

Analysis

To assess the risk of dying by suicide, we calculated SMRs and tested for their statistical significance. The SMR is the ratio of the observed number of veteran suicides to the expected number of suicides based on non- veteran suicide rates adjusted for race and calendar year. Veteran and nonveteran suicide rates were calculated with the number of decedents obtained from the NVDRS. Popula- tion estimates for those aged 18 and older who resided in the NVDRS states were based on VA veteran population estimates40 and the Amer- ican Community Survey (ACS) 2003 to 2008.41

The ACS participants were asked “has this person ever served on active duty in the U.S. Armed Forces, military Reserves, or National Guard?” Those who responded “no, never served in the military” were considered non- veterans, and those who responded “yes, now on active duty” or “yes now training for Reserves/ National Guard” were classified as active duty military personnel.42 Veteran population esti- mates included former and current military personnel (i.e., veterans, active duty personnel, National Guard, and Reserves). Veteran suicide rates were computed using population estimates that included current (derived from ACS data) and former (from the VA population data) military personnel. Nonveteran suicide rates were computed using ACS data.

The age-specific demographic characteristics were examined, and the prevalence of the precipitating suicide circumstances among the veteran decedents was estimated. Unadjusted age differences were assessed using logistic regression, with the youngest age group (18---34 years) used as the reference category. Next, a logistic regression was performed for each circumstance while adjusting for the effects of race, marital status, region of resi- dence, and place of residence (urban---rural). Again, the youngest age group (18---34 years) was used as the reference category. All analyses were performed using the Statistical Package for the Social Sciences (SPSS; version 19.0; SPSS, Chicago, Illinois).

RESULTS

The risk of suicide among male veterans compared with nonveterans was statistically significantly greater in all age groups, except for

those aged 65 years and older. Specifically, the SMRs for veterans aged 18 to 34, 35 to 44, 45 to 64, and 65 years and older were 1.26 (P < .05), 1.12 (P < .05), 1.04 (P = .05), and 0.98 (P > .05), respectively.

Table 1 presents the demographic charac- teristics of the veteran suicide decedents for all ages and in each age group. The findings showed that older veteran decedents were more likely to be White, married, and to have died in a rural area than were their younger counterparts.

Table 2 also shows that the antecedents of suicide appeared to be different across age groups. Unadjusted results had a similar pat- tern of findings as in the adjusted model for most of the circumstances and age groups.

In the adjusted model, compared with older veteran decedents, younger decedents were more likely to have had intimate partner problems as well as financial, legal, and occu- pational difficulties. Nearly 1 of every 2 youn- ger veteran suicide decedents (aged 18---34 years) experienced relationship problems shortly before death. By contrast, older veteran decedents were more likely to have had health problems. It was interesting that the middle- aged veterans (aged 35---44) were more likely to have received a mental health diagnosis, whereas the older veterans (aged ‡ 65 years) were more likely to be suspected of being depressed at the time of death. For all age groups, depression was the most common di- agnosis (> 70%). Posttraumatic stress disorder (PTSD) was reported in 7% of the younger veterans. Notable differences were also evident

in relation to alcohol dependence (i.e., 14.9% and 6.2% among those aged 18---34 and ‡ 65 years vs 23.8% and 24.3% among those aged 35---44 and 45---64 years, respectively). Evi- dence of acute alcohol use (BAC ‡ 0.08) was present at the time of death in about one third of the younger veterans but in less than 10% of the elderly veterans.

The results also showed age differences in suicidal behaviors. Older and middle-aged men were less likely to have had a history of suicide attempts than were their youngest counter- parts. Another notable difference was evident in relation to suicide methods. Firearm-related suicides were far more common among the older veterans (83.0%). Other variations across the 4 age groups were minor or did not form a clear or coherent pattern.

DISCUSSION

The risks of suicide (SMRs) among veterans were higher than among nonveterans across all age groups, except for the oldest veterans. Our estimates of suicide risks were consistent with those of other studies that found an increased risk of suicide among veterans,15---18

but disagreed with the findings of Kang and Bullman6 and Miller et al.18 The present findings were based on current data, accounted for active duty or reserve military personnel, and might reflect recent changes in veterans’ suicide rates. In addition, the large number of suicides in the present study provided substantial statistical power that might not have been available in earlier research.

TABLE 1—Demographic and Geographic Characteristics of Young, Middle-Aged, and Older

Male Veteran Suicide Decedents: National Violent Death Reporting System, 2003–2008

Aged 18–34

Years

(n = 845), %

Aged 35–44

Years

(n = 1035), %

Aged 45–64

Years

(n = 3266), %

Aged ‡ 65 Years

(n = 3294), % All Ages

White (vs non-White) 80.2 86.9 93.6 97.3 92.3

Married (vs not married) 34.4 47.3 48.6 51.1 48.0

Region of residence

Northeast 5.9 8.2 9.9 11.3 9.8

Midwest 5.1 9.7 8.1 6.4 7.3

South 62.5 54.8 52.5 55.0 54.7

West 26.5 27.3 29.6 27.4 28.1

Nonmetropolitan area (vs metropolitan) 19.6 19.4 25.2 27.2 24.7

RESEARCH AND PRACTICE

Supplement 1, 2012, Vol 102, No. S1 | American Journal of Public Health Kaplan et al. | Peer Reviewed | Research and Practice | S133

Our results also demonstrated that the cir- cumstances preceding suicide varied by age group. Although most veterans used firearms, older men were more likely to complete suicide with a firearm. This finding was consistent with that of Kaplan et al.,20 who showed that older veterans were more likely to own and use a firearm than were younger veterans and non- veteran elderly men. This finding suggested that it is worthwhile for health care providers to probe for access to firearms among all depressed or suicidal older patients. In recognition of the problem of the availability and use of firearms, the VA adopted guidelines for VA providers to address the importance of means restriction in managing suicidal veterans.43 These guidelines specify that rather than removing firearms, which might lead to conflict, a responsible person (e.g., family member) should safely store the weapons. The VA also initiated a Veteran Family Safety Program that provides free cable gunlocks at all VA Medical Centers for veterans and their family

members. In the United States, many veterans are eligible for health care through the VA system. However, only a minority of veterans use VA programs,44 although a higher percentage of veterans of Operation Enduring Freedom/ Operation Iraqi Freedom (approximately 50%) have accessed services because of broader eligi- bility criteria. Thus, to address the needs of the broader veteran population, non-VA providers should be following these guidelines and might wish to address gun safety and storage issues.

Reports of relationship problems among younger veteran suicide decedents under- scored the need for more extensive resources to assist veterans in coping with family con- cerns.7 Those who work with veterans should consider moving beyond the standard “danger signs” of suicide risk (e.g., major depressive disorder or PTSD) and address the role that life crises play in triggering suicidal behavior. Therefore, interventions that focus on develop- ing practical interpersonal skills could avert

a suicidal crisis. Furthermore, Kapur et al.21

recently observed that the main strategy used in the United Kingdom to date included the encouragement of appropriate help-seeking be- havior once individuals left the armed forces. In the United Kingdom, all citizens, including vet- erans, are served through the National Health Service.

Many young veteran decedents had elevated BACs at the time of death. Nearly one third of the youngest veteran suicide decedents had a BAC ‡ 0.08. It is well accepted that acute alcohol use is associated with suicidal behavior, and some consider this relationship to be causal.45 Other indicators of substance abuse were also elevated in the young and middle-aged veteran suicide decedents compared with the older veteran suicide decedents. For example, according to the NVDRS, evidence of alcohol dependence appeared more prevalent in the middle-aged than in the youngest and oldest groups. Earlier studies suggested that Vietnam-era

TABLE 2—Suicide Risk and Precipitating Circumstances Among Young, Middle-Aged, and Older Male Veteran

Suicide Decedents: National Violent Death Reporting System, 2003–2008

Aged 35–44 Years Aged 45–64 Years Aged ‡ 65 Years

Aged 18–34 Years, % % AOR (95% CI) % AOR (95% CI) % AOR (95% CI) All Ages

Mental health

Ever treated for a mental health problem 29.8 36.3 1.25* (1.01, 1.53) 35.2 1.19 (1.00, 1.41) 25.1 0.72*** (0.60, 0.86) 30.9

Diagnosed with a mental health problem 33.7 41.1 1.29* (1.06, 1.58) 41.4 1.28** (1.08, 1.52) 30.6 0.79** (0.67, 0.94) 36.4

Current treatment of a mental health problem 24.4 28.8 1.15 (0.92, 1.43) 30.4 1.24* (1.03, 1.50) 21.9 0.79* (0.66, 0.96) 26.3

Current depressed mood 36.1 44.0 1.30** (1.07, 1.59) 45.2 1.33** (1.13, 1.58) 48.1 1.52*** (1.29, 1.80) 45.3

Suicidal behaviors

Disclosed intent to complete suicide 25.2 28.9 1.13 (0.90, 1.40) 26.3 0.98 (0.81, 1.18) 28.1 1.10 (0.91, 1.32) 27.2

Previous suicide attempts 16.6 19.8 1.12 (0.87, 1.44) 14.6 0.79* (0.63, 0.98) 6.9 0.33*** (0.26, 0.42) 12.4

Left a suicide note 29.5 32.6 1.17 (0.95, 1.45) 36.3 1.38*** (1.16, 1.65) 32.6 1.19 (0.99, 1.42) 33.7

Firearm suicide 61.3 53.4 0.78* (0.64, 0.95) 63.7 1.16 (0.98, 1.37) 83.5 3.51*** (2.92, 4.21) 69.9

Substance use

Substance problem other than alcohol 14.7 14.4 0.99 (0.75, 1.31) 9.8 0.67** (0.52, 0.85) 0.9 0.05*** (0.03, 0.08) 7.4

Alcohol dependence 15.7 24.0 1.61*** (1.26, 2.06) 24.5 1.61*** (1.30, 2.00) 6.3 0.33*** (0.26, 0.42) 16.4

BAC ‡ 0.08 32.5 33.0 0.96 (0.75, 1.23) 27.9 0.75** (0.61 0.93) 8.8 0.17*** (0.13, 0.22) 22.5 Life events

Crisis 38.3 35.8 0.88 (0.72, 1.07) 27.9 0.60*** (0.51, 0.71) 23.6 0.48*** (0.41, 0.58) 28.2

Financial problem 11.2 16.7 1.53** (1.16, 2.03) 16.7 1.48** (1.16, 1.88) 5.0 0.39*** (0.29, 0.52) 11.6

Health problem 5.3 11.7 2.41*** (1.66, 3.51) 26.8 6.43*** (4.62, 8.95) 66.7 36.09*** (25.92, 50.25) 38.4

Job problem 15.4 18.8 1.25 (0.97, 1.62) 15.8 0.97 (0.78, 1.22) 2.0 0.11*** (0.08, 0.15) 10.8

Criminal problem 13.5 17.4 1.38* (1.05, 1.81) 11.1 0.86 (0.67, 1.10) 1.9 0.14*** (0.10, 0.20) 8.5

Intimate partner problem 49.7 47.3 0.77* (0.63, 0.94) 25.8 0.27*** (0.23, 0.32) 8.6 0.07*** (0.06, 0.09) 24.1

Note: AOR = adjusted odds ratio; BAC = blood alcohol concentration; CI = confidence interval. Odds ratios were adjusted for race, marital status, metropolitan level, and region of residence. The age group 18–34 years was used as the reference category. *P < .05; **P < .01; ***P < .001.

RESEARCH AND PRACTICE

S134 | Research and Practice | Peer Reviewed | Kaplan et al. American Journal of Public Health | Supplement 1, 2012, Vol 102, No. S1

veterans, now middle aged, were particularly vulnerable to substance abuse and other psy- chiatric conditions.46---49 Specifically, several studies showed a relationship between PTSD and substance abuse problems among Vietnam veterans.50,51 A question of some importance was whether substance abuse prevention and treatment programs that focus on young and middle-aged veterans could reduce suicide in this population. It should be noted that the VA currently has suicide prevention care managers at each facility who coordinate the care of suicidal VA users and work to ensure that evidence-based treatments are accessed and used.

Our study also pointed to the burden of physical health problems among older veterans who died by suicide. Physical health problems were associated with suicide in the general population52---54; however, further analyses of the NVDRS revealed that physical illness was more prevalent among older veterans than among nonveterans (data not shown). Although elderly veterans were more likely than younger veterans to be described as currently having a depressed mood, evidence of mental health treatment was less common among the older than the younger veterans. Consistent with ear- lier findings,21 middle-aged and older veterans were less likely to have received treatment for mental health problems compared with nonvet- erans in the same age groups (data not shown). The lower prevalence of mental health problems reported among older veterans might be related to family members’ lack of knowledge of the decedent’s mental health status and treatment, or could be related to less contact with family or reluctance to divulge information on a diagnosis or treatment because of stigma. Suicide preven- tion programs for older veterans within the medical system might be more likely to be successful if they focus on primary (physical) health care rather than exclusively on mental health care.

This study made an important and unique contribution to the literature regarding the circumstances preceding suicide among male veterans of different age groups. The findings of our study added to knowledge of the role that life stressors play in suicide, including relationship and intimate partner problems among younger veterans, substance abuse among middle-aged veterans, and physical

illness among older veterans. In addition, our results also showed that the use of firearms was the most common method of suicide among veterans of all ages (especially among older veterans). Of note, the rate and likelihood of firearm use were higher among veterans than among nonveterans in all age groups.20

The present study had some potential limi- tations. First, questions were raised about the accuracy of the designation of veteran status on death certificates. However, data from the National Mortality Followback Survey (NMFS) showed a high rate of agreement between death certificates and proxy-derived informa- tion regarding the decedent’s military service (j = 0.91; data not shown).55 Second, veteran status on death certificates did not distinguish among service members who were on active duty, those who were in the National Guard or Reserves, and those who had been discharged. Further examination of the NMFS revealed that among those classified as veterans on the death certificates, only 10% of those aged 18 to 34 years, 5% of those aged 35 to 44 years, 0.7% of those aged 45 to 64 years, and 0.6% of those aged 65 years and older were on active duty or in the Reserves at the time of death. Of note, although the NMFS data suggested that veteran status on death certificates was valid, the data were gathered in 1993 and might or might not be relevant today. Third, the Blue Ribbon Work Group on Suicide Prevention in the Veteran Population19 observed that veterans’ deaths might be more accurately classified as suicides (and correspondingly less likely to be considered “undetermined” deaths) because there was gen- erally more information available about veteran deaths than about nonveteran deaths. For ex- ample, veterans were likely to have health in- surance, family support, and social support in the community (e.g., VA and veterans’ organizations) and, thus, more likely than nonveterans to have collateral data. However, a supplementary anal- ysis of the NVDRS revealed that the rates of “undetermined” deaths across age groups were nearly identical for veteran and nonveteran de- cedents. Fourth, not all coroners or medical examiners and law enforcement personnel rou- tinely collected mental health, substance abuse, toxicology, and firearm information; thus, the prevalence of important circumstances associ- ated with suicide cases might have been under- estimated.56 The collection of postmortem data

was particularly difficult because some states have independent county coroner systems rather than a centralized medical examiner system.57

However, 63% of the states participating in the NVDRS have a centralized state medical exam- iner system compared with only 15% of non- participating states.58 Fifth, as noted earlier, many of the suicide circumstances were derived from proxy information and might be unreli- able.56,59 However, several studies using psy- chological autopsy data showed that proxy re- sponses were valid.60---63 Finally, this study was descriptive, and no causal inferences could be made.

Despite these limitations, data from the NVDRS had numerous strengths. First, the NVDRS is the only surveillance system for violent deaths in the United States. Second, the NVDRS collects information from multiple sources to characterize violent deaths as op- posed to using only death certificates (which contain limited data). Third, the data on nearly all suicide deaths include precipitating circum- stances. Finally, although the NVDRS states are not necessarily representative of the nation, the populations of these 16 states are similar in terms of gender, age, ethnic/racial composition, urban/rural characteristics, and overall suicide rates to the nation as a whole.

The evidence presented here indicated that the risk of dying by suicide was significantly higher among nonelderly male veterans than among male nonveterans. Although the results of our analysis showed that the relative risk of suicide (vs nonveterans) was generally higher for younger compared with older veterans, preventive interventions should reach a broad population, including different age groups and those who do not use VA health care services.64 Our findings have important impli- cations for the development and implementation of preventive interventions. Primary care and behavioral health providers both inside and outside the VA system need to be attentive to age-specific circumstances (in addition to signs of suicidal intent and access to firearms) that may lead to suicide. The use of highly lethal suicide methods, such as firearms, reduces the opportu- nity to intervene and rescue, and firearm access among high-risk populations needs to be addressed before crises develop.65 Addressing the diversity of precipitating suicide circum- stances across age groups may both advance

RESEARCH AND PRACTICE

Supplement 1, 2012, Vol 102, No. S1 | American Journal of Public Health Kaplan et al. | Peer Reviewed | Research and Practice | S135

causal theories related to suicide risk and result in more efficacious preventive interventions. j

About the Authors Mark S. Kaplan and Nathalie Huguet are with Portland State University, Portland, OR. Bentson H. McFarland is with Oregon Health and Science University, Portland. Marcia Valenstein is with the Department of Veterans Affairs Ann Arbor Health System and the University of Michigan, Ann Arbor. Correspondence should be sent to Mark S. Kaplan,

DrPH, Professor of Community Health, Portland State University, P.O. Box 751, Portland, OR 97207 (e-mail: [email protected]). Reprints can be ordered at http://www. ajph.org by clicking the “Reprints/Eprints” link. This article was accepted August 30, 2011.

Contributors M. S. Kaplan, B. H. McFarland, and N. Huguet designed and conceptualized the study, analyzed and interpreted the data, and wrote the article. M. S. Kaplan and N. Huguet acquired the data. N. Huguet performed the statistical analysis. B. H. McFarland supervised statistical analysis. M. Valenstein contributed to the drafts and critical revisions of the article. All the authors approved the final draft.

Acknowledgments This research was supported by the American Founda- tion for Suicide Prevention. It used data from NVDRS, a surveillance system designed by the Centers for Disease Control and Prevention’s (CDC) National Center for Injury Prevention and Control. The findings are based, in part, on the contributions of the 16 funded states that collected violent death data and the contributions of the states’ partners, including personnel from law enforce- ment, vital records, medical examiners/coroners, and crime laboratories.

None of the authors has any interests that might be interpreted as influencing the research.

Note. The American Foundation for Suicide Preven- tion and the CDC had no role in the design and conduct of the study; in the analysis and interpretation of the data; or in the preparation, review, or approval of the article.

Human Participant Protection The Human Subject Review Committee at Portland State University reviewed and approved the project.

References 1. Goldsmith SK, Pellmar TC, Kleinman AM, Bunney WE. Reducing Suicide: A National Imperative. Washing- ton, DC: National Academies Press; 2002.

2. Nock MK, Borges G, Bromet EJ, Cha CB, Kessler RC, Lee S. Suicide and suicidal behavior. Epidemiol Rev. 2008;30(1):133---154.

3. Helmkamp JC. Suicides in the military: 1980-1992. Mil Med. 1995;160:45---50.

4. Shaffer D. Suicide and Suicide Prevention in the Military’ Forces; Report of a Consultation. Washington, DC: Office of the Secretary of Defense; 1998.

5. Allen JP, Cross G, Swanner J. Suicide in the Army: a review of current information. Mil Med. 2005;170 (7):580---584.

6. Kang HK, Bullman TA. Is there an epidemic of suicides among current and former U.S. military person- nel? Ann Epidemiol. 2009;19(10):757---760.

7. Institute of Medicine (IOM). Returning Home from Iraq and Afghanistan: Preliminary Assessment of Read- justment Needs of Veterans, Service Members, and Their Families. Washington, DC: The National Academies Press; 2010.

8. Kuehn BM. Soldier suicide rates continue to rise: military, scientists work to stem the tide. JAMA. 2009; 301(11):1111---1113.

9. Department of Veterans Affairs Office of Inspector General. Implementing VHA’s Mental Health Strategic Plan Initiatives for Suicide Prevention. Washington, DC: Department of Veterans Affairs; 2007.

10. Hampton T. Research, law address veterans’ suicide. JAMA. 2007;298(23):2732.

11. House Committee on Veterans’ Affairs Hearing. The Truth About Veterans’ Suicide, May 6 2008. Washing- ton, DC: House of Committee on Veteran’s Affairs. Available at: http://veterans.house.gov/hearings. Ac- cessed April 20, 2009.

12. Senate Committee on Veterans’ Affairs Hearing. Mental Health Care and Suicide Prevention for Veterans, March 3 2010, Washington, DC: Senate Committee on Veterans’ Affairs. Available at: http://veterans.senate. gov/hearings.cfm. Accessed April 20, 2009.

13. Peake JB. (Secretary of Veteran Affairs). The Truth about Veterans’ Suicides before the House Committee on Veterans Affairs, June 5, 2008. Washington, DC: House of Committee on Veteran’s Affairs. Available at: http:// www.gpo.gov/fdsys/pkg/CHRG-110hhrg43052/html/ CHRG-110hhrg43052.htm. Accessed January 4, 2012.

14. Maynard C, Boyko EJ. Datapoints: suicide rates in the Washington State veteran population. Psychiatr Serv. 2008;59:1245.

15. Shen X, Millet L, Kohn M. Violent deaths in Oregon: 2005: Oregon Department of Human Services, Portland, Oregon; 2007. Available at: http://www.oregon.gov/ DHS/ph/ipe/nvdrs. Accessed April 20, 2009.

16. McCarthy JF, Valenstein M, Kim HM, Ilgen M, Zivin K, Blow FC. Suicide mortality among patients receiving care in the Veterans Health Administration Health System. Am J Epidemiol. 2009;169(8):1033---1038.

17. Kaplan MS, Huguet N, McFarland BH, Newsom JT. Suicide among male veterans: a prospective population- based study. J Epidemiol Community Health. 2007;61 (7):619---624.

18. Miller M, Barber C, Azrael D, Calle EE, Lawler E, Mukamal KJ. Suicide among US veterans: a prospective study of 500,000 middle-aged and elderly men. Am J Epidemiol. 2009;170(4):494---500.

19. United States Department of Veterans Affairs. Re- port of the Blue Ribbon Work Group Report on Suicide Prevention in the Veteran Population, June 30 2008. Available at: http://www.mentalhealth.va.gov/suicide_ prevention/Blue_Ribbon_Report-FINAL_June-30-08. pdf. Accessed January 4, 2012.

20. Kaplan MS, McFarland BH, Huguet N. Firearm suicide among veterans in the general population: find- ings from the National Violent Death Reporting System. J Trauma. 2009;67(3):503---507.

21. Kapur N, While D, Blatchley N, Bray I, Harrison K. Suicide after leaving the UK armed forces---a cohort study. PLoS Med. 2009;6:e26.

22. Centers for Disease Control and Prevention. Na- tional Violent Death Reporting System (NVDRS) Coding Manual Revised [Online], 2008. National Center for Injury Prevention and Control, Centers for Disease Control and Prevention (producer). Available at: http:// www.cdc.gov/ncipc/pub-res/nvdrs-coding/vs2/default. htm. Accessed January 4, 2012.

23. International Classification of Diseases. 10th Revision. Geneva, Switzerland: World Health Organization; 1994.

24. Fontana A, Rosenheck R. An etiological model of attempted suicide among Vietnam theater veterans: prospective generalization to a treatment-seeking sample. J Nerv Ment Dis. 1995;183(6):377---383.

25. Kausch O, McCormick RA. Suicide prevalence in chemical dependency programs: preliminary data from a national sample, and an examination of risk factors. J Subst Abuse Treat. 2002;22(2):97---102.

26. Lehmann L, McCormick RA, McCracken L. Suicidal behavior among patients in the VA health care system. Psychiatr Serv. 1995;46(10):1069---1071.

27. Hartl TL, Rosen C, Drescher K, Lee TT, Gusman F. Predicting high-risk behaviors in veterans with PTSD. J Nerv Ment Dis. 2005;193(7):464---472.

28. Thompson R, Kane V, Cook JM, Greenstein R, Walker P, Woody G. Suicidal ideation in veterans re- ceiving treatment for opiate dependence. J Psychoactive Drugs. 2006;38(2):149---156.

29. Tiet QQ, Ilgen MA, Byrnes HF, Moss RH. Suicide attempts among substance use disorder patients: steps towards a decision tree. Alcohol Clin Exp Res. 2006;30: 998---1005.

30. Zivin K, Kim HM, McCarthy JF, et al. Suicide mortality among individuals receiving treatment for de- pression in the Veterans Affairs health system: associa- tions with patient and treatment setting characteristics. Am J Public Health. 2007;97(12):2193---2198.

31. Jakupcak M, Conybeare D, Phelps L, et al. Anger, hostility and aggression among Iraq and Afghanistan war veterans reporting PTSD and subthreshold PTSD. J Trauma Stress. 2007;20(6):945---954.

32. Freeman TW, Roca V, Moore WM. A Comparison of chronic combat-related posttraumatic stress disorder (PTSD) patients with and without a history of suicide attempt. J Nerv Ment Dis. 2000;188(7):460---463.

33. Lish JD, Zimmerman M, Farber NJ, Kuzma MA, Plescia G. Suicide screening in a primary care setting at a Veteran’s Affairs Medical Center. Psychosomatics. 1996; 37:413---424.

34. Benda BB. Gender differences in predictors of suicidal thoughts and attempts among homeless veterans that abuse substances. Suicide Life Threat Behav. 2005; 35(1):106---116.

35. Rodell DE, Benda BB, Rodell L. Suicidal thoughts among homeless alcohol and other drug abusers. Alcohol Treat Q. 2003;21:57---74.

36. Thoresen S, Mehlum L, Røysamb E, Tønnessen A. Risk factors for completed suicides in veterans of peace- keeping: repatriation, negative life events, and marital status. Arch Suicide Res. 2006;10(4):353---363.

37. MacLean A, Elder GH. Military service in the life course. Annu Rev Sociol. 2007;33:175---196.

38. Desai RA, Dausey DJ, Rosenheck RA. Mental health service delivery and suicide risk: the role of individual patient and facility factors. Am J Psychiatry. 2005;162 (2):311---318.

RESEARCH AND PRACTICE

S136 | Research and Practice | Peer Reviewed | Kaplan et al. American Journal of Public Health | Supplement 1, 2012, Vol 102, No. S1

39. United States Department of Agriculture Economic Research Service. Measuring rurality: Rural-urban con- tinuum codes, 2003. Available at: http://www.ers.usda. gov/briefing/rurality/ruralurbcon. Accessed April 20, 2009.

40. United States Department of Veteran Affairs. Vet- eran Population 2007. Available at: http://www1.va. gov/vetdata. Accessed April 20, 2009.

41. United States Department of Commerce. Bureau of the Census. American Community Survey (ACS): Public Use Microdata Sample (PUMS), 2006. ICPSR22101-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research; 2008.

42. US Census Bureau. 2007 American Community Survey (ACS) Questionnaire. Available at: http://www. census.gov/acs/www/SBasics/SQuest/SQuest1.htm. Accessed April 20, 2009.

43. Stanley B, Brown GK. The Safety Plan Treatment Manual to Reduce Suicide Risk: Veteran Version. Wash- ington, DC: United States Department of Veterans Af- fairs; 2008.

44. National Center for Veterans Analysis and Statistics. (10/27/2008). VA benefits and health care utilization. Available at: http://www.va.gov/VETDATA/Pocket- Card/4X6_spring10_sharepoint.pdf. Accessed January 4, 2012.

45. Cherpitel CJ, Borges GL, Wilcox HC. Acute alcohol use and suicidal behavior: a review of the literature. Alcohol Clin Exp Res. 2004;28(Suppl 5):18S---28S.

46. Boudewyns PA, Woods MG, Hyer L, Albrecht WJ. Chronic combat-related PTSD and concurrent substance abuse: Implications for treatment of this frequent “dual diagnosis.” J Trauma Stress. 1991;4(4):549---560.

47. Brooks MS, Laditka SB, Laditka JN. Long-term effects of military service on mental health among veterans of the Vietnam war era. Mil Med. 2008;173: 570---575.

48. Boehmer TK, Flanders WD, McGeehin MA, Boyle C, Barrett DH. Post-service mortality in Vietnam vet- erans: 30-year follow-up. Arch Intern Med. 2004;164 (17):1908---1916.

49. Goldberg J, Eisen SA, True WR, Henderson WG. A twin study of the effects of the Vietnam conflict on alcohol drinking patterns. Am J Public Health. 1990;80 (5):570---574.

50. Boscarino J. Current excessive drinking among Vietnam veterans: a comparison with other veterans and nonveterans. Int J Soc Psychiatry. 1981;27(3):204---212.

51. McFall M, Fontana A, Raskind M, Rosenheck R. Analysis of violent behavior in Vietnam combat veteran psychiatric inpatients with posttraumatic stress disorder. J Trauma Stress. 1999;12(3):501---517.

52. Kaplan MS, McFarland BH, Huguet N, Newsom JT. Physical illness, functional limitations, and suicide risk: a population-based study. Am J Orthopsychiatry. 2007; 77:56---60.

53. Conwell Y, Lyness JM, Duberstein P, et al. Com- pleted suicide among older patients in primary care practices: a controlled study. J Am Geriatr Soc. 2000; 48:23---29.

54. Conwell Y, Duberstein PR, Hirsch JK, et al. Health status and suicide in the second half of life. Int J Geriatr Psychiatry. 2010;25:371---379.

55. United States Department of Health and Human Services, National Center for Health Statistics. National Mortality Followback Survey, 1993. ICPSR version.

Hyattsville, MD: U.S. Department of Health and Human Service; 2005.

56. Logan J, Hill HA, Black ML, et al. Characteristics of perpetrators in homicide-followed-by-suicide incidents: National Violent Death Reporting System---17 US States, 2003-2005. Am J Epidemiol. 2008;168(9):1056---1064.

57. Karch DL, Dahlberg LL, Patel N. Surveillance for violent deaths---National Violent Death Reporting System, 16 States, 2007. MMWR Surveill Summ. 2010;59:1---50.

58. National Research Council. Strengthening Forensic Science in the United States: A Path Forward. Washington, DC: The National Academies Press, 2009.

59. Bossarte RM, Simon TR, Barker L. Characteristics of homicide followed by suicide incidents in multiple states, 2003-04. Inj Prev. 2006;12(Suppl 2):ii33---ii38.

60. Conner KR, Conwell Y, Duberstein PR. The validity of proxy-based data in suicide research: a study of patients 50 years of age and older who attempted suicide. II. Life events, social support and suicidal behavior. Acta Psychiatr Scand. 2001;104(6):452---457.

61. Conner KR, Duberstein PR, Conwell Y. The validity of proxy-based data in suicide research: a study of patients 50 years of age and older who attempted suicide. I. Psychiatric diagnoses. Acta Psychiatr Scand. 2001;104 (3):204---209.

62. Brent DA, Perper JA, Moritz G, et al. The validity of diagnoses obtained through the psychological autopsy procedure in adolescent suicide victims: use of family history. Acta Psychiatr Scand. 1993;87(2):118---122.

63. Kelly TM, Mann JJ. Validity of DSM-III-R diagnosis by psychological autopsy: a comparison with clinician ante-mortem diagnosis. Acta Psychiatr Scand. 1996;94 (5):337---343.

64. Bruce ML. Suicide risk and prevention in veteran populations. Ann N Y Acad Sci. 2010;1208:98---103.

65. Mann JJ, Currier D. Prevention of suicide. Psychiatr Ann. 2007;37(5):331---339.

RESEARCH AND PRACTICE

Supplement 1, 2012, Vol 102, No. S1 | American Journal of Public Health Kaplan et al. | Peer Reviewed | Research and Practice | S137