Characteristics of Suicides Among US Army Active Duty Personnel in 17 US States From 2005 to 2007

Abstract

Suicides are increasing among active duty US Army soldiers. To help focus prevention strategies, we characterized 56 US Army suicides that occurred from 2005 to 2007 in 17 US states using 2 large-scale surveillance systems. We found that intimate partner problems and military-related stress, particularly job stress, were common among decedents. Many decedents were also identified as having suicidal ideation, a sad or depressed mood, or a recent crisis before death. Focusing efforts to prevent these forms of stress might reduce suicides among soldiers.

In recent years, the suicide rate has increased among US Army active duty personnel.^1,2 The estimated suicide rate for this population nearly doubled from 2004 to 2008 (from 10.8 to 20.2 per 100 000).^1,2

Mental health conditions, substance abuse problems, certain physical health problems (e.g., cancer, chronic pain), and financial, legal, and relationship problems are risk factors for suicide among civilian and military populations.^3–14 Building on this research, we used data from 2 large-scale surveillance systems to assess the frequency of these factors as well as other military-related stresses (e.g., recent combat exposure, job problems, disciplinary proceedings) among suicide decedents who were on active duty in the US Army and residing in the United States to determine the most prevalent circumstances preceding suicide for this population. A better understanding of the most common preceding circumstances among the number of known risk factors for suicide might help focus military suicide prevention initiatives.^15,16

METHODS

We obtained data for active duty US Army suicide decedents who died during 2005 to 2007. Two large-scale surveillance systems, the National Violent Death Reporting System (NVDRS) and the Department of Defense Suicide Event Report (DoDSER) were linked to comprehensively characterize the decedents.

NVDRS uses coroner or medical examiner and toxicology reports, law enforcement records, and death certificates to provide details on suicides, such as decedent demographic information, mechanism or weapon information, and preceding health and stressful life-event circumstance information.¹⁷ During the study, NVDRS collected data from 17 US states; therefore, case inclusion for this study was limited to suicide cases that had death certificates filed in one of those states. The 17 NVDRS states were Alaska, California, Colorado, Georgia, Kentucky, Maryland, Massachusetts, New Jersey, New Mexico, North Carolina, Oklahoma, Oregon, Rhode Island, South Carolina, Utah, Virginia, and Wisconsin. Data collected in NVDRS was statewide with the exception of California, which collected data in only 5 counties (Los Angeles, Riverside, San Francisco, Alameda, and Santa Clara). NVDRS has been described in detail elsewhere.^17,18

DoDSER data provide details on suicides for all active duty service members. The DoDSER is part of a suicide surveillance program that standardizes retrospective suicide surveillance efforts across the US service branches. The DoDSER was first launched in 2008; however, the 2005–2007 data we used were collected from the Army Suicide Event Report (ASER), the predecessor to the DoDSER that collected data only for US Army suicide deaths. The DoDSER’s web-based data collection process was modeled after the ASER and was developed by the National Center for Telehealth and Technology in collaboration with the Department of Defense’s Suicide Prevention and Risk Reduction Committee’s Suicide Prevention Program Managers representing all of the service branches. This system provides many details unavailable in NVDRS, such as the decedent’s military background, family history, health service utilization, combat exposure, and military disciplinary history.¹⁹ DoDSER data also provide details on circumstances preceding suicide. DoDSERs are completed by behavioral health providers within 60 days of the suicide; each report requires information from medical, mental health, and personnel records, any other relevant documents, and interviews with the decedent’s coworkers, supervisors, friends, family members, and any other acquaintances or involved law enforcement and health professionals, as appropriate.¹⁹

The decedent population was initially identified in the DoDSER database. Cases were linked to NVDRS using incident variables (i.e., state and date of death) and decedent demographic variables (i.e., age, gender, race/ethnicity, marital status, veteran status, and occupation). No personal identifying information was linked into the final dataset. Fifty-nine decedents were identified in the DoDSER, and 56 (90%) cases were linked and thereby included in the study.

RESULTS

Table 1 shows that most decedents were males; of White, non-Hispanic race/ethnicity; less than 30 years old; married; and in the enlisted ranks. Approximately 46% of the decedents had children, and half of these decedents had children residing with them. Sixty-one percent of the decedents died in their personal residences, and 55% used a firearm in the incident. Over a third of the decedents left evidence suggesting their suicide was planned, and 21% left suicide notes, which also suggests premeditation.

TABLE 1—

Demographic and Other Background Characteristics of Active Duty US Army Suicide Decedents and Characteristics of the Suicide Events: 2005–2007

Variable	No. (%)
Age, y
18–24	21 (37.5)
25–29	14 (25.0)
30–39	13 (23.2)
40–49	6 (10.7)
50–59	…
Gender
Male	53 (94.6)
Female	3 (5.4)
Race/ethnicity
White, non-Hispanic	42 (75.0)
Black, non-Hispanic	9 (16.1)
American Indian/Alaskan Native	…
Asian/Pacific Islander	…
Hispanic	…
Marital status
Married	34 (60.7)
Never married	16 (28.6)
Widowed, divorced, or separated	4 (7.1)
Single unspecified or unknown	…
Household and parental factors
Decedent resided alone	17 (30.4)
Decedent had children	26 (46.4)
Children resided with decedent (% is calculated among decedents who had children)	13 (50.0)
US Army status
Regular	43 (76.8)
Reserve	4 (7.1)
National Guard	8 (14.3)
Other	…
Duty status
Active duty	45 (80.4)
Active Guard/Reserve	3 (5.4)
Active duty for training duty	4 (7.1)
Other	4 (7.1)
Pay grade
E1–E2	6 (10.7)
E3	8 (14.3)
E4	16 (28.6)
E5	8 (14.3)
E6	6 (10.7)
E7	4 (7.1)
E8–E9	3 (5.4)
W1–5	…
O1–10	4 (7.1)
Location of death
Personal residence	34 (60.7)
Residence of family or friend	6 (10.7)
Automobile (away from residence)	5 (8.9)
Other	11 (19.6)
Weapon/mechanism used
Firearm	31 (55.4)
Poisoning	7 (12.5)
Hanging, strangulation	15 (26.8)
Other or unknown	3 (5.4)
Other event characteristicsa
Left evidence suggesting the event was planned	20 (35.7)
Left a suicide note	12 (21.4)

Note. E = enlisted ranks; O = commissioned officer ranks; W = warrant officer ranks. The Sample size was n = 56. Ellipses indicate that items with less than 3 counts (< 5% of the population) were suppressed to prevent potential identification of decedents.

^aCategories are not mutually exclusive.

Table 2 shows the prevalence of various health and stress-related circumstances preceding death. The most common circumstances were intimate partner problems (45%) and military-related stress (41%); current job problems and combat experiences were the most common military specific circumstances. Many decedents showed symptoms of mental health distress (e.g., 36% communicated their intent to self harm, 32% were identified as having a depressed mood) or had a recent crisis (32%). Twenty-three percent of the decedents received a mental health diagnosis. Alcohol was involved in over a quarter of the incidents. Alcohol and substance abuse, physical health, criminal and civil legal, and financial problems were also evident among some decedents.

TABLE 2—

Health Related Characteristics of Active Duty US Army Suicide Decedents and Stressful Life-Event Circumstances Preceding Death: 2005–2007

Variable	No. (%)
Health–related factors
Suicidal ideation—disclosed intent of self harma	20 (35.7)
Current depressed mood	18 (32.1)
Substance use at time of incident
Alcohol useb	16 (28.6)
Drug usec (% is calculated among those tested)	8 (30.8)
Current mental health problem	13 (23.2)
Diagnosesd (% is calculated among those who had current mental health problems)
Depression/dysthymia	9 (69.2)
Posttraumatic stress disorder	3 (23.1)
Other	3 (23.1)
Unknown	…
Alcohol or other substance abuse problems	7 (12.5)
Current physical health probleme	4 (7.1)
Stressful life event factorsf
Recent intimate partner problems	25 (44.6)
Any military-related stressful circumstancesg	23 (41.1)
Military specific circumstancesh (% is calculated among those who had any military-related stress)
Current job-related problems	14 (60.9)
Experienced combat in last deployment	9 (39.1)
Subject to administrative separation	5 (21.7)
Subject to AWOL proceedings	4 (17.4)
Subject to medical evaluation board	3 (13.0)
Subject to courts martial proceedings	…
Recent crisis (within 2 wk of death)	18 (32.1)
Subject to other civil criminal or Article 15 proceedings	8 (14.3)
Recent civil legal issues	8 (14.3)
Recently perpetrated interpersonal violence	5 (8.9)
Recent financial problems	4 (7.1)
Recent other relationship problems	…
Recently was a victim of interpersonal violence	…

Note. AWOL = absent-without-leave. The sample size was n = 56. Ellipses indicate that items with less than 3 counts (< 5% of the population) were suppressed to prevent potential identification of decedents.

^aThe decedent communicated potential for self harm. This variable excluded leaving a suicide note.

^bAlcohol use was determined by toxicologic tests and whether there was evidence of use (e.g., witnesses or investigative reports that state the victim was seen drinking before the incident).

^cDrug use was determined by toxicologic tests alone. Toxicologic tests were conducted for 26 decedents.

^dDiagnostic categories are not mutually exclusive.

^ePhysical health problems refer to only medical problems perceived to have precipitated the suicide.

^fStressful life-event factors are not mutually exclusive.

^gThe variable “any military-related stressful circumstances” did not include those undergoing Article 15 proceedings as a circumstance because these decedents could have been actually undergoing civilian criminal proceedings based on how information for this variable is collected. However, over 90% of those who were identified as undergoing Article 15 proceedings also had other military-related stresses before death that were included in this variable; therefore, the proportion of decedents identified as having any military-related stressful circumstances was not largely underestimated.

^hCategories are not mutually exclusive.

DISCUSSION

The range in rank distribution among suicide decedents reflected the general Army population,²⁰ and the decedent demographics were similar to other, mostly civilian, suicide populations.^21–23 The finding that intimate partner problems were the most common circumstances preceding suicide death was also similar to findings in other studies that described characteristics of a male suicide decedent population^21–23; however, the proportion of decedents with preceding job problems was concerning. Among mostly male civilian suicide decedent populations during the same data years, only 12.3% to 12.6% of decedents had reported job problems.^21–23 Similar to other suicide populations, many decedents were identified as sad, depressed, or suicidal by family members, friends, coworkers, clinicians, and other acquaintances, which further shows the difficulty of knowing when and how to help someone in need of mental health services.^21–23

These findings must be viewed cautiously. These results neither provided national representation nor representation of suicides that occurred overseas. Causality could not be inferred between the circumstances and the suicides, and all circumstance information was gleaned from previous reports and interviews, which might not reflect all information known about the incidents. Mental and medical health information was obtained from sources for NVDRS, which include coroner or medical examiners, family members, and friends of the victims. These informants might not have known all of the decedents’ health information; therefore, some health conditions might have been underestimated.

This research used details from 2 federal surveillance systems to not only characterize US Army suicide incidents but also to help describe the most prevalent circumstances that commonly precede death. Future studies are planned to compare the characteristics of this population to those of soldiers who died by suicide overseas and civilian suicide decedents to further determine their unique suicide circumstances. The findings in this report suggest focusing military suicide prevention efforts toward building positive intimate partner relationships, increasing coping skills to handle job-related problems, increasing access to mental health or substance abuse treatment, and providing support for soldiers currently in treatment.