Journal of Affective Disorders 235 (2018) 341–347

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Depression and impulsiveness among soldiers who died by suicide: A T

psychological autopsy study
Leah Shelefa,b, Neta Koremb, Nirit Yavnaic, Rinat Yedidyab, Keren Ginatb, Golan Shahard,
⁎
Assaf Yacobie,f,
a
Psychology Branch, Israeli Air Force, Ramat Gan, Israel
b
Mental Health Unit, Medical Corps, Israel Defense Force, Ramat Gan, Israel
c
Medical Corps, Israel Defense Force, Ramat Gan, Israel
d
Department of Psychology, Ben-Gurion University of the Negev, Beer-Sheva, Israel
e
Department of Psychiatry, Tel Aviv Sourasky Medical Center, Israel
f
Tel Aviv University Sackler Faculty of Medicine, Tel Aviv, Israel

A R T I C LE I N FO A B S T R A C T

Keywords: Introduction: Despite the accumulated knowledge about suicide, suicidal acts remain difficult to predict, and
Depression many suicides are acted out impulsively.
Impulsivity Methods: We performed a psychological autopsy study based on inquiries about the deaths of all male soldiers
Suicide aged 18–21 years who served in the Israeli army and died by suicide between 2009 and 2013 (n = 69). The study
Military
population was first divided into two groups: those who had depressive disorder (n = 31); and those who did not
Soldiers
(n = 38). Socio-demographic characteristics of the subjects and the characteristics of the suicidal act were
Adolescents
compared. Afterwards, the study population was re-divided by the presence or absence of impulsive personality
traits (n = 22, and n = 47, respectively), and investigated for distinct suicidal behavior features.
Results: No significant socio-demographic differences were found between the depressed and non-depressed
suicide victims. The depressed group had showed more signs of planning the act (47% vs. 23%), and had ex-
pressed suicidal ideation in the days preceding the suicide (51.6% vs. 21%). One third of the subjects were found
to have an impulsive personality trait, with significantly more histories of disciplinary issues, violence and
cluster B personality disorders. Alcohol use during the act was significantly more prevalent among impulsive
than non-impulsive subjects (45.4% vs. 14.9%).
Conclusion: Identification of distinct clinical groups of suicide victims among young males might help clinicians
evaluate high risk cases, and may provide valuable opportunities to alleviate and prevent these events in the
future.

1. Introduction
Suicide is the second leading cause of death worldwide among
adolescents and young adults aged 15–24 (Sullivan et al., 2015). In
contrast to the stable trend in the rate of suicide among civilians in the
US, the suicide rate in the US army has dramatically increased
(Archuleta et al., 2014; Nock et al., 2014). This evidence suggests that
soldiers might be a particularly relevant population for suicide re-
search, and Israeli soldiers might be particularly relevant for such re-
search because of compulsory military service for all citizens who reach
the age of 18. Hence, most of the country's healthy adolescent popu-
lation serves in the military and is under army medical surveillance
(Shelef et al., 2016). Research on suicide in this population is vital
owing to the high levels of stress and availability of weapons in the
military, and because the risk of suicide is at its peak at these ages
(Nock et al., 2008; Shelef et al., 2016).
The vulnerability-stress model proposes that an individual has un-
ique biological, psychological, and social elements, including strengths
and vulnerabilities for dealing with stress (Zubin and Spring, 1977).
This model may be helpful in understanding the phenomenon of sui-
cide, because it distinguishes vulnerability factors that predispose
people to be at high risk for suicide (e.g., presence of mental illness, and
prior suicidal behavior), as well as stress factors that trigger suicidal
behaviors among those who are vulnerable (e.g., hopelessness, adverse
life events; (Brown et al., 2000; Buchman-Schmitt et al., 2017;
Mann et al., 1999). Unipolar depression features prominently among

⁎
Corresponding author at: Department of Psychiatry, Tel Aviv Sourasky Medical Center, 10 Dafna St., Tel-Aviv, Israel.
E-mail address: assafy@tlvmc.gov.il (A. Yacobi).

https://doi.org/10.1016/j.jad.2018.04.068
Received 25 October 2017; Received in revised form 14 March 2018; Accepted 7 April 2018
Available online 10 April 2018
0165-0327/ © 2018 Elsevier B.V. All rights reserved.
L. Shelef et al.
these vulnerability factors, as it has been found to constitute the most
common diagnosis of individuals who have taken their lives by suicide
(Zalsman et al., 2016). Mood disorders, alcohol or substance abuse,
psychotic, and personality disorders have also been associated with
high risk for suicide in military as well as civilian settings (Mann et al.,
1999; Shelef et al., 2017).
In addition, there is growing evidence that specific psychological
factors which are not related to any particular mental disorder are
highly relevant to understanding the association between psycho-
pathology and suicidality (Nock et al., 2014). For example, in our
group, cognitive and emotional difficulties were found to be associated
with suicidal ideation in mentally healthy soldiers (Shelef et al., 2014).
Stressful life events have also been found to have a significant dose-
response effect on the occurrence of suicidal behavior (Stein et al.,
2010), and soldiers may experience military-related stressors (e.g.,
combat exposure and adjustment difficulties) or personal difficulties
(Hyman et al., 2012; Kuehn, 2009). In addition to the above-mentioned
factors, major risk factors such as male gender, prior suicidal or self-
harm behavior, and past family history of suicidal behavior
(Borges et al., 2010), as well as protective factors such as social support
and mental health treatment are also noteworthy (Beautrais, 2003;
Yacobi et al., 2013).
The interaction between prior vulnerability and stressful events is
still being explored, in an attempt to understand how these factors in-
teract to lead a certain individual to engage in a suicidal act at a given
time (Nock et al., 2013; Schapir et al., 2016). Research findings indicate
that in the past, it was believed that over 90% of suicides were asso-
ciated with diagnosable mental illness (Nock et al., 2008) or, at the very
least, with subclinical psychiatric symptoms, and that the vast majority
of suicidal individuals showed prominent clinical symptoms (Joiner
et al., 2017). However, recent psychological-autopsy studies have
challenged these findings, and suggest that up to 50% of all suicides are
not associated with mental disorders (Milner et al., 2013;
Pridmore, 2015). These studies emphasize the role of social and psy-
chological factors in the pathogenesis of suicidal behavior, and argue
against using mental disorders as the main target of treatment and
preventive interventions.
Despite the accumulated knowledge about suicide prevention stra-
tegies, it is still difficult to predict suicidal acts (Mann et al., 2005;
Zalsman et al., 2016). Research indicates that more than half of all
suicide attempters may be characterized as impulsive, and that a sig-
nificant portion of suicidal acts are carried out impulsively
(Rimkeviciene et al., 2015; Simon et al., 2001). Therefore, identifica-
tion of the factors that are associated with impulsive suicide acts may
shed light on different preventive strategies that can be used among this
subgroup of suicidal individuals (Baca-Garcia et al., 2005). Studies on
the relationship between impulsivity and the medical seriousness of
Table 1
Subjects Demographic Divided to Depressive and Non-Depressive Disordersa.
Depressive disorders (n = 31)
N %
Time in Service (months) 17.5 ( ± 11.59)
Country of Birth Native 24 77.4
Immigrant 7 22.6
Education (years) 12
Religion Secular Jewish 26 83.9
Religious 3 9.7
Non-Jewish 2 6.5
Service in combat unit 18 58.0
Method of suicide shooting 30 96.8
hanging 1 3.2
Place of suicide - at the base 12 38.7
a
Based on diagnosis made either before the suicidal event or by post-mortem psychiatric evaluation. Note: Chi-square was used for dichotomous variables. Fisher's
Exact Test was used for variables that are not dichotomous.
342
Journal of Affective Disorders 235 (2018) 341–347
suicide attempts have yielded conflicting findings. Recent reviews have
cast doubt on the role of impulsivity in suicidal behavior (Anestis et al.,
2014), although impulsive suicide ideators/attempters might simply
constitute a minority group among suicidal individuals. Thus, studies
have found that individuals who make impulsive suicide attempts re-
ported lower expectations of death as a result of their attempt, had less
suicide ideation, and were less depressed than non-impulsive suicide
attempters (Baca-Garcia et al., 2005; Simon et al., 2001; Wyder and De
Leo, 2007). Spokas et al. (2012) used the Suicide Intention Scale (SIS)
to compare individuals who make impulsive suicide attempts with
those who make premeditated attempts, and found that impulsive sui-
cide attempters were less depressed and expected that their attempts to
be less lethal. However, the lethality of suicide attempts was similar for
both groups. The authors therefore concluded that the presence of
impulsive behavior might increase the risk of suicide completion even
in the absence of depression and hopelessness (Spokas et al., 2012).
However, there is a lack of knowledge about the relationship between
impulsive behavior and the characteristics of suicide completers, be-
cause only a few studies have been carried out among this population
(Brent et al., 1994; Dumais et al., 2005; Maser et al., 2002; Zouk et al.,
2006). Existing studies have shown that compared with premeditated
suicides, impulsive suicides are characterized by a higher prevalence of
aggressive behavior, substance abuse, and psychiatric comorbidity,
particularly personality disorders (clusters B and C).
The current study aimed to examine mental illness and impulsivity
in suicide victims. We examined all cases of male soldiers in the Israeli
Defense Force (IDF) who took their lives by suicide during the period
2009–2013 (69 cases). The study hypothesis was that there are two
different types of suicides: the first, planned suicides, which are the
result from depression; the second, impulsive suicide in cases that had
past histories of impulsive pattern of behavior. We began by dividing
the study population into two groups: those who had depressive dis-
order at the time of the suicide, and those who did not. Subjects were
characterized in terms of socio-demographic parameters. In the second
phase of the study we concentrated on the suicidal act itself, and the
two groups were compared in light of their affective state before the act
of suicide. Finally, we focused on impulsive personality traits:
Specifically, we examined whether the soldiers had prior clinically
identifiable features and whether their suicidal act had distinctive
features.
2. Methods
2.1. Study population & data source
The study population included all cases of death by suicide of male
soldiers in compulsory military service, during the period 2009–2013 (a
Non-depressive disorders (n = 38) P Χ2
N %
16 ( ± 11.90) –
22 57.9
16 42.1 0.12 8.7
11.6 0.06 –
30 78.9
3 7.9
5 13.2 0.65 –
19 54.0 0.63 0.32
33 86.8
5 13.2 0.21 6.6
19 50 0.11 2.6
L. Shelef et al.
total of 69 cases). Female suicides were excluded due to the small po-
pulation size (n = 8). All of the soldiers were single and had no known
serious medical or mental conditions.
Table 1 shows the demographic characteristics of the study popu-
lation, which was divided to two groups: depressed subjects, and non-
depressed subjects. All of the soldiers were between 18 and 21 years of
age, and the average length of their service was 16.7 months
(SD = 11.5). Two thirds of the soldiers were Israeli-born; most were
secular Jews (81%), and 10% non-Jewish. Seventy-eight percent of the
participants had finished high school before enlisting in the military,
and 53% had been serving in combat units at the time of the suicide.
Almost all of the soldiers (91%) used a rifle to commit suicide, but less
than half (44.9%) committed suicide at home. Only 20% had received
mental health treatment before recruitment, and 40.5% had received
treatment during their service.
In the IDF, an independent committee of inquiry is appointed to
investigate every case of death by suicide. The committee investigates
the history of the solider as well as the events leading up to the suicide,
in an attempt to identify the factors that contributed to the event. The
committee bases its reports on testimonies given by family members,
commanders, and fellow soldiers, as well as on information gathered by
the military police about the soldier's character, personal background,
and details concerning the last couple of days before the suicide, in-
cluding the final hours before the act.
In this study, we performed retrospective data analysis based on the
reports of the committee of inquiry. In addition, data were gathered
from the IDF's demographic and medical databases (which include the
patient's mental health record). The data collected from these sources
were coded by a clinical psychologist, as part of a Master's degree thesis
on suicidality of soldiers, who had been blinded as to the hypothesis
and aims of this particular study.
2.2. Measures
2.2.1. Psychiatric and psychological assessment
In addition to prior known psychiatric diagnoses, all of the cases
examined in the study had undergone a post-hoc psychiatric analysis by
the chief military psychiatrist as part of the routine suicide investiga-
tion. Notably, psychiatric diagnoses established using proxy-based
evaluation has been found to be reliable (Conner et al., 2001). Psy-
chiatric diagnoses were based on the ICD-10 manual. In the present
analysis, three main diagnostic groups were used: (1) Cluster B per-
sonality disorders are characterized by marked difficulties in impulse
control and high levels of sensitivity in interpersonal relationships.
These include Narcissistic, Borderline, Histrionic, and Anti-Social Per-
sonality Disorders. We decided to focus on this group of personality
disorders due to their high prevalence in the present study (49% of the
subjects), whereas only 9 subjects (13%) had been diagnosed with other
types of personality disorders; (2) Depressive disorders, including major
depressive disorder and dysthymia; and (3) Adjustment disorder, de-
fined by the DSM-V as a maladaptive reaction to an identifiable psy-
chosocial stressor or stressors, that occurs within three months after the
onset of that stressor. The existence of impulsive personality trait was
based on the psychiatric reports investigating the incidents. Impulsive
personality trait was defined as having at least one of the following: (1)
a psychiatric diagnosis of impulse control disorder; (2) cluster B per-
sonality disorders which included specific criteria for impulsivity; (3)
impulsive patterns of behavior including: known history of repetitive
behaviors of acting without planning or acting with no regards to the
consequences (Lewis et al., 2016, Zouk et al., 2006). Other character-
istics examined were prior known violent behavior or disciplinary
problems recorded in the soldier's military file.
2.2.2. Assessment of the impulsivity of suicidal act
In light of the substantial disagreement about the valid criteria for
defining an impulsive suicide attempt (Rimkeviciene et al., 2015),
343
Journal of Affective Disorders 235 (2018) 341–347
especially in psychological autopsy studies, specific criteria for de-
termining the impulsivity of the suicidal event as well as an overall SIS
score (Beck et al., 1974; Conner, 2004) filled by an independent re-
searcher were used in the present study. Specific criteria for impulsivity
included: (1) the level of planning the act known to the others in the
victim's environment (not known planning, partial planning, or full
planning); (2) suicidal expressions less than three hours before the act
or a week before the act, which include suicidal threats, intimation, or
gestures; (3) the presence of suicide note; (4) whether the subject was
alone during the act; and (5) the possibility of receiving medical at-
tention. In addition, the use of alcohol during the act was examined.
Because the use of other substances was found in only one case, it is not
mentioned in the Results section.
The SIS is an interview-based measure of the seriousness of the in-
tent to attempt suicide among patients who have actually attempted
suicide. There is abundant evidence suggesting that the SIS is a reliable
tool for assessment of the intent to die and the impulsivity of suicidal
acts (Conner, 2004). This questionnaire consists of 15 items (score
ranging from 0 to 30), which quantify the verbal and nonverbal beha-
vior of suicide attempters prior to and during the most recent suicide
attempt. Based on cut-off values mentioned in prior research literature,
as well as on cluster analysis and multidimensional scaling, we used a
cut-off score of 20 to distinguish low and high intent to die. The SIS
questionnaire has been shown to have a high Cronbach's alpha relia-
bility (0.95). The reliability of the questionnaire used in our study was
0.80 (Beck et al., 1974; Brown et al., 2000). The possible reasons for the
suicidal act were categorized as: a desire to end suffering, or a desire to
avenge or punish someone else (parents, spouses, commanders).
2.3. Statistical analysis
SPSS version 20.0 for Windows was used for all of the analyses.
Univariate analysis included a Chi Square test for categorical in-
dependent variables (demographic, personal characteristics and psy-
chiatric diagnosis), which was used to detect differences between
groups. Fishers' Exact Test was used to assess high intent- low-intent
differences. For categorical variables with more than two groups, each
group was compared with the other groups. For continues variables,
Mann Whitney test or Ttest were conducted. The level of statistical
significance was set to p = .05.
2.4. Ethics approval
The institutional review board of the IDF Medical Corps approved
the study and waived the requirement for informed consent on the basis
of preserving the participants' anonymity.
3. Results
The subjects were initially divided into groups – depressive disorder
(n = 31; 44.9%), and non-depressive disorder (n = 38; 55.1%). This
was based on a diagnosis made either before they died or on a post-
mortem psychiatric evaluation (Table 1). There were no statistically
significant differences in the demographic characteristics of each group,
including number of years of education and religiosity, or in the method
of suicide used (Table 1).
The suicidal attributions of depressive and non-depressive suicides
are presented in Table 2. As can be seen in the table, most of the soldiers
in both groups did not have a history of prior suicide attempts or self-
injurious behaviors. However, a significantly higher percentage of the
soldiers in the non-depressed group had a history of past self-injury
than those in the depressed group (18.4% vs. 3.2%; p = .02), whereas
only slightly more of the non-depressed soldiers had either threatened
suicide or actually made suicide attempts in the past (9.7% vs. 13.1%,
p = .42; 12.9% vs. 21%, p = .05).
In regard to warning signs and planning, there were no significant
L. Shelef et al. Journal of Affective Disorders 235 (2018) 341–347

Table 2
Differences in Suicidal Attributions between Depressive and Non-Depressive Disorders Groups.
Depressive disorders (n = 31) Non-depressive disorders (n = 38) P Χ2

N % N %

Prior suicide attempt 4 12.9 8 21.0 0.05 2.3

Prior suicide threat 3 9.7 5 13.1 0.42 0.65
Prior self-injury 1 3.2 7 18.4 0.02 11.2
Physical Complaints (3 months) 9 29 9 23.6 0.39 0.75
Suicidal expressions (1–7 days) 16 51.6 8 21.0 0.01 20.2
Suicidal expressions (< 3 h) 31 100 24 63.1 <0.001 34.8
Level of planning Without planning 2 6.4 18 4704 <0.001 42.7
Partial planning 15 48.3 11 28.9 0.04 7.9
Full planning 14 45.2 9 23.7 <0.001 10.2
Suicide note 17 54.8 9 23.7 0.01 20.3
Alone during suicide 27 87.1 25 65.8 0.03 12.6
Medical help possible 3 9.7 13 34.2 0.01 14.5
End of suffering as suicide motivation 31 100 32 84.2 0.01 17.1
Alcohol use during suicide 3 10 14 36.8 0.004 16.6
Cluster A Personality Disorder 2 6.4 2 5.2 0.7 0.13
Cluster B Personality Disorder 11 35.4 23 60.5 0.04 12.6
Cluster C Personality Disorder 8 25 7 18.4 0.26 1.3
Anxiety Disorders 1 3.2 1 2.6 0.8 0. 64
Impulsive trait 4 12.9 18 47.4 0.001 28.3
SIS Score > 20 25 80.6 16 42.1 0.001 16.8

Note: Pearson chi-square coefficients are presented.

difference in physical complaints between the depressed group and
non-depressed groups (29% vs 23.6, p = .39). However, significantly
more of the soldiers in the depressed group had expressed suicidal
ideation and gestures in the days preceding the suicide than those in the
non-depressed group (51.6% vs. 21%; p = .01). Forty five percent of the
soldiers in the depressed group had left signs of fully planning the
suicide, compared with 23% of the non-depressed group (p < .001).
Soldiers in the depressed group left more suicide notes than those in the
non-depressed group (54.8% vs. 23.7%, p = .01); more of them had
been alone during the act (87% vs. 65.8%, p = .035), they used less
alcohol during the act (10% vs. 36.8%, p = .004) and did it in a way
that made medical help less possible (9.7% vs. 34.2%, p = .01). There
was only one incident of drug use during the suicidal act (marijuana) by
a non-depressed subject.
Regarding the reason for suicide, the desire to end suffering was
appraised as the motivation in 100% of the depressed subjects, com-
pared to 84.2% of those in the non-depressed group (p = .006).
As for co-morbid diagnoses, there was no significant difference in
the prevalence of anxiety disorders (3.2% vs. 2.6, p = .8) or personality
disorders, other than cluster B personality disorders. Overall, the per-
centage of subjects with a high SIS score (> 20) was significantly
higher in the depressed group than in the non-depressed group (80.6%
vs. 42.1%, p = .001), and they were found to have significantly fewer
impulsive characteristics than their non-depressed peers (12.9% vs.
47.4%. p = .001).
Thirty two percent of the subjects in the study were characterized as
having impulsive traits (Table 3). A significantly higher percentage of
those with impulsive traits had a history of past disciplinary issues in
the army and violent behavior records compared with those who did
not have impulsive traits (63.6% vs. 19.1%. p < .001; and 18.2% vs.
2.1%. p = .038 of the impulsive and non-impulsive subjects, respec-
tively). Regarding clinical diagnoses, a significantly higher percentage
of the subjects in the impulsive trait group had been diagnosed with
cluster B personality disorders (77.3% vs. 36.2%, p = .004), but not
adjustment disorder (50% vs.48.9, p = .39). As for mental health
treatment, the percentage of soldiers with impulsive traits who were
treated before military service was significantly higher than the per-
centage of non-impulsive soldiers (40.9% vs. 10.6%, p = .01). How-
ever, the differences between the two groups were not significant with
regard to either treatment during their entire military service (45.4%
vs. 38.3%, p = .38) or within 3 months of committing suicide (22.7%
vs. 19.1%, p = .53).
During the week before the suicide, the impulsive subjects had ex-
perienced significantly more stressful events that were not related to
their military service (an average of 1.82 ± 0.5 vs. 0.98 ± 0.3 events,
p = .047), but only slightly more stressful events within their service
(1.54 ± 0.4 vs. 1.04 ± 0.5 events, p = .06). During the day of the
suicide, 45.4% of the subjects in the impulsive trait group had ex-
pressed suicidal ideation, compared to 78.7% of those in the non-im-
pulsive trait group (p < .001). However, during the last three hours
before the event 54.4% of those in the impulsive trait group expressed
suicidal thoughts, compared to only 21.3% of those in the non-im-
pulsive group (p < .007).
Within the period of 3 months before the suicide, impulsive subjects
had more changes in behavior (extreme and sudden changes including
anger outbursts, becoming reclusive or acting out) and psychomotor
agitation (18% vs.6.4%, p = .11; 27.3% vs.12.7%, p = .02, respec-
tively). During the week before the suicide, however, these differences
were not significant, and both groups had an increase incidences in
behavioral changes and psychomotor agitation.
Revenge was found to be a major reason for suicide in the impulsive
group, but not in the non-impulsive group (77.3% vs. 25.5%; p = .002),
and the use of alcohol was significantly more prevalent among the
impulsive subjects (45.4% vs. 14.9%; p = .008). The percentage of non-
impulsive subjects with a high total SIS score (> 20) was significantly
higher among the non-impulsive subjects than among the impulsive
subjects (70.2% vs. 36.3%; p = .008).
3.1. Discussion
The findings in our study indicate that the characteristics of young
male soldiers who killed themselves and had depression were similar to
those who did not have a major depressive disorder in terms of de-
mographic background, education, and religiosity (Table 1). These re-
sults are consistent with previous studies that compared demographic
characteristics (such as age, gender, marital status and education) of
individuals who made suicidal attempts with different levels of de-
pression or levels of intent to die (Spokas et al., 2012; Wojnar et al.,
2009). Moreover, the finding that almost all of the subjects used fire-
arms to kill themselves is in accordance with suicide prevention stra-
tegies in the Israeli military (Shelef et al., 2016; Zalsman et al., 2016).
These strategies hypothesize that reducing availability to firearms

344
L. Shelef et al. Journal of Affective Disorders 235 (2018) 341–347

Table 3
Impulsive trait and impulsive suicide.
Impulsive trait (n = 22) Non-Impulsive trait (n = 47) P Χ2

N % N %

Disciplinary problems 14 63.6 9 19.1 0.0003 40.8

Violent behavior 4 18.2 1 2.1 0.04 12.6
Cluster B personality disorder 17 77.3 17 36.2 0.004 41.6
Adjustment disorder 11 50 23 48.9 0.39 0.02
Mental health treatment before service 9 40.9 5 10.6 0.01 24
Mental health treatment during service 10 45.4 18 38.3 0.38 1.2
Mental health treatment within 3 months 5 22.7 9 19.1 0.53 3.9
Average of stressful events 1.54 (0.4) 1.04 (0.5) 0.06 –
within service within 7 days (SD)
Average of stressful events 1.82 (0.5) 0.98 (0.3) 0.047 –
outside of service within 7 days (SD)
Changes in behavior (3 months) 4 18.2 3 6.4 0.011 6.45
Changes in behavior (7 days) 11 50 19 40.4 0.17 1.9
Psychomotor agitation (3 months) 6 27.3 6 12.7 0.02 5.7
Psychomotor agitation (7 days) 8 36.4 13 27.6 0.18 1.76
Suicidal expressions more than 3 h 10 45.4 37 78.7 0.004 22.5
Suicidal expressions less than 3 h 12 54.4 10 21.3 0.01 23.7
Revenge as suicide motivation 17 77.3 12 25.5 0.002 53.7
Alcohol use during suicide 10 45.4 7 14.9 0.01 22
SIS Score > 20 8 36.3 33 70.2 0.01 29

Note: Chi-square was used for dichotomous variables. Fisher's Exact Test was used for variables that are not dichotomous.
SD = Standard Deviation

might reduce suicide rates in the military.
Among this population of young soldiers, most of the suicide victims
in our study did not have past histories of self-injurious behavior
(n = 8, 11.6%), suicidal threats (n = 8, 11.6%) or suicide attempts
(n = 12, 17.3%). The depressed group had even less of a history of
these types of behaviors than the non-depressed group (Table 2). This
finding is contradictory with the results of previous studies on suicide,
which found that persons with previous self-harm behavior are at
higher risk for suicide than the general population (Beautrais, 2001).
Moreover, in our study most of the non-depressed subjects and only half
of the depressed expressed any suicidal thoughts and intentions during
the week before the suicide (Table 2). These findings emphasize the
difficulties in identifying high-risk cases for complete suicides, even
during the days leading up to the act.
The suicidal act performed by the subjects in the depressed group
seems to be premeditated with a higher degree of intent to die, in-
cluding: planning, leaving a note, making sure they are alone during the
act and that they will be unlikely to receive medical help (Table 2). This
is consistent with previous findings on serious non-impulsive suicide
attempts (Spokas et al., 2012), as well as with recent reviews which
highlight the planned nature of many suicides (Anestis et al., 2014).
Post-mortem analysis revealed that the depressed subjects were all
driven by feelings of hopelessness and a desire to end their suffering,
complying with similar findings in previous studies (Goldston et al.,
2001; Hunter and O'Connor, 2003).
The characteristics of the suicidal act among the non-depressed
subjects in the current study were found to be different from those of
the depressed subjects. Despite the fact that their act was lethal enough
to result in death, there are indications that the intent of the non-de-
pressed subjects was not as high as the intent of those in the depressed
group (Table 1). This result is consistent with the findings of
Spokas et al. (2012) on suicide attempters. They did not plan the act as
fully, used more alcohol, and performed the act in a manner that left a
greater chance for them to be found and receive medical attention.
These findings indicate that the suicides committed by the non-de-
pressed group were more impulsive in nature, and are supported by the
lower scores of the non-depressed subjects on the SIS questionnaire
(Conner, 2004). Examination of the histories of the non-depressed
subjects indicates that almost half of them (47%) had impulsive per-
sonality traits, compared with 12.9% of those in the depressed group
(Table 2). These findings are congruent with the results of past studies
(Marttunen et al., 1998). They support the theory that there are two
types of suicidal acts (premeditated and impulsive), which correlate
with two types of background characteristics (depressed or impulsive
personalities) among young male suicide completers who did not have
debilitating mental illness. While in their study, Zouk et al. (2006) in-
cluded depressive disorder NOS in the depressive group, and concluded
that both premeditated and impulsive suicide completers experienced
similar depressive symptoms. In this study we included only major
depression and dysthymia in the depressive suicide group. Our con-
clusion is, therefore, that raising the cut-off symptoms to major de-
pression and dysthymia distinguishes effectively between the two
groups.
Suicidal soldiers with impulsive personality traits were found to
have more Cluster B personality disorders as well as histories of violent
behavior and disciplinary issues. High prevalence of cluster B disorders
has been found in previous research, especially when the suicidal acts
were a quick response to an inter-personal conflict (Dumais et al., 2005;
Pompili et al., 2011; Zouk et al., 2006).
Examination of the days and hours before the suicide reveals that
the soldiers in the impulsive group had experienced twice as many
stressful events outside of the military than their non-impulsive coun-
terparts (e.g., family disputes, financial difficulties, and difficulties in
intimate relationships). Additionally, about half of the soldiers ex-
amined in the study, both impulsive and non-impulsive, had experi-
enced adjustment disorder during their service, indicating that they had
considerable difficulty conforming to their new environment. Despite
these difficulties, it is concerning that only 20% of the soldiers had
received any kind of mental health treatment before recruitment, less
than half had been treated during their service (Table 3). Only 10% of
the non-impulsive group had ever been treated before recruitment, and
38% had been treated during their service. Furthermore, a fifth of the
subjects in both groups had met with a mental health professional
during the 3 months leading to the suicides, possibly indicating that the
mental difficulties of 80% of the subjects in the study had gone un-
noticed. These findings might be attributed to well-known difficulties in
seeking help by traditional mental health services (Wang et al., 2007).
Similar difficulties are described both in other military settings and
among populations under stress, therefore alternative prevention
models have been implemented, such as peer support, improved

345
L. Shelef et al.
handling of suicidal crises by unit command, media, and phone/In-
ternet-based interventions have been recommended (Nock et al., 2014;
Shelef et al., 2016).
Interestingly, those with impulsive personality trait had shown
more behavioral changes (anger outbursts, acting out or becoming re-
clusive) and psychomotor agitation in the three months period before
the suicides compared with non-impulsive subjects (Table 3). In the
week before the events, both impulsive and non-impulsive subjects had
more behavioral changes and psychomotor agitation, with no sig-
nificant difference between the groups. On the day of the suicide, most
of the non-impulsive soldiers (78.7%) had expressed suicidal ideation,
whereas less than half of the impulsive soldiers had expressed such
thoughts. This trend was reversed during the last few hours before the
act, when only 21% of the non-impulsive group expressed suicidal
ideation and intent, as opposed to more than half of the impulsive
group. This finding might indicate a sudden onset of suicidal thoughts
and urges in the impulsive group, whereas the non-impulsive subjects
became quiet because they had already decided to execute their plan
(Wyder and De Leo, 2007).
The suicide of the impulsive subjects was found to have occurred in
the context of interpersonal relationships (Table 3). Most of them
(77.3%) acted out of revenge and wanted to punish others for their
suffering, in contrast to their non-impulsive peers, who acted out of
high suicidal intent. The use of alcohol during the suicidal act perhaps
served as an additional risk factor which lowered inhibitions, increased
aggression, and focused attention on suicide (Hufford, 2001).
3.2. Study limitations
The Israeli military recruits all healthy 18–21 year-old males in the
country, under mandatory law. This unique situation allows for in-
depth research of rare phenomena such as complete suicide, but also
limit the ability to extrapolate to adolescents in other cultures or sol-
diers in countries where voluntary service is applied. Studies comparing
Israeli and American soldiers have found that suicidality is more
common among young new male recruits, who also tend to suffer from
high rates of mental health difficulties (Bryan and Rudd, 2012;
Gilman et al., 2014).
Psychological autopsy studies are subject to the inherent metho-
dological limitation of using proxy-based information. To minimize this
limitation, much of the information in this study was based either on
prior documented data or post-hoc independent psychiatric evaluation.
The validity of this procedure with regard to the type of variables used
in the current study has been well demonstrated in previous research
(Conner et al., 2012). Another limitation of this study is the small size of
the study population. Despite the anticipated low power of statistical
analysis, statistically significant differences were found in most of the
variables examined, which are indicative of marked differences in the
factors among the study groups.
3.3. Conclusions and future directions
These days most of the strategies for preventing suicide involve
detection and treatment of depression (Mann et al., 2005;
Zalsman et al., 2016). It is possible that over-emphasizing depression
and psychiatric diagnosis may lead clinicians to underestimate the
importance of impulsivity in adolescent and young soldiers at risk for
suicide, thus missing valuable opportunities to prevent these events
(McGirr and Turecki, 2007; Spirito and Esposito-Smythers, 2006).
To our knowledge this study is the first to show a clear connection
between levels of suicidal intent and pre-morbid states, either depres-
sion or impulsive personality traits, in individuals who died by suicide.
We hope that future studies will examine if these patterns reproduce in
other populations (civilians, younger adolescents, elderly etc.).
Furthermore, the findings in this study emphasize the need for deeper
understanding of impulsive personality traits in suicide-prevention
346
Journal of Affective Disorders 235 (2018) 341–347
research.
Conflict of Interest
Authors declare no conflict of interest.
Contributors
All of the listed authors in the article have taken an active part in the
making of the study or writing the article. The study was designed and
planned by Leah Shelef and Assaf Yacobi. Leah Shelef, Neta Corem and
Nirit Yavnai had full access to all of the data in the study and take
responsibility for the integrity of the data and the accuracy of the data
analysis. Statistical analysis was done by Nirit Yavnai and Assaf Yacobi.
The analysis of the data, interpretations and drafting of the manu-
script were done collectively by all of the authors. The final critical
revisions of the manuscript for were prepared by the joint effort of Leah
Shelef, Golan Shahar and Assaf Yacobi.
Lastly, the data presented in the article has not been published or
presented in other places.
Funding
This study was not funded or supported financially by any group or
organization.
Acknowledgment
We would like to express our gratitude to the medical corps of the
Israeli military for its support in making this study possible.
Supplementary materials
Supplementary material associated with this article can be found, in
the online version, at doi:10.1016/j.jad.2018.04.068.
References
Anestis, M.D., Soberay, K.A., Gutierrez, P.M., Hernandez, T.D., Joiner, T.E., 2014.
Reconsidering the link between impulsivity and suicidal behavior. Pers. Soc. Psychol.
Rev. 18, 366–386.
Archuleta, D., Jobes, D.A., Pujol, L., Jennings, K., Crumlish, J., Lento, R.M., Brazaitis, K.,
Moore, B.A., Crow, B., 2014. Raising the clinical standard of care for suicidal soldiers:
an army process improvement initiative. US Army Med. Dep. J. 55–66.
Baca-Garcia, E., Diaz-Sastre, C., Garcia Resa, E., Blasco, H., Braquehais Conesa, D.,
Oquendo, M.A., Saiz-Ruiz, J., de Leon, J., 2005. Suicide attempts and impulsivity.
Eur. Arch. Psychiatry Clin. Neurosci. 255, 152–156.
Beautrais, A., 2003. Suicide in New Zealand II: a review of risk factors and prevention. N.
Z. Med. J. 116, U461.
Beautrais, A.L., 2001. Suicides and serious suicide attempts: two populations or one.
Psychol. Med. 31, 837–845.
Beck, R.W., Morris, J., Lester, D., 1974. Suicide notes and risk of future suicide. JAMA
228, 495–496.
Borges, G., Nock, M.K., Haro Abad, J.M., Hwang, I., Sampson, N.A., Alonso, J., Andrade,
L.H., Angermeyer, M.C., Beautrais, A., Bromet, E., Bruffaerts, R., de Girolamo, G.,
Florescu, S., Gureje, O., Hu, C., Karam, E.G., Kovess-Masfety, V., Lee, S., Levinson, D.,
Medina-Mora, M.E., Ormel, J., Posada-Villa, J., Sagar, R., Tomov, T., Uda, H.,
Williams, D.R., Kessler, R.C., 2010. Twelve-month prevalence of and risk factors for
suicide attempts in the World Health Organization World Mental Health Surveys. J.
Clin. Psychiatry 71, 1617–1628.
Brent, D.A., Johnson, B.A., Perper, J., Connolly, J., Bridge, J., Bartle, S., Rather, C., 1994.
Personality disorder, personality traits, impulsive violence, and completed suicide in
adolescents. J. Am. Acad. Child. Adolesc. Psychiatry 33, 1080–1086.
Brown, G.K., Beck, A.T., Steer, R.A., Grisham, J.R., 2000. Risk factors for suicide in
psychiatric outpatients: a 20-year prospective study. J. Consult. Clin. Psychol. 68,
371–377.
Bryan, C.J., Rudd, M.D., 2012. Life stressors, emotional distress, and trauma-related
thoughts occurring in the 24 h preceding active duty U.S. soldiers' suicide attempts. J.
Psychiatr. Res. 46, 843–848.
Buchman-Schmitt, J.M., Chu, C., Michaels, M.S., Hames, J.L., Silva, C., Hagan, C.R.,
Ribeiro, J.D., Selby, E.A., Joiner Jr., T.E., 2017. The role of stressful life events
preceding death by suicide: Evidence from two samples of suicide decedents.
Psychiatry Res. 256, 345–352.
Conner, K.R., 2004. A call for research on planned vs. unplanned suicidal behavior.
L. Shelef et al.
Suicide Life Threat Behav. 34, 89–98.
Conner, K.R., Beautrais, A.L., Brent, D.A., Conwell, Y., Phillips, M.R., Schneider, B., 2012.
The next generation of psychological autopsy studies: part 2. Interview procedures.
Suicide Life Threat Behav. 42, 86–103.
Conner, K.R., Duberstein, P.R., Conwell, Y., 2001. 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. 104, 204–209.
Dumais, A., Lesage, A.D., Alda, M., Rouleau, G., Dumont, M., Chawky, N., Roy, M., Mann,
J.J., Benkelfat, C., Turecki, G., 2005. Risk factors for suicide completion in major
depression: a case-control study of impulsive and aggressive behaviors in men. Am. J.
Psychiatry 162, 2116–2124.
Gilman, S.E., Bromet, E.J., Cox, K.L., Colpe, L.J., Fullerton, C.S., Gruber, M.J., Heeringa,
S.G., Lewandowski-Romps, L., Millikan-Bell, A.M., Naifeh, J.A., Nock, M.K.,
Petukhova, M.V., Sampson, N.A., Schoenbaum, M., Stein, M.B., Ursano, R.J., Wessely,
S., Zaslavsky, A.M., Kessler, R.C., 2014. Sociodemographic and career history pre-
dictors of suicide mortality in the United States Army 2004-2009. Psychol. Med. 44,
2579–2592.
Goldston, D.B., Daniel, S.S., Reboussin, B.A., Reboussin, D.M., Frazier, P.H., Harris, A.E.,
2001. Cognitive risk factors and suicide attempts among formerly hospitalized ado-
lescents: a prospective naturalistic study. J. Am. Acad. Child Adolesc. Psychiatry 40,
91–99.
Hufford, M.R., 2001. Alcohol and suicidal behavior. Clin. Psychol. Rev. 21, 797–811.
Hunter, E.C., O'Connor, R.C., 2003. Hopelessness and future thinking in parasuicide: the
role of perfectionism. Br. J. Clin. Psychol. 42, 355–365.
Hyman, J., Ireland, R., Frost, L., Cottrell, L., 2012. Suicide incidence and risk factors in an
active duty US military population. Am. J. Public Health 102 (Suppl 1), S138–S146.
Joiner, T.E., Buchman‐Schmitt, J.M., Chu, C., 2017. Do undiagnosed suicide decedents
have symptoms of a mental disorder? J. Clin. Psychol. 73, 1744–1752. http://dx.doi.
org/10.1002/jclp.22498.
Kuehn, B.M., 2009. Soldier suicide rates continue to rise: military, scientists work to stem
the tide. JAMA 301 (1111), 1113.
Lewis, K.C., Meehan, K.B., Cain, N.M., Wong, P.S., 2016. Within the confines of character:
A review of suicidal behavior and personality style. Psychoanalytic Psychol. 3,
179–202.
Mann, J.J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., Hegerl, U.,
Lonnqvist, J., Malone, K., Marusic, A., Mehlum, L., Patton, G., Phillips, M., Rutz, W.,
Rihmer, Z., Schmidtke, A., Shaffer, D., Silverman, M., Takahashi, Y., Varnik, A.,
Wasserman, D., Yip, P., Hendin, H., 2005. Suicide prevention strategies: a systematic
review. JAMA 294, 2064–2074.
Mann, J.J., Waternaux, C., Haas, G.L., Malone, K.M., 1999. Toward a clinical model of
suicidal behavior in psychiatric patients. Am. J. Psychiatry 156, 181–189.
Marttunen, M.J., Henriksson, M.M., Isometsa, E.T., Heikkinen, M.E., Aro, H.M.,
Lonnqvist, J.K., 1998. Completed suicide among adolescents with no diagnosable
psychiatric disorder. Adolescence 33, 669–681.
Maser, J.D., Akiskal, H.S., Schettler, P., Scheftner, W., Mueller, T., Endicott, J., Solomon,
D., Clayton, P., 2002. Can temperament identify affectively ill patients who engage in
lethal or near-lethal suicidal behavior? A 14-year prospective study. Suicide Life
Threat Behav. 32, 10–32.
McGirr, A., Turecki, G., 2007. The relationship of impulsive aggressiveness to suicidality
and other depression-linked behaviors. Curr. Psychiatry Rep. 9, 460–466.
Milner, A., Sveticic, J., De Leo, D., 2013. Suicide in the absence of mental disorder? A
review of psychological autopsy studies across countries. Int. J. Soc. Psychiatry 59,
545–554.
Nock, M.K., Borges, G., Bromet, E.J., Alonso, J., Angermeyer, M., Beautrais, A., Bruffaerts,
R., Chiu, W.T., de Girolamo, G., Gluzman, S., de Graaf, R., Gureje, O., Haro, J.M.,
Huang, Y., Karam, E., Kessler, R.C., Lepine, J.P., Levinson, D., Medina-Mora, M.E.,
Ono, Y., Posada-Villa, J., Williams, D., 2008. Cross-national prevalence and risk
factors for suicidal ideation, plans and attempts. Br. J. Psychiatry 192, 98–105.
Nock, M.K., Deming, C.A., Fullerton, C.S., Gilman, S.E., Goldenberg, M., Kessler, R.C.,
McCarroll, J.E., McLaughlin, K.A., Peterson, C., Schoenbaum, M., Stanley, B., Ursano,
R.J., 2013. Suicide among soldiers: a review of psychosocial risk and protective
factors. Psychiatry 76, 97–125.
Nock, M.K., Stein, M.B., Heeringa, S.G., Ursano, R.J., Colpe, L.J., Fullerton, C.S., Hwang,
I., Naifeh, J.A., Sampson, N.A., Schoenbaum, M., Zaslavsky, A.M., Kessler, R.C., 2014.
347
Journal of Affective Disorders 235 (2018) 341–347
Prevalence and correlates of suicidal behavior among soldiers: results from the Army
Study to Assess Risk and Resilience in Servicemembers (Army STARRS). JAMA
Psychiatry 71, 514–522.
Pompili, M., Innamorati, M., Szanto, K., Di Vittorio, C., Conwell, Y., Lester, D., Tatarelli,
R., Girardi, P., Amore, M., 2011. Life events as precipitants of suicide attempts among
first-time suicide attempters, repeaters, and non-attempters. Psychiatry Res. 186,
300–305.
Pridmore, S., 2015. Mental disorder and suicide: a faulty connection. Aust. N. Z. J.
Psychiatry 49, 18–20.
Rimkeviciene, J., O'Gorman, J., De Leo, D., 2015. Impulsive suicide attempts: a systematic
literature review of definitions, characteristics and risk factors. J. Affect Disord. 171,
93–104.
Schapir, L., Zalsman, G., Hasson-Ohayon, I., Gaziel, M., Morag-Yaffe, M., Sever, J.,
Weizman, A., Shoval, G., 2016. Suicide, satisfaction with life, and insight capacity
among adolescents with mental disorders. Crisis 37, 347–352.
Shelef, L., Fruchter, E., Mann, J.J., Yacobi, A., 2014. Correlations between interpersonal
and cognitive difficulties: Relationship to suicidal ideation in military suicide at-
tempters. Eur. Psychiatry 29, 498–502.
Shelef, L., Tatsa-Laur, L., Derazne, E., Mann, J.J., Fruchter, E., 2016. An effective suicide
prevention program in the Israeli Defense Forces: A cohort study. Eur. Psychiatry 31,
37–43.
Shelef, L., Tomer, G., Tatsa-Laur, L., Kedem, R., Bonne, O., Fruchter, E., 2017. Risk factors
for suicide in the Israeli army between the years 1992-2012: A case-control study.
Eur. Psychiatry 39, 106–113.
Simon, O.R., Swann, A.C., Powell, K.E., Potter, L.B., Kresnow, M.J., O'Carroll, P.W., 2001.
Characteristics of impulsive suicide attempts and attempters. Suicide Life Threat
Behav. 32, 49–59.
Spirito, A., Esposito-Smythers, C., 2006. Attempted and completed suicide in adolescence.
Annu. Rev. Clin. Psychol. 2, 237–266.
Spokas, M., Wenzel, A., Brown, G.K., Beck, A.T., 2012. Characteristics of individuals who
make impulsive suicide attempts. J. Affect Disord. 136, 1121–1125.
Stein, D.J., Chiu, W.T., Hwang, I., Kessler, R.C., Sampson, N., Alonso, J., Borges, G.,
Bromet, E., Bruffaerts, R., de Girolamo, G., Florescu, S., Gureje, O., He, Y., Kovess-
Masfety, V., Levinson, D., Matschinger, H., Mneimneh, Z., Nakamura, Y., Ormel, J.,
Posada-Villa, J., Sagar, R., Scott, K.M., Tomov, T., Viana, M.C., Williams, D.R., Nock,
M.K., 2010. Cross-national analysis of the associations between traumatic events and
suicidal behavior: findings from the WHO World Mental Health Surveys. PLoS One 5,
e10574.
Sullivan, E.M., Annest, J.L., Simon, T.R., Luo, F., Dahlberg, L.L., 2015. Suicide trends
among persons aged 10-24 years–United States, 1994-2012. MMWR Morb. Mortal.
Wkly. Rep. 64, 201–205.
Wang, P.S., Angermeyer, M., Borges, G., Bruffaerts, R., Tat Chiu, W., G, D. E., G., Fayyad,
J., Gureje, O., Haro, J.M., Huang, Y., Kessler, R.C., Kovess, V., Levinson, D., Nakane,
Y., Oakley Brown, M.A., Ormel, J.H., Posada-Villa, J., Aguilar-Gaxiola, S., Alonso, J.,
Lee, S., Heeringa, S., Pennell, B.E., Chatterji, S., Ustun, T.B., 2007. Delay and failure
in treatment seeking after first onset of mental disorders in the World Health
Organization's World Mental Health Survey Initiative. World Psychiatry 6, 177–185.
Wojnar, M., Ilgen, M.A., Czyz, E., Strobbe, S., Klimkiewicz, A., Jakubczyk, A., Glass, J.,
Brower, K.J., 2009. Impulsive and non-impulsive suicide attempts in patients treated
for alcohol dependence. J. Affect Disord. 115, 131–139.
Wyder, M., De Leo, D., 2007. Behind impulsive suicide attempts: indications from a
community study. J. Affect Disord. 104, 167–173.
Yacobi, A., Fruchter, E., Mann, J.J., Shelef, L., 2013. Differentiating army suicide at-
tempters from psychologically treated and untreated soldiers: A demographic, psy-
chological and stress-reaction characterization. J. Affect Disord.
Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone,
M., Carli, V., Hoschl, C., Barzilay, R., Balazs, J., Purebl, G., Kahn, J.P., Saiz, P.A.,
Lipsicas, C.B., Bobes, J., Cozman, D., Hegerl, U., Zohar, J., 2016. Suicide prevention
strategies revisited: 10-year systematic review. Lancet Psychiatry 3, 646–659.
Zouk, H., Tousignant, M., Seguin, M., Lesage, A., Turecki, G., 2006. Characterization of
impulsivity in suicide completers: clinical, behavioral and psychosocial dimensions.
J. Affect Disord. 92, 195–204.
Zubin, J., Spring, B., 1977. Vulnerability–a new view of schizophrenia. J. Abnorm.
Psychol. 86, 103–126.