Suicide and Life-Threatening Behavior 37(4) August 2007
2007 The American Association of Suicidology
367
The Association between Life Events and
Suicide Intent in Self-Poisoners with and
without a History of Deliberate Self-Harm:
A Preliminary Study
Catherine Crane, D Phil, J. Mark. G. Williams, DSc, Keith Hawton, DSc,
Ella Arensman, PhD, Heidi Hjelmeland, Dr Polit, Unni Bille-Brahe, MA,
Paul Corcoran, MSc, Diego De Leo, MD, PhD, FRANZCP,
Sandor Fekete, MD, PhD, Onja Grad, PhD, Christian Haring, PhD,
Ad J.F.M. Kerkhof, PhD, Jouko Lonnqvist, MD, PhD, Konrad Michel, MD,
Ellinor Salander Renberg, PhD, Armin Schmidtke, MD, PhD,
Cornelius van Heeringen, MD, PhD, and Danuta Wasserman, MD, PhD
The associations between life events in the 12 months preceding an episode
of self-poisoning resulting in hospital attendance (the index episode), and the suicide intent of this episode were compared in individuals for whom the index episode was their first, episode and in individuals in whom it was a recurrence of
DSH. Results indicated a significant interaction between independent life events,
repetition status, and gender in the prediction of suicide intent, the association
between life events and intent being moderated by repetition status in women
only. The results provide preliminary evidence to suggest the presence of a sui-
Catherine Crane, D Phil, J. Mark. G. Williams, DSc, and Keith Hawton, DSc, are with the
Centre for Suicide Research, University of Oxford Department of Psychiatry; Ella Arensman, PhD, is
with the National Suicide Research Foundation, Cork, Ireland; Heidi Hjelmeland, Dr Polit, is with
the Department of Social Work and Health Science, Norwegian University of Science and Technology,
Trondheim, and Division of Mental Health, Norwegian Institute of Public Health, Oslo, Norway; Unni
Bille-Brahe, MA, was with Centre for Suicidological Research, Odense, Denmark; Paul Corcoran,
MSc, is with the National Suicide Research Foundation, Cork, Ireland; Diego De Leo, MD, PhD,
FRANZCP, is with the University of Padua, Italy (now at Australian Institute for Suicide Research and
Prevention, Griffith University, Australia); Sandor Fekete, MD, PhD, is with the Department of Psychiatry, University of Pecs, Hungary; Onja Grad, PhD, is with the University Psychiatric Hospital, Ljubljana,
Slovenia; Christian Haring, PhD, is with the Psychiatric State Hospital Hall in Tirol, Austria; Ad J.F.M.
Kerkhof, PhD, is with the Department of Clinical Psychology, Vrije Universiteit, The Netherlands; Jouko
Lonnqvist, MD, PhD, is with National Public Health Institute, Department of Mental Health and Alcohol
Research and Department of Psychiatry, University of Helsinki, Finland; Konrad Michel, MD, is with
the University Psychiatric Services, University of Bern, Switzerland; Ellinor Salander Renberg, PhD, is
with the Department of Clinical Sciences, Division of Psychiatry, Umea University, Sweden; Armin
Schmidtke, MD, PhD, is with Department of Clinical Psychology, Clinic for Psychiatry and Psychotherapy, University of Wurzburg, Germany; Cornelius van Heeringen, MD, PhD, is with the Unit for Suicide Research, Department of Psychiarty,Ghent University, Belgium; and Danuta Wasserman, MD, PhD,
is with Centre for Suicide Research and Prevention, Karolinska Institute, Sweden.
The authors wish to thank the many patients and research colleagues who participated in this
research. This research was supported in part by the Wellcome Trust GR067797.
Address correspondence to Catherine Crane, Department of Psychiatry, Warneford Hospital,
Warneford Lane, Oxford, OX3 7JX, UK. E-mail: Catherine.crane@psych.ox.ac.uk
368
Life Events and Suicide Intent
cidal process in women, in which the impact of negative life events on suicide
intent diminishes across episodes.
Approximately half of all individuals presenting to hospital following an episode of deliberate self-harm (DSH; self-poisoning or selfinjury) have a history of previous self-harm
(Sakinofsky, 2000), with 16% of patients engaging in a further episode within the next
12 months, rising to approximately 23% after
four years or more (Owens, Horrocks, &
House, 2002). Rates of repetition are higher
in individuals with a previous history of DSH
at index than in those without (e.g., Wilkinson & Smeeton, 1987). Despite this fact, relatively few studies have focused specifically
on changes in the nature of DSH across episodes. For example, although stressful life
events and social problems are prominent in
the recent histories of suicidal individuals,
and although there have been several controlled trials of problem-solving therapies for
deliberate self-harm patients (see Townsend
et al., 2001), it is not clear whether the impact or significance of negative life events
and problems changes over consecutive episodes of self-harming behavior1 (e.g., Arensman, McAuliffe, Corcoran, & Perry, 2004).
1. There is considerable debate about the
most appropriate way to categorize and refer to
self-harming and suicidal behaviors. For the purposes of the current study the focus is on episodes
of self-poisoning which fall within the definition
of parasuicide adopted by the World Health Organization (1986). This defines parasuicide as: “an
act with non-fatal outcome in which an individual
deliberately initiates a non-habitual behaviour
that, without intervention from others, will cause
self-harm, or deliberately ingests a substance in
excess of the prescribed or generally recognized
therapeutic dosage, and which is aimed at realizing changes which the subject desired, via the actual or expected physical consequences.” This definition, while excluding habitual self-damaging
behaviors, acknowledges the inherent difficulties
in making assumptions about the often mixed motivations and intent of any given act of self-harm.
Because the term deliberate self-harm has superseded parasuicide as the preferred term in the liter-
One possibility is that episodes of selfharm bring about neurobiological, emotional, cognitive, or interpersonal changes
that directly increase risk of recurrence, a
phenomenon referred to as kindling, or ‘episode sensitization’ (Post, 1992; Segal, Williams, Teasdale, & Gemar, 1996). In recent
years a large number of studies have examined episode sensitization processes in major
depressive disorder (MDD: e.g., Post, 1992;
Segal et al., 1996; Kendler, Karkowski, &
Prescott, 1999; Kendler, Thornton, & Gardner, 2000, 2001; Kendler, Thornton, & Prescott, 2001). These studies have indicated that
previous episodes of MDD contribute directly to an individual’s risk of recurrence
(Kessing, Andersen, & Andersen, 2000), with
stressful life events appearing to be more important in triggering first depressive episodes
than those occurring later in the course of
the disorder (e.g., Mitchell, Parker, Gladstone, Wilhelm, & Austin, 2003; Lewinsohn,
Allen, Seeley, & Gotlib, 1999). The majority
of DSH patients are suffering from an affective disorder at the time of presentation to
hospital (e.g., Haw, Hawton, Housten, &
Townsend, 2001; Suominen et al., 1991) and
it is possible that a similar process of episode
sensitization may take place across episodes
of DSH. Indeed, Joiner (2002) has made this
point, suggesting that while enduring predispositions (e.g., genetic or temperamental factors) contribute to ongoing risk of suicidal
behavior, episodes of deliberate self-harm
may themselves increase risk of repetition, in
part as a result of cognitive sensitization (the
increased activity and accessibility of suiciderelated cognitive structures across consecutive episodes).
Drawing on this work it is hypothesized that with each suicidal crisis the associations will be strengthened between low
ature this is adopted, where appropriate, throughout the paper.
Crane et al.
mood, feelings of hopelessness, and the desire to engage in self-harming behavior. As a
consequence, although early episodes of DSH
may occur in response to the external stress
of major negative life events and problems,
later episodes may be more easily triggered,
with relatively minor hassles and/or endogenous cognitive and affective changes gaining
the capacity to reactive suicidal ideation and
deliberate self-harm (see Joiner & Rudd,
2000; Williams, Barnhofer, Crane, & Beck,
2005). Such a process would result in a diminishing association between major life
events and deliberate self-harm across episodes, the association being less pronounced
in those with prior episodes than in those engaging in DSH for the first time.
There have been relatively few empirical studies examining episode sensitization in
deliberate self-harm and the findings of those
that have been conducted are mixed. An early
study (Clark, Gibbons, Fawcett, & Scheftner,
1989) concluded that a heterogeneity model,
which is based on the assumption that individuals differ in their propensity towards suicidal behavior as the result of stable predispositions, was sufficient to explain data on
repetition of suicide attempts in a mixed psychiatric sample, arguing against the presence
of episode sensitization. However, no account was taken of the number of episodes
participants had engaged in prior to their entry into the study, a factor which may critically influence the likelihood of observing
episode sensitization (see Kendler et al.,
2000, for discussion in the case of MDD).
Two more recent studies have provided more support for the episode sensitisation hypothesis. The first ( Joiner & Rudd,
2000) indicated that multiple attempters
showed a weaker association between life
event stress and suicidal ideation than those
who had harmed themselves for the first time
or who had experienced suicidal ideation
only. The second, a 12 month prospective
study, examined the occurrence of death ideation, death wishes, suicide contemplation,
and deliberate self-harm in a large general
population sample (Neeleman, Graaf, & Vol-
369
leberg, 2004). Negative life events were more
closely associated with the occurrence of suicidality in individuals without a prior history
of suicidal ideation or behavior, whereas
mental illness was more closely associated
with suicidality in individuals with a prior
history. Both these studies broadly support
the idea of a suicidal process in which suicidal
ideation and self-harming behavior become
more autonomous of external negative life
events over time. However, the first study
grouped together suicide ideators and individuals who had made a single suicide attempt, reporting that findings were more
pronounced if comparison was restricted to
the distinction between ideators and multiple
attempters ( Joiner & Rudd, 2000). Therefore, it is not clear whether single attempters
and multiple attempters actually differ significantly in the association between life events
and suicide intent. Secondly, the sample on
which this study was based (predominantly
young, male, military personnel) is unusual
and it is not clear whether a similar effect of
prior history of deliberate self-harm on the
association between life events and the severity of suicidality would be observed in other
populations.
The current paper utilizes data on recent life events and the suicide intent of an
episode of DSH resulting in presentation to
hospital. The data was collected as part of a
large multi-center study of psychological and
socio-demographic factors associated with
repetition of deliberate self-harm (the European Parasuicide Study Interview Schedule,
EPSIS). As part of this interview participants
completed the Beck Suicide Intent Scale
(Beck, Schuyler, & Herman, 1974), which assesses the severity of an episode of DSH, in
terms of the objective circumstances and the
subjective wish to die. Episode sensitization
can be operationalized as a diminishing association across episodes between the severity
of the DSH episode (level of suicide intent)
and the burden of negative life events present
at the time, or alternatively as a difference in
the strength of such an association between
first ever DSH patients and repeaters. In or-
370
der to examine the hypothesis that DSH is
subject to episode sensitization processes it is
necessary to focus analyses on individuals in
whom there is likely to be a relatively strong
association between life events and degree of
suicide intent at the outset (to allow for potential diminution of the association across
episodes). Post (1992) and Kendler, Thornton, and Gardner (2001) suggest that some
people will be prekindled showing lower levels
of association between negative life events
and suicide intent from the outset. In patients
with major depression this pre-kindling is observed in those with a past history of trauma
(e.g., Hammen, Henry, & Daley, 2000; Beatson & Taryan, 2003), where first onsets of
depression are less closely associated with
major life events. Deliberate self-harm patients with high rates of past trauma (e.g.,
borderline patients) show a similar pattern
(e.g., Stanley, Brodsky, Groves, & Mann,
2004). Alcohol and substance misuse, which
increase impulsivity of self-harm episodes
(e.g., Suominen, Isometsa, Ostamo, & Lonnqvist, 1997; Borges, Walters, & Kessler, 2000),
are also likely to increase the importance of
immediate contextual factors and limit the
contribution of major negative life events. As
a result, we examined patients without a history of sexual trauma or co-morbid alcohol/
substance misuse at the time of self-poisoning, in order to ensure that analyses are focused on those in whom episode sensitisation
(as operationalized in this study) is most
likely to be observed.2,3
2. Sexual abuse/assault was defined as a report of one of the following experiences in childhood or later life (>12 months prior to index):
forced sexual intercourse by parents, forced sexual
intercourse by siblings, rape by another person,
being forced to endure other sexual activity by
parents, and being forced to endure other sexual
activity by siblings, forced sexual intercourse by a
partner, other forced sexual activity by a partner,
having to prostitute oneself, and being forced into
prostitution by a partner.
3. To check the validity of this assumption
we also examined the data on all self-poisoners,
including those with current alcohol or substance
abuse and a history of sexual trauma. In this
broader sample there was no significant associa-
Life Events and Suicide Intent
Data on the occurrence of a large
number of life events and difficulties were
collected from study participants. Some life
events and difficulties may arise as a consequence of ongoing distress and mental illness
(e.g., loneliness, problems in intimate relationships, unemployment), as a consequence
of temperamental factors (e.g., impulsivity,
aggression, neuroticism) which overlap with
predispositions to psychiatric disorder (see
Kendler et al., 1999 for a discussion), or as a
consequence of factors related directly to an
individual’s own risk of recurrent deliberate
self-harm (for example, the psychiatric illness
of family members, e.g., Brent, Bridge, Johnson, & Connolly 1996; Johnson, Brent,
Bridge, & Connolly, 1998). Such events potentially confound an individual’s history and
risk of repeating deliberate self-harm, with
the way that self-harm manifests itself (for
example, the lethality or impulsivity of an episode). Therefore, we conducted analyses examining relations between suicide intent and
the experience of life events that can be considered to occur relatively independently of an
individual’s mental state or risk of self-harm:
deaths and physical illnesses in the close social network.4
METHOD
Participants
All participants were interviewed as
part of the WHO/EURO Multicentre Study
on Parasuicide, initiated by the World Health
Organization (Platt et al., 1992). Data was
collected from 16 of the centers which participated in the repetition-prediction part of
tion between suicide intent and number of life
events (total, independent) in either first episode
patients or in repeaters.
4. We also explored the association between suicide intent and a total life events score.
However, no significant associations between this
variable and suicide intent in either first ever
DSH patients or repeaters were identified, so for
clarity these analyses are omitted.
Crane et al.
the study (EPSIS Study) (Bern, Switzerland;
Cork, Ireland; Emilia-Romagna, Italy; Ghent,
Belgium; Hall, Austria; Helsinki, Finland;
Leiden, Netherlands; Ljubljana, Slovenia;
Odense, Denmark; Oxford, England; Padova, Italy; Pecs, Hungary; Sør-Trøndelag,
Norway; Stockholm, Sweden and Umea,
Sweden; Wuerzburg, Germany). Studies
have confirmed the similarities between centers in psychological characteristics such as
intentions involved in the parasuicide (DSH)
(e.g., Hjelmeland et al., 2002), supporting the
validity of combining data across centers.
Further details of the participating centers
can be found in a report on the WHO/
EURO Multicentre Study published by the
World Health Organization (1999) and a
summary of recent research findings in De
Leo, Bille-Brahe, Kerkhof, and Schmidtke
(2004). Readers should note that, given the
large scale of this study, a number of publications have arisen from the data collected.
However, no previous studies have examined
episode sensitization processes or the association between life events and suicide intent as
a function of history of deliberate self-harm.
General data on the occurrence of life events
and their association to measures of mood
and self-esteem can be found in Arensman
and Kerkhof (2004).
Individuals were suitable for inclusion
in the EPSIS study if they presented to hospital having engaged in an act fulfilling the
WHO/EURO definition of parasuicide (the
Index episode). Individuals fulfilling these
criteria took part in interviews, usually within
one week of their act. In some centers all eligible individuals were approached. In others
a random sample of potential eligible patients
were identified and invited to participate.
Further information concerning the methods
of the EPSIS study can be found in BilleBrahe et al., (1996) and Bille-Brahe, Schmidtke,
Kerkhof, and De Leo (2004).
Measures
Participants completed a structured interview schedule, the European Parasuicide
Interview Schedule (EPSIS), detailed in Ker-
371
khof and colleagues (1993), which incorporated several self-report questionnaires. Only
the aspects of the interview schedule relevant
to the current analyses are discussed below.
Demographic Information. Participants
provided information on their age, sex, marital status, level of education, and number of
children.
History of Deliberate Self-Harm. Participants provided information on the number of previous episodes of DSH they had
engaged in, based on their self-report. Only
episodes falling within the WHO definition
of parasuicide were counted. Because recollection of the total number and timing of
previous episodes of DSH may be subject to
bias, it was considered more likely that individuals would be accurate in their recall of
whether or not any episodes of DSH had occurred. For this reason most analyses use a
dichotomized variable to identify those with
and those without a previous episode of DSH.
Life Events. Individuals completed an
inventory to indicate (by ticking a box to say
yes or no to each item) whether or not they
had experienced each of a large number of
negative life events (related to parents, siblings, children, partners, other close social
ties, and events and difficulties relating only
to the individual themselves) in the 12
months preceding the episode of DSH that
resulted in their presentation to hospital and
inclusion in the study, the Index episode (see
Kerkhof et al., 1993, for further details). A
score was calculated which corresponded to
the number of independent life events experienced, of a possible eleven.5 These events,
deaths, and illnesses in the social network, as
well as being independent, are likely to be
experienced as severe and distressing to most
people.
Suicide Intent. Suicide Intent was
5. The independent life events were: death
of mother, death of father, parental chronic illness, death of a sibling, sibling’s chronic illness,
death of a partner, partner’s chronic illness, death
of participant’s child, child’s chronic illness, death
of another close person, participant’s own physical
illness.
372
Life Events and Suicide Intent
measured using the Beck Suicide Intent Scale
(Beck, Schuyler, & Herman, 1974). The
items refer both to objective features of the
suicide attempt (e.g., whether or not the individual communicated their wish to die to
others, whether precautions were taken
against discovery) and level of subjective intent (e.g., whether the individual believed
that the actions they took were likely to result in death, whether the individual considered that they had made a serious attempt to
end their life). The internal reliability in the
current sample was good (α = .86).
Hopelessness. Participants’ levels of
hopelessness were assessed with the Beck
Hopelessness Scale (Beck, Weissman, Lester,
& Trexler, 1974). Internal reliability in the
current sample was good (α = .83).
Depression. The Beck Depression Inventory (Beck, Ward, Mendelson, Mock, &
Erbauch, 1961) was used to assess the level
of depression over the previous two weeks.
Internal reliability in the current sample was
good (α = .80).
Statistical Analysis
The primary research question, the
nature of the association between suicide intent and life events in first-ever DSH patients
and repeaters, was explored using hierarchical linear regression. Because some of the
variables of interest (e.g., life events) departed from normality, both parametric
(Pearson’s) and nonparametric (Spearman’s)
correlations were computed to confirm that
the findings were the same in each case. All
analyses were conducted using SPSS version 13.
RESULTS
Participant Characteristics
Data was collected from 609 individuals (243 males, 366 females) who presented
to hospital following an episode of selfpoisoning and did not (a) self-report current
problems with alcohol, (b) report use of her-
oin or cocaine in the preceding week, and (c)
report a history of sexual victimization. The
mean age of participants was 36.0 years
(SD = 15.4). For a total of 285 individuals
(47%, 111 males, 174 females) the index
DSH episode was their first (first-evers), with
323 individuals (53%; 132 males, 191 females) reporting a prior history of DSH (repeaters). One individual had missing data
concerning number of previous episodes of
DSH. The proportions of females was the
same amongst first-evers and repeaters, χ2 (1) =
.23, p > .60. Amongst repeaters, 153 individuals (47%; 61 males, 92 females) reported
one previous episode, 100 individuals (31%;
43 males, 57 females) two to four previous
episodes, and 70 individuals (21%; 28 males,
42 females) five or more previous episodes.
First-evers and repeaters did not differ in
number of children, χ2 (4) = 7.18, p > .10,
marital status, χ2 (4) = 7.37, p > .10, or educational level χ2 (2) = 1.89, p > .30.
Initial Analyses
Level of Suicide Intent. The mean suicide intent scale score was M = 14.0 (SD =
7.0, range 0–30), which represents a relatively high level of intent which may reflect
the focus on self-poisoners, since some other
forms of self-harm, e.g., self-cutting, appear
to be associated with lower levels of suicide
intent (Hawton, Harriss, Simkin, Bale, &
Bond, 2004). Univariate analysis of variance
(ANOVA) indicated a main effect of repetition status, F (1, 601) = 8.23, p < .01, with
first-evers (M = 13.1, SD = 6.9) having lower
mean suicide intent scores than repeaters
(M = 14.8, SD = 7.0). There was also a main
effect of gender, F (1, 601) = 5.73, p < .05,
with males having higher mean suicide intent
scores (M = 14.9, SD = 7.1) than females
(M = 13.5, SD = 6.9), although the absolute
difference in scores was small. There was no
interaction between gender and repetition
status.
Level of Depression. The mean BDI
score was 21.8 (SD = 11.5, range 0–54), representing depressive symptoms of moderate
severity. A univariate ANOVA indicated a
Crane et al.
significant main effect of repetition status, F
(1, 584) = 15.13, p < .001, resulting from
higher levels of depression in repeaters (M =
23.4, SD = 11.2) than in first-evers (M = 19.9,
SD = 11.6). There was no significant main effect of gender and no interaction between
gender and repetition status.
Level of Hopelessness. The mean level
of hopelessness was 12.0 (SD = 4.6, range 3–
21), which is relatively high and somewhat
higher than the level reported for samples of
suicide ideators and attempters (Beck &
Steer, 1993). This may reflect the exclusion
of substance misusers since Beck and Steer
report that levels of hopelessness are much
lower in these groups. Univariate ANOVA
again indicated a main effect of repetition
status, F (1, 572) = 13.74, p < .001, resulting
from lower levels of hopelessness in firstevers (M = 11.3, SD = 4.5) than repeaters
(M = 12.7, SD = 4.7). There was no main effect of gender and no interaction between
gender and repetition status.
Number of Independent Life Events.
The mean number of independent life events
(deaths/illness) reported by participants was
relatively low (M = 0.8, SD = 0.9, range = 0
to 5), reflecting the relative rarity of these serious events. Univariate ANOVA identified
no significant main effects or interactions, indicating that the occurrence of these life
events was similar in individuals with and
without prior episodes of deliberate selfharm, as would be expected if such events occur independently of an individual’s mental
state.
Correlations
Correlations were calculated between
the above variables and are shown in Table 1.
Across the sample as a whole there was no
significant association between the measures
of independent negative life events and suicide intent. The sample was then divided into
individuals presenting with their first episode
of self-poisoning and individuals who had a
prior history of deliberate self-harm at index.
A significant (if small) association between
independent life events and suicide intent
373
was observed in first-evers, r(283) = .15 p =
.01; rs(283) = .16, p < .01, but not in repeaters, r(320) = −.07, p = ns; rs (320) = −.04, p =
ns. A comparison of the strengths of the correlation coefficients between independent life
events and suicide intent in the two groups
yielded a significant difference, Z = 2.68, p <
.004, which was investigated further using hierarchical regression.
Regression Analysis
To examine whether repetition status
moderated the association between independent life events and suicide intent, a hierarchical multiple regression analysis was conducted with suicide intent entered as the
dependent variable. Number of life events
and repetition status were entered at Step 1.
Gender and hopelessness were also entered
at Step 1 due to their significant relationships
with level of suicide intent.6 At Step 2, the
two-way interaction, Life Events × Repetition Status was added. At Step 3, the threeway interaction Life Events × Gender × Repetition Status was added (see Table 2 for a
summary of the regression model). The results indicated that at Step 1, the overall regression model was significant, R 2 = .05, F(4,
567) = 6.94, p < .001, with repetition status,
gender, and hopelessness, but not life events
entering as significant predictors of suicide
intent. At Step 2, following the inclusion of
the two-way interaction, Life Events × Repetition status there was a significant improvement in the model, ∆R 2 = .01, ∆F(1) = 6.82,
p < .01, with repetition status, hopelessness,
gender, life events, and the two-way interaction all significantly contributing to the prediction of suicide intent. At Step 3, there was
a further significant improvement in variance
explained, ∆R 2 = .013, ∆F(1) = 7.92, p = .005.
The three-way interaction between Life
Events × Gender × Repetition Status entered
6. Due to the high degree of association
between BHS and BDI scores only one of these
measures were entered into the regression analysis. Since BHS score showed the stronger correlation with suicide intent this variable was selected.
374
Life Events and Suicide Intent
TABLE 1
Intercorrelations Between Life Events, Suicide Intent, Depression,
and Hopelessness
Spearman’s r
1. Independent events
2. Suicide Intent
3. BDI
4. BHS
Pearson’s r
1
2
3
4
1
—
.04
—
.13**
.13**
.07
.17***
.59***
2
3
54
.05
.13***
.13**
.10**
.18***
.59***
*p ≤ .05, **p ≤ .01, ***p ≤ .001
as a significant predictor of suicide intent,
overriding the contribution of the two-way
interaction, which was no longer significant.
Hopelessness, gender, and life events remained as significant predictors of suicide intent. Overall, the model accounted for approximately 7% of the variance in suicide
intent and was highly significant, F(6, 565) =
7.20, p < .001.
Correlations Clarifying
Three-Way Interaction
To clarify the nature of the three-way
interaction between repetition status, gender,
and life events in the prediction of suicide intent, the correlations between number of life
events and suicide intent were computed separately for men and women presenting with
TABLE 2
Summary of Hierarchical Regression Analysis
for Prediction of Suicide Intent
Step 1
Step 2
Step 3
Variable
B
SE B
β
Hopelessness
Life Events (L)
Repetition Status (S)
Gender (G)
Hopelessness
Life Events
Repetition Status
Gender
L×S
Hopelessness
Life Events
Repetition Status
Gender
L×S
L×S×G
.22
.29
1.38
−1.28
.21
1.15
1.39
−1.41
−1.59
.20
1.15
1.38
−1.35
−.28
−2.34
.06
.31
.58
.58
.06
.45
.57
.58
.61
.06
.45
.57
.58
.77
.83
.15***
.04
.10*
−.09*
.14**
.16**
.10*
−.10*
−.16**
.13***
.16**
.10*
−.10*
−.03
−.17**
Note. Step 1, ∆R 2 = .047, ∆ F (4, 567) = 6.94, p <
.001; Step 2, ∆ R 2 = .011, ∆ F (2, 566) = 6.82, p = .009;
Step 3, ∆ R 2 = .013, ∆ F (1, 564) = 7.92, p = .005.
*p < .05; **p ≤ .01, ***p ≤ .001
Crane et al.
375
a first episode and presenting with a repetition. As shown in Table 3, the association
between suicide intent and number of life
events in men was similar for first episodes
and for repetitions, being positive but nonsignificant in both cases. In contrast, in
women, number of life events was associated
with higher levels of suicide intent in those
presenting with a first episode of DSH and
lower levels of suicide intent in those presenting with a repetition. As shown in Table 3 there appears to be a change in the
strength and direction of the correlation between life events and suicide intent from
those women in whom the index episode
was their first, to those with five or more
episodes (see Table 4).
DISCUSSION
There has been increasing interest in
the possibility that episode sensitization processes are a feature of repeated deliberate
self-harm (e.g., Joiner, 2002). It is suggested
that across episodes minor stressors and endogenous factors (such as spontaneous fluctuations in sad mood) become capable of
triggering suicidal ideation and self-harming
behavior (e.g., Williams et al., 2005). If this
occurs it would be expected that the association between major life events and level of
suicide intent would diminish across epi-
TABLE 3
Intercorrelations Between Suicide Intent and Life
Events in Men and Women with and without
a Prior History of Deliberate Self-Harm
Men
r
rs
Women
First
ever
Repeater
First
ever
Repeater
.14
.14
.14
.17
.19*
.20**
−.22**
−.20**
*p < .05; **p < .01
TABLE 4
Intercorrelations Between Suicide Intent and Life
Events Score in Women According to Number
of Prior Episodes of Deliberate Self-Harm
Number of prior episodes
of DSH
Coefficient
r
rs
0
1
2–4
5+
.19*
.20**
−.20
−.18
−.21
−.21
−.36*
−.30*
*p ≤ .05 **p ≤ .01
sodes. A first step in investigating this issue is
to establish whether there are in fact significant differences between first-ever episodes
and repeat episodes in this regard. The current study identified limited evidence in favor
of the episode sensitization hypothesis. A significant three-way interaction was identified
between independent life events, repetition
status, and gender in the prediction of suicide
intent, with the association between independent life events and suicide intent being
moderated by participants’ repetition status
(first-ever, repeater), but only in women.
Women presenting with a first episode of deliberate self-harm showed a significant positive correlation between life events and suicide intent, while female repeaters showed a
significant negative correlation. When female
repeaters with one repeat, two to four repeats
and five or more repeats were examine separately, the relationship between number of
independent negative life events and suicide
intent was negative in all cases but significant
only for those with five or more prior episodes.
The levels of association between the
variables of interest in this study were low
and so the results should be interpreted with
caution. However, if taken at face value, the
findings have at least two possible interpretations. They may indicate the presence of a
suicidal process in women in which the impact of negative life events on suicide intent
diminishes across episodes, or suggest that
376
there are stable differences between women
engaging in single episodes of DSH and
women engaging in repeated episodes in the
association between independent negative
life events and the severity of a suicidal episode.
A number of issues should be borne in
mind when considering the results of this
study. First, as discussed above, although statistically significant and similar to levels observed in other comparable studies (e.g.,
Joiner & Rudd, 2000), the results of this
study are likely to have limited clinical significance. Although the relationship between
suicide intent and independent life events
may differ in women as a function of history
of DSH, a large number of other factors
must also contribute to the severity of individual episodes since the majority of the variance in suicide intent remained unexplained.
Thus, although the results indicate that the
relationship between life events and suicide
intent may alter subtly across episodes at a
population level, they do not provide strong
evidence to suggest a suicidal process observable at the level of the individual patient.
Second, we excluded from the study
individuals with current alcohol problems/
drug use or a history of past sexual trauma.
These individuals constitute a significant
proportion of the deliberate self-harm patients presenting to hospital, but were excluded on theoretical grounds, since it is
likely that major negative life events are less
important as determinants of suicide intent
from the outset. As a consequence, the generalizability of the findings is limited to a
sub-group of self-poisoners without these
characteristics.
Third, information on Axis I and Axis
II diagnoses was not available, preventing the
inclusion of diagnostic variables as potential
moderators of the life events-suicide intent
association. It is likely that different diagnostic sub-groups within the DSH population
would show differing patterns of change in
association between life events and suicide
intent across episodes. For example, a diminishing contribution of major life events to
Life Events and Suicide Intent
suicide intent might be most relevant to
those whose DSH occurs in the context of
uncomplicated major depression, where episode sensitization processes have most frequently been observed, rather than to DSH
patients as a whole.
Finally, a number of correlation analyses were conducted to explore the nature of
the interaction between gender, life events,
and repetition status in the prediction of suicide intent and it should be noted that if a
Bonferroni correction was applied conservatively, to reduce the family-wise error rate to
below alpha = .05 across all analyses conducted, that association would no longer be
statistically significant. Therefore, the results
should be interpreted cautiously.
Despite these limitations, the current
study also has strengths. It reports data on a
large sample drawn from across Europe, and
represents the first study to examine the extent to which the association between life
events and the severity of a verified episode
of DSH varies with number of prior DSH
episodes. The fact that the study provides
only limited evidence of sensitization is itself
important, since there has been increasing
speculation that sensitization processes may
contribute to DSH (e.g., Joiner, 2002, Joiner
& Rudd, 2000; Williams et al., 2005). However, it is likely to be premature to rule out
the contribution of episode sensitization processes in DSH without further studies. In
particular, future research might consider not
only the occurrence of life events or problems and their relationship to episodes of
DSH, but also the extent to which problemsolving skills protect against further episodes of DSH in first episode patients and
repeaters. Such studies would have the potential to shed light not only onto the nature
of the suicidal process, but would also have
important implications for treatment, indicating for whom problem-solving interventions (focused on resolution of external
problems) are likely to be most efficacious
and in whom treatment might be better focused on the ability to tolerate aversive internal states.
Crane et al.
377
REFERENCES
Arensman, E., & Kerkhof, A. (2004).
Negative life events and non-fatal suicidal behaviour. In De Leo, D., Bille-Brahe, U., Kerkhof, A.,
& Schmidtke, a. (Eds). Suicidal behaviour: Theories
and research findings. Gottingen, Germany:
Hogrefer & Huber.
Arensman, E., McAuliffe, C., Corcoran,
P., & Perry, I. (2004). Comment on Tyrer et al.,
2003. Psychological Medicine, 34, 1143–1144.
Beatson, J., & Taryan, S. (2003). Predisposition to depression: The role of attachment.
Australian and New Zealand Journal of Psychiatry,
37, 219–225.
Beck, A. T., Schuyler, D., & Herman, J.
(1974). Development of suicide intent scales. In A.
T. Beck, H.L.P. Resnik, & D. J. Lettieri (Eds). The
prediction of suicide. Bowie, Maryland: Charles
Press.
Beck, A. T., & Steer, R. A. (1993). Manual
for the Beck Hopelessness Scale. San Antonio: The
Psychological Corporation.
Beck, A. T., Ward, C. H., Mendelson,
M., Mock, J., & Erbauch, J. (1961). An inventory
for measuring depression. Archives of General Psychiatry, 4, 561–571.
Beck, A. T., Weissman, A., Lester, D., &
Trexler, L. (1974). The measurement of pessimism: The Hopelessness Scale. Journal of Consulting and Clinical Psychology, 6, 861–865.
Bille-Brahe, U., Kerkhof, A., De Leo,
D., Schmidke, A., Crepet, P., Lonnqvist, J., et
al. (1996). A repetition-prediction study on European parasuicide populations. Crisis, 17, 22–31.
Bille-Brahe, U., Schmidtke, A., Kerkhof, A., & De Leo, D. (2004). The WHO/Euro
multi-centre study on suicidal behaviour: Its background, history, aims and design. In De Leo, D.,
Bille-Brahe, U., Kerkhof, A.J.F.M., & Schmidtke,
A. Eds. Suicidal behaviour. Theories and research
findings. Goettingen, Germany: Hogrefe & Huber
Publishers.
Borges, G., Walters, E. E., & Kessler,
R. C. (2000). Associations of substance use, abuse,
and dependence with subsequent suicidal behaviour. American Journal of Epidemiology, 151, 781.
Brent, D. A., Bridge, J., Johnson, B. A.,
& Connollly, J. (1996). Suicidal behaviour runs
in families. Archives of General Psychiatry, 53,
1145–1152.
Clark, D. C., Gibbons, R. D., Fawcett,
J., & Scheftner, W. A. (1989). What is the
mechanism by which suicide attempts predispose
to later suicide attempts? A mathematical model.
Journal of Abnormal Psychology, 98, 42–49.
De Leo, D., Bille-Brahe, U., Kerkhof,
A., & Schmidtke, A. (Eds.). (2004). Suicidal behaviour: Theories and research findings. Gottingen,
Germany: Hogrefer & Huber.
Hammen, C., Henry, R., & Daley, S. E.
(2000). Depression and sensitization to stressors
among young women as a function of childhood
adversity. Journal of Consulting and Clinical Psychology, 68, 782–787.
Haw, C., Hawton, K., Houston, K., &
Townsend, E. (2001). Psychiatric and personality
disorders in deliberate self-harm patients. British
Journal of Psychiatry, 178, 97–104.
Hawton, K., Harriss, L., Simkin, S.,
Bale, E., & Bond, A. (2004). Self-cutting: Patient
characteristics compared to self-poisoners. Suicide
and Life-Threatening Behavior, 34, 199–207.
Hjelmeland, H., Hawton, K., Nordvik,
A. H., Bille-Brahe, U., De Leo, D., Fekete, S.,
et al. (2002). Why people engage in parasuicide:
A cross-cultural study of intentions. Suicide and
Life-Threatening Behaviour, 32, 380–393.
Johnson, B. A., Brent, D. A., Bridge, J.,
& Connolly, J. (1998). The familial aggregation
of adolescent suicide attempts. Acta Psychiatrica
Scandinavica, 97, 18–24.
Joiner, T. E. (2002). The trajectory of suicidal behaviour over time. Suicide and Life-Threatening Behaviour, 32, 33–41.
Joiner, T. E., & Rudd, M. D. (2000). Intensity and duration of suicidal crises vary as a
function of previous suicide attempts and negative
life events. Journal of Consulting and Clinical Psychology, 68, 909–916.
Kendler, K. S., Karkowski, L. M., &
Prescott, C. A. (1999). Causal relationship between stressful life events and the onset of major
depression. American Journal of Psychiatry, 156,
837–841.
Kendler, K. S., Thornton, L. M., &
Gardner, C. O. (2000). Stressful life events and
previous episodes in the etiology of major depression in women: An evaluation of the kindling hypothesis. American Journal of Psychiatry, 157,
1243–1251.
Kendler, K. S., Thornton, L. M., &
Gardner, C. O. (2001). Genetic risk, number of
previous depressive episodes and stressful life
events in predicting onset of major depression.
American Journal of Psychiatry, 158, 582–586.
Kendler, K. S., Thornton, L. M., &
Prescott, C. A. (2001). Gender differences in the
rates of exposure to stressful life events and sensitivity to their depressogenic effects. American
Journal of Psychiatry, 158, 587–593.
Kerkhof, A.J.F.M., Bernasco, W., BilleBrahe, U., Platt, S., & Schmidtke, A. (1993). In
378
Bille-Brahe, U., Bjerke, T., Crepet, P., De Leo,
D., Haring, C., Hawton, K., et al. (Eds.). WHO/
EURO Multicentre Study on Parasuicide. Facts and
Figures (pp. 53–118). Copenhagen: World Health
Organisation Regional Office for Europe.
Kessing, L. V., Andersen, E. W., & Andersen, P. K. (2000). Predictors of recurrence in
affective disorder—analyses accounting for individual heterogeneity. Journal of Affective Disorders,
57, 139–145.
Lewinsohn, P. M., Allen, N. B., Seeley,
J. R., & Gotlib, I. A. (1999). First onset versus
recurrence in depression: Differential processes of
psychosocial risk. Journal of Abnormal Psychology,
108, 483–489.
Mitchell, P. B., Parker, G. B., Gladstone, G. L., Wilhelm, K., & Austin, M. V.
(2003). Severity of stressful life events in first and
subsequent episodes of depression: Relevance of
depressive sub-type. Journal of Affective Disorders,
73, 245–252.
Neeleman, J., De Graaf, R., & Volleberg, W. (2004). The suicidal process: Prospective
comparison between early and later stages. Journal
of Affective Disorders, 82, 43–52.
Owens, D., Horrocks, J., & House, A.
(2002). Fatal and nonfatal repetition of self-harm.
British Journal of Psychiatry, 181, 193–199.
Platt, S., Bille-Brahe, U., Kerkhof, A.,
Schmidtke, A., Bjerke, T., Crepet, P., et al.
(1992). Parasuicide in Europe. The WHO EURO
Multicentre on Parasuicide 1. Introduction and
preliminary analysis for 1989. Acta Psychiatrica
Scandinavica, 85, 97–104.
Post, R. M. (1992). Transduction of psychosocial stress into the neurobiology of recurrent
affective disorder. American Journal of Psychiatry,
149, 999–1009.
Sakinofsky, I. (2000). Repetition of suicidal behaviour. In K. Hawton & K. van Heeringen (Eds.), The International Handbook of Suicide
and Attempted Suicide (pp. 385–404), Chichester,
England: John Wiley & Sons, Ltd.
Segal, Z. V., Williams, J.M.G., Teasdale, J. D., & Gemar, M. (1996). A cognitive sci-
Life Events and Suicide Intent
ence perspective on kindling and sensitization in
recurrent affective disorder. Psychological Medicine,
26, 371–380.
Stanley, B., Brodsky, B., Groves, S., &
Mann, J. J. (2004, August). Distinguishing suicidal
behaviour and precipitants in borderline personality
disorder. Paper presented at the 10th European
Symposium on Suicide and Suicidal Behaviour,
Copenhagen, Denmark.
Suominen, K., Henriksson, M., Suokas,
J., Isometsa, E., Ostamo, A., & Lonnqvist, J.
(1991). Mental disorders and co-morbidity in attempted suicide. Acta Psychiatrica Scandinavica, 94,
234–240.
Suominen, K., Isometsa, E., Ostamo, A.,
& Lonnqvist, J. (1997). Hopelessness, impulsiveness, and intent among suicide attempters with
major depression, alcohol dependence, or both.
Acta Psychiatrica Scandinavica, 96, 142–149.
Townsend, E., Hawton, K., Altman,
D. G., Arensman, E., Gunnel, D., Hazell, P., et
al. (2001). The efficacy of problem-solving treatments after deliberate self-harm: Meta-analysis of
randomized controlled trials with respect to depression, hopelessness, and improvement in problems. Psychological Medicine, 31, 979–988.
Wilkinson, G., & Smeeton, N. (1987).
The repetition of parasuicide in Edinburgh 1980–
1981. Social Psychiatry, 22, 14–19.
Willliams, J.M.G., Barnhofer, T., Crane,
C., & Beck, A. T. (2005). Problem solving deteriorates following mood challenge in formerly depressed patients with a history of suicidal ideation.
Journal of Abnormal Psychology, 114, 421–431.
World Health Organization. (1999).
WHO/EURO Multicentre Study on Parasuicide.
Facts and Figures. (2nd ed.). Copenhagen: World
Health Organization Regional Office for Europe.
World Health Organization. (1986).
Summary report, Working group in prevention practices in suicide and attempted suicide. Copenhagen:
WHO Regional Office for Europe.
Manuscript Received: December 27, 2005
Revision Accepted: November 1, 2006