Content uploaded by Julia J Rucklidge
Author content
All content in this area was uploaded by Julia J Rucklidge on Jan 26, 2015
Content may be subject to copyright.
ORIGINAL ARTICLE
The relationship between ADHD symptomatology and self-harm,
suicidal ideation, and suicidal behaviours in adults: a pilot study
Mairin R. Taylor •Joseph M. Boden •
Julia J. Rucklidge
Received: 11 November 2013 / Accepted: 26 April 2014 / Published online: 8 May 2014
ÓSpringer-Verlag Wien 2014
Abstract The aim of this study was to explore whether
individuals with attention-deficit/hyperactivity disorder
(ADHD) are at risk of harm over the lifespan due to
increased rates of self-harm, suicidal ideation and suicidal
behaviour, and whether this association is mediated by
psychosocial factors. Sixty-six adults (43 men, 23 women;
18–65 years) participated in this study involving clinical
interview and retrospective self-report measures of ADHD
symptoms, self-harm/suicidal behaviour, mental health
disorders, and coping style measures. Significant associa-
tions were found between ADHD symptom severity and
self-reported histories of self-harm behaviour, suicidal
ideation, and suicide attempts (all pvalues \.05). These
relationships between self-destructive behaviours and
ADHD symptom severity were found to be significantly
and differentially mediated by psychosocial variables (all
pvalues\.05) including comorbidity (mood, anxiety, drug,
and alcohol abuse disorders) and emotion-focussed coping
style. This study suggests that linkages between self-inju-
rious behaviour and ADHD symptomatology may be due
primarily to comorbid mental health disorders and emo-
tion-focussed coping. The identification of these mediating
factors and processes may potential pathways for
intervention in reducing suicide and self-harm risk amongst
those with ADHD symptoms.
Keywords ADHD Self-harm Suicide Comorbidity
Coping
Introduction
Attention-deficit/hyperactivity disorder (ADHD) is a
developmental disorder often persisting into adulthood
(Biederman et al. 2006) that is associated with poor
impulse control and comorbid psychopathology (Hes-
slinger et al. 2003). Those adults who continue to be sig-
nificantly affected by ADHD symptomatology are more
likely to engage in a number of high-risk behaviours such
as dangerous driving (Murphy and Barkley 1996; Ramos
Olazagasti et al. 2013), substance abuse (Biederman et al.
2006; Klein et al. 2012), and sexual risk-taking behaviour
(Barkley and Gordon 2002).
In addition to risk-taking behaviour, a number of studies
have demonstrated linkages between ADHD and self-
directed harmful behaviours in adults. In a review of the
literature, Impey and Heun (2012) concluded that there was
compelling evidence of positive relationships between
ADHD and risk to the self. In addition, a number of studies
have found significantly higher rates of suicidal ideation
amongst ADHD populations (Haavisto et al. 2005; Arias
et al. 2008), in addition to higher rates of suicide attempts
(Lam 2002; Westmoreland et al. 2010; Semiz et al. 2008).
Further, a number of studies have identified a relation-
ship between ADHD in young people and self-harm
behaviours. For example, Izutsu et al. (Izutsu et al. 2006)
found higher values of childhood hyperactivity amongst
those who reported self-harm behaviours in an adolescent
M. R. Taylor J. J. Rucklidge
Department of Psychology, University of Canterbury,
Christchurch, New Zealand
M. R. Taylor
LightBox Psychology Services, Christchurch, New Zealand
J. M. Boden (&)
Department of Psychological Medicine, Christchurch School of
Medicine and Health Sciences, University of Otago,
PO Box 4345, Christchurch 8140, New Zealand
e-mail: joseph.boden@otago.ac.nz
123
ADHD Atten Def Hyp Disord (2014) 6:303–312
DOI 10.1007/s12402-014-0139-9
population, and Galera et al. (2008) found that childhood
combined-type ADHD predicted suicidality in adolescent
boys. Barkley (2006) reported that adolescents with ADHD
had a significantly higher rate of attempted suicide than
control group members. Of that number, approximately
46 % required hospitalization, suggesting that the attempts
made tend to be serious. In addition, in a prospective study
of young adults who had been psychiatrically hospitalized,
Goldston et al. (2009) assessed the relationship between
persistence of psychiatric diagnoses and suicidal behav-
iours. The authors reported a relationship between con-
temporaneous diagnoses of ADHD and suicide attempts
amongst adolescents (whilst controlling for other disorders
and demographic factors), a relationship which strength-
ened into adulthood. Also, Chronis-Tuscano et al. (2010)
found that childhood ADHD predicted both depression and
suicide attempts in adolescence.
A question arises from the literature as to the nature of
the linkages between ADHD and self-harm/suicidal
behaviour. It could be argued that there is a direct linkage
between ADHD and self-harm. For example, it could be
argued that the linkages between ADHD and self-harm
may be due to intervening variables such as mental health
disorders, or coping style. Research suggests that ADHD
may associated with a number of comorbid disorders, such
as depression, anxiety, and substance abuse (Hesslinger
et al. 2003; Biederman et al. 1996; Fischer et al. 2002),
although the nature of the linkages between ADHD and
comorbid mental health disorders is not entirely clear
(Connor et al. 2003a; Jensen et al. 2001). However, the
evidence more clearly suggests that mental health disorders
are in turn significant risk factors for suicidal behaviours
(Beautrais et al. 1996). Furthermore, evidence suggests that
ADHD may be associated with maladaptive forms of
interpersonal coping (Hampel et al. 2008), in particular
emotion-focussed coping, which may in turn be related to
increased risks of suicidal behaviour (Edwards and Holden
2001; Horwitz et al. 2011).
Aims of the study
Against this background, the aims of the present pilot study
were to explore linkages between ADHD symptomatology
and self-harm and suicidal behaviour, using data from a
preliminary case–control study of ADHD and outcomes in
a sample of adults. It was hypothesized that higher levels of
ADHD symptomatology would predict increased risks of
self-harm and suicidal behaviour. We hypothesized that
there would be both direct and indirect relationships from
ADHD to self-harm and suicidal ideation/behaviour, via
psychosocial factors such as mental health disorders
including anxiety, depression, and substance abuse; and
emotion-focussed coping style.
Methods
Sample
Sixty-six participants were recruited as part of a wide-
ranging study of adult ADHD via a participant pool from
existing studies at the University of Canterbury (New
Zealand); advertisements on campus, in local media; and
referral from community mental health treatment services.
Exclusion criteria for the study were (a) IQ under 70; (b) a
history of psychotic illness; (c) a history of significant
traumatic brain injury; (d) diagnoses of pervasive devel-
opmental disorder; or (e) being unable to provide corrob-
orating information for the ADHD assessment (e.g. parent
or partner completed measures and/or recent clinical
diagnosis of ADHD by a trained mental health
professional).
Of the participants, 26 % were university students.
Participants’ mean age was 31.9 years (SD =1.6). The
research methods used in this study were approved by both
the host university and regional Health and Disability
ethics boards. Written consent was obtained prior to
interview and reviewed to ensure that participants were
fully informed.
Although the study consisted of two groups [those
meeting criteria for adult ADHD (n=35; 23 males, 12
females) and those not meeting criteria (n=31; 20 males,
11 females)], for the purposes of the present investigation,
the two groups were combined to form a single sample.
Tests of group membership (ADHD/no ADHD) 9covar-
iate interactions were performed to ensure that the strength
of association between covariates and outcomes did not
differ across the two groups. In no case was a statistically
significant interaction observed (all pvalues [.05).
Measures
ADHD symptomatology
All participants were administered structured interviews in
order to assess ADHD symptomatology and inclusion/
exclusion criteria. In addition, the Conners’ Adult ADHD
Rating Scales (CAARS: Conners et al. 1999) was used to
screen for the presence of attentional difficulties as well as
to assess severity (a=.86–.92). In this study, a continuous
predictor variable of ADHD symptomatology was derived
from the severity scores of the CAARS Index G T-scores
(ADHD DSM-IV Symptom Total). The average of the self-
report and observer report scores was obtained to form an
overall continuous measure of ADHD symptoms. As
expected, this continuous variable is consistent with our
dichotomous group variable as demonstrated by the high-
point biserial correlation between group membership and
304 M. R. Taylor et al.
123
ADHD severity scores on the CAARS (rpbð64Þ¼:85;
p\.001). Participants were then grouped into a four-group
independent variable (ADHD Severity) based on quartile
ranking of the above score of ADHD symptomatology
severity (\25, 26–50, 51–75, 76 %[). These quartiles were
used as the independent variable in the analyses reported
below for the purposes of reporting clarity.
Self-destructive behaviours
Three variables were chosen to measure lifetime self-
destructive behaviours, including deliberate self-harm
behaviours (without suicidal intent), recurrent suicidal
ideation, and past suicide attempts. Self-harm behaviours
were assessed with the Deliberate Self-Harm Inventory
(DSHI; Gratz 2001), which was administered to all partic-
ipants. The DSHI is a self-report questionnaire that assesses
an individual’s history of self-harm behaviours as well as
the frequency, severity duration, and type of those behav-
iours. Deliberate self-harm was defined as some self-inju-
rious behaviour performed without the intention of ending
one’s own life. In this study, the Self-harm dependent var-
iable was a dichotomous variable derived from a positive
response on any questions of the DSHI (Gratz 2001).
Recurrent suicidal intention was assessed by Q.9 of the
mood episodes section of the SCID-I (First et al. 2002):
‘‘Were things so bad that you were thinking a lot about
death or that you would be better off dead? What about
thinking of hurting yourself?’’ This SCID-I screening
question related to both current and past mood episodes
and was asked of all participants (regardless of presence of
mood disorder). Responses to this question formed the
Suicidal Ideation dependent variable in this study, a
dichotomous variable whereby negative and positive
responses were coded 0 and 1, respectively.
If participants indicated any suicidal ideation, they were
further assessed concerning suicide attempts using the
following question: ‘‘Did you do anything to hurt your-
self?’’ (First et al. 2002). Responses to the above question
were coded to record the number of self-reported suicide
attempts (ranging from 0 to 2). These responses formed the
continuous dependent variable: suicide attempts. A suicide
attempt was defined as some act performed with the
intention to end one’s life with some belief in the lethality
of the method used.
Potential confounding factors
Demographic information
Demographic variables including age, gender, ethnicity,
and occupation were assessed. Of the participants, 72.7 %
were European New Zealanders, 12.1 % Chinese, and
9.1 % Indian (other ethnicities were represented by only
three individuals, 4.5 %).
The occupational responses from all participants were
converted into SES scores using the New Zealand Socio-
Economic Index (NZSEI; Davis et al. 1997); a measure
which involves the assignation of a score based on one of
97 coded occupational groups, which range from 10 to 90
(10 being the lowest and 90 representing the highest
occupational group). The participants were classified as
being low or high SES on the basis of their occupation
being lower or higher than the midpoint of the scale. On the
basis of this classification, 60.6 % of the participants were
classified as being low SES, due primarily to the number of
students in the sample.
Intellectual functioning
The Wechsler Abbreviated Scales of Intelligence (WASI;
Wechsler 1999) were administered to gauge general levels
of intellectual functioning (m=120.2; SD =13.4).
History of child abuse
The Childhood Trauma Questionnaire (CTQ: Bernstein and
Fink 1998) was administered to all participants, covering
the period of childhood and adolescence. The CTQ asses-
ses the occurrence of childhood trauma, differentiating
between emotional, physical, and sexual abuse whilst
excluding experiences of non-abuse-related traumatic
events such as death of a parent. For the purposes of the
present investigation, the measures of emotional, physical
and sexual abuse were combined to form a composite
measure of abuse exposure ranging from 0 to 5 (m=1.71;
SD =1.51).
History of conduct problems
A history of conduct problems from childhood to early
adulthood was assessed using questions from the Con-
ners’Adult ADHD Diagnostic Interview for DSM-IV
(CAADID: Epstein et al. 2001) to assess delinquent and
conduct-disordered behaviour. On the basis of this ques-
tioning, 25.8 % of the sample reported a history of conduct
problems.
Potential mediating factors
Comorbid psychopathology
All participants were administered the Structured Clinical
Interview for DSM-IV-TR Axis 1 Disorders, Research
Version (SCID-I) in order to assess for the presence of
Attention problems and risk of harm 305
123
lifetime significant psychiatric symptomatology (Ramirez
Basco et al. 2000). In addition to the SCID-I, the partici-
pants completed a further three psychometric measures in
order to provide normative (and severity) measures of
mood, anxiety, and substance abuse symptoms. These
included: 1) The Depression, Anxiety and Stress Scale
(DASS; Lovibond and Lovibond 1995), a 42-item ques-
tionnaire which assesses an individual’s level of symptoms
relating to depression, anxiety and stresses; 2) the Michi-
gan Alcoholism Screening Test (MAST; Selzer 1971); and
3) the Drug Abuse Screening Test (DAST; Skinner 1982).
The MAST is a self-administered questionnaire that mea-
sures current and lifetime alcohol-related problems
(a=0.86). The DAST is a self-report measure that
assesses an individual’s current and lifetime use of classi-
fied and prescription drugs. Participant responses on the
above measures were pooled to form a mental health dis-
order mediator variable, whereby each participant was
allocated a cumulative score for each comorbid psychiatric
criterion they met (i.e. 1 point for each of the following:
mood disorder, anxiety disorder, alcohol abuse disorder,
drug abuse disorder) resulting in a range of scores from 0 to
4. Participants met each criterion if their SCID-I scores
were commensurate with their scores in relation to the cut-
off scores of each of the measures mentioned above
(Lovibond and Lovibond 1995; Selzer 1971; Skinner
1982). Of the sample, 56 % met criteria for a mood dis-
order, 37.9 % met criteria for an anxiety disorder, 33.3 %
met criteria for an alcohol use disorder, and 22.7 met cri-
teria for an illicit drug abuse disorder.
Emotion-focussed coping style
The Coping Inventory for Stressful Situations (CISS; En-
dler and Parker 1990) was administered to all participants.
The CISS is a self-report measure which measures an
individual’s level of coping associated with stressful events
as well as coping style. In this study, participants’ scores on
the emotion-focussed coping style index score were used as
a mediator variable.
Statistical analyses
The data were analysed over several steps. In the first step
of the analyses, the bivariate associations between the
quartile measure of ADHD and the three outcomes (self-
harm; suicidal ideation; suicide attempts) were modelled
using logistic regression (for the measures of self-harm and
suicidal ideation) and Poisson regression (for suicide
attempts). In order to examine the extent to which these
associations could be explained by confounding factors, the
demographic factors, measures of IQ and conduct disorder
symptomatology, and exposure to child abuse were entered
into each of the models individually.
In the second step of the analyses, the associations
between the potential mediating factors (the count measure
of mental health disorders and the emotion-focussed cop-
ing style measure) and the three outcome measures were
also modelled using logistic and negative binomial
regression modelling.
In the third step of the analyses, we examined the pos-
sible mediating role of mental health disorders and emo-
tion-focussed coping style in the linkages between ADHD
and the outcomes. In order to examine this, it was neces-
sary to model both the direct and indirect pathways from
ADHD to self-harm/suicidality using a meditation analytic
approach. An issue arising, however, from many of the
more common meditational approaches (such as the Sobel
test: Sobel 1982) is that these approaches assume a normal
distribution amongst both predictors and outcomes (see
Hayes for a discussion of these issues: 2009).
One approach that does not assume normality amongst
variables is the bootstrapping of indirect effects via latent
variable structural equation modelling (SEM; Muthen and
Muthen 2007). In this approach, bootstrapping is used to
estimate bias-corrected confidence intervals for each direct
and indirect effect in the model, thereby reducing the risk
of Type II error and increasing the power of the model to
detect effects. Furthermore, these approaches allow the
specification of more complex models with multiple
mediating pathways (see below). The bootstrapping
approach is particularly appropriate for the present analy-
ses as they employ a mixture of variable scales, including
dichotomous outcomes and ordinal and continuous pre-
dictors, and in two cases, the models employ two inter-
vening variables simultaneously.
There were three steps to testing this mediation model:
(1) the independent variable (ADHD Severity) was
regressed onto each of the three dependent variables (self-
harm, suicidal ideation, and suicidal attempts), (2) the
independent variable was also regressed onto the mediator
variables (mental health disorders and emotion-focussed
coping style), and (3) the mediator variables were, in turn,
regressed onto the dependent variable (self-destructive
behaviours). If the regressions between the variables were
found to be significant in all steps 1–3, then a fourth step
involved testing mediation. A result is found to be signif-
icant if the independent variable (ADHD) becomes non-
significant (full mediation) or reduced in significance
(partial mediation) when the mediator is added to the
regression between the independent and dependent
variables.
Bootstrapping latent variable models were fitted using
MPlus V.5 (Muthen and Muthen 2007). In these models,
effects were estimated for the direct pathway between
306 M. R. Taylor et al.
123
ADHD and each outcome, as well as the indirect pathways
via mental health disorders and emotion-focussed coping,
using weighted least squares and Delta parameterization.
The models also provided tests of statistical significance
for each direct and indirect pathway in the model.
In addition, to examine the extent to which the data were
robust to alternative representations of the measure of
ADHD, supplementary analyses were conducted in which
the four-group measure of ADHD was replaced with the
dichotomous measure representing whether the individual
met criteria for a diagnosis of ADHD, and the analyses
above repeated.
Results
Sample characteristics
Table 1shows the sample classified into four ADHD
severity score quartiles. For each quartile, the mean ADHD
severity score, the number of participants meeting criteria
for ADHD, and the number of participants in the quartile
are provided. The data clearly show that the ADHD quar-
tile scores represent increasing levels of ADHD.
Associations between ADHD and outcome measures
(self-harm; suicidal ideation; suicide attempts)
Table 2also shows the sample divided into quartiles on
ADHD severity score. For each quartile, the table displays
the percentage of participants classified as demonstrating
self-harm or suicidal ideation, and the mean number of
suicide attempts. The table also displays the parameter
estimates and tests of significance for the logistic and
Poisson regression models for the associations between
ADHD quartile and outcomes. The table shows that
increasing levels of ADHD symptoms were significantly
(p \.05) associated with increased risks of: self-harm;
suicidal ideation; and suicide attempts.
Testing for potential confounding (gender; ethnicity;
IQ; socio-economic status; history of child abuse;
conduct disorder) in the association between ADHD
and outcome measures
In order to examine the possibility that the associations
between ADHD and outcomes could be explained by
confounding factors, the models described above were
extended to include the following variables gender, eth-
nicity, IQ, socio-economic status, history of child abuse,
and conduct disorder symptomatology. In no case was a
potentially confounding factor statistically significant (all
pvalues [.05).
Associations between potential mediating factors
(mental health disorders; emotion-focussed coping
style) and outcome measures
As noted in Methods, two variables were chosen as
potential mediating factors in the analyses (mental health
disorders; emotion-focussed coping style). The Pearson
product moment correlations for each of these with the
quartile measure of ADHD were .50 (p\.0001) and .41
(p\.001), respectively. Table 3shows the associations
between the two mediating factors (the count measure of
mental health disorders; and emotion-focussed coping) and
the three outcomes. The tables show that:
1. Increasing rates of mental health disorders were
significantly (p\.05) associated with increasing risk
of self-harm; suicidal ideation; and suicide attempts.
2. Higher scores on the measure of emotion-focussed
coping were significantly (p\.05) associated with
increasing risk of self-harm; suicidal ideation; and
suicide attempts.
Table 1 Characteristics of sample
Clinical characteristics ADHD score quartiles
1–25 % 26–50 % 51–75 % 76–100 %
Mean (SD) ADHD
severity score
a
40.68
(3.71)
52.41
(3.33)
66.26
(5.33)
81.50
(3.61)
% met criteria for
ADHD
0.0 18.8 94.1 100.0
n17 16 17 16
a
‘‘ADHD severity score’’ indicates an average of ADHD symptom
scores derived from self-report, and observer report. Scores ranged
from 34.5 to 88.5
Table 2 Associations between ADHD severity score and outcomes
Outcome ADHD score quartiles
1–25 % 26–50 % 51–75 % 76–100 %
Self-harm (%) 23.5 37.5 41.2 62.5
nof cases 4 6 7 10
B=.52, SE =.24, p\.05
Suicidal ideation (%) 23.5 12.5 58.8 62.5
nof cases 4 2 10 10
B=.74, SE =.26, p\.01
Mean suicide attempts
(SD)
.06
(.24)
.06 (.25) .12 (.33) .50 (.82)
nof cases 1 1 2 5
B=.88, SE =.36, p\.05
Attention problems and risk of harm 307
123
Tests of mediation in the association between ADHD
and outcome measures
As noted above, the extent to which mental health disor-
ders and emotion-focussed coping mediated the linkages
between ADHD and each of the three outcomes was
examined using bootstrapping of indirect effects via latent
variable SEM (Muthen and Muthen 2007). In this proce-
dure, the data were modelled with a single direct pathway
between ADHD and each outcome, self-harm, and two
indirect pathways between ADHD and outcomes, the first
via the mental health disorder count measure and the sec-
ond via the emotion-focussed coping measure. For the
measure of self-harm, the results of this modelling showed:
1. There were statistically significant pathways between
ADHD and both mental health disorders (b=1.13,
SE =.32, p\.0001) and emotion-focussed coping
(b=12.33, SE =2.73, p\.0001), and statistically
significant pathways to self-harm from both mental
health disorders (b=.43, SE =.16, p\.01) and
emotion-focussed coping (b=.05, SE =.01,
p\.0001). However, there was no evidence of a
statistically significant direct pathway from ADHD to
self-harm (b=-.5, SE =.37, p[.10).
2. Overall tests of the mediating pathways via mental
health disorders and emotion-focussed coping were
found to be statistically significant (mental health
disorders: b=.48, SE =.20, p\.05; emotion-
focussed coping: b=.61, SE =.19, p\.01). A test
of the total indirect effect via both mediating factors
was also found to be statistically significant (b=1.09,
SE =.29, p\.0001).
The results of these analyses suggest that the linkages
between ADHD and self-harm were mediated by mental
health disorders and emotion-focussed coping. Those
individuals with higher ADHD scores were at greater risk
of self-harm, and this risk could be largely explained by a
greater burden of mental health disorder comorbidity, and a
greater tendency to use emotion-focussed coping methods
amongst those with higher ADHD scores (Fig. 1).
Similar to the aforementioned model for self-harm, the
mediation model for suicidal ideation used a single direct
pathway between ADHD and suicidal ideation, and two
indirect pathways between ADHD and suicidal ideation,
the first via the mental health disorder count measure, and
Table 3 Associations between mediating variables (mental health disorders; emotion-focussed coping) and outcomes
Outcome Mental health disorders
0(n=20) 1 (n=14) 2 (n=17) 3 (n=10) 4 (n=5)
Self-harm (%) 20.0 28.6 52.9 60.0 80.0
B=.68, SE =.23, p\.01
Suicidal ideation (%) 5.0 21.4 64.7 70.0 80.0
B=1.18, SE =.30, p\.0001
Mean suicide attempts (SD) 0.0 (0.0) .14 (.36) .24 (.56) .30 (.67) .60 (.89)
B=.69, SE =.28, p\.05
Outcome Emotion-focussed coping score quintile
1–20 % (n=11) 21–40 % (n=14) 41–60 % (n=12) 61–80 % (n=13) 81–100 % (n=13)
Self-harm (%) 0 21.4 66.7 53.8 61.5
B=.72, SE =.22, p\.01
Suicidal ideation (%) 18.1 14.2 41.7 46.2 76.9
B=.73, SE =.23, p\.01
Mean suicide attempts (SD) .09 (.30) 0.0 (0) .25 (.62) .15 (.38) .46 (.78)
B=.54, SE =.27, p\.05
ADHD
Emotion
Coping
Self-
Harm
Mental
Health
Fig. 1 Mediation model showing the single direct pathway between
ADHD and self-harm (a), and two indirect pathways between ADHD
and self-harm, the first via the mental health disorders variable (b), the
second via the emotion coping variable (c). The total indirect effect
via both mediating factors is also shown (d). Notes:*p\.05,
**p\.01
308 M. R. Taylor et al.
123
the second via the emotion-focussed coping measure (see
Fig. 2). The results of this modelling showed:
1. The same statistically significant pathways between
ADHD and both mental health disorders and emotion-
focussed coping applied to this model as in the
previous model. Statistically significant pathways to
suicidal ideation from both mental health disorders
(B=.61, SE =.12, p\.0001) and emotion-focussed
coping (B=.03, SE =.02, p\.05) were also evi-
dent. Similar to the model above, there was no
evidence of a statistically significant direct pathway
from ADHD to suicidal ideation (B=.17, SE =.43,
p[.10).
2. Overall tests of the mediating pathways via mental
health disorders were again found to be statistically
significant (B=.68, SE =.26, p\.01) and margin-
ally significant via emotion-focussed coping: (B=.39,
SE =.20, p=.056). A test of the total indirect effect
on suicidal ideation via both mediating factors was
also found to be statistically significant (B=1.07,
SE =.33, p\.001).
The results of these analyses suggest that the linkages
between ADHD and suicidal ideation were also mediated
by mental health disorders and emotion-focussed coping.
Those individuals with higher ADHD scores were at
greater risk of suicidal ideation, and this risk could be
largely explained by a greater burden of mental health
disorder comorbidity, and a greater tendency to use emo-
tion-focussed coping methods amongst those with higher
ADHD scores.
In the final procedure, the same methods of bootstrap-
ping of indirect effects via latent variable SEM (Muthen
and Muthen 2007) were utilized to examine the extent to
which mental health disorders mediated the relationship
between ADHD and number of past suicide attempts. Ini-
tial model-fitting iterations found that emotion-focused
coping was not a statistically significant (p[.05)
mediator. Therefore, emotion-focussed coping was omitted
from any further modelling. In the final fitted model, the
data were modelled with a single direct pathway between
ADHD and suicide attempts, and a single indirect pathway
between ADHD and suicide attempts via the mental health
disorder count measure (see Fig. 3). The results of this
modelling showed:
1. The same statistically significant pathways between
ADHD and mental health disorders were repeated in
the present model as in the two models described
previously (above). A statistically significant pathway
to suicide attempts from mental health disorders
(B=.43, SE =.14, p\.003) was also evident.
Again, similar to the models above, there was no
evidence of a statistically significant direct pathway
from ADHD to suicide attempts (B=.27, SE =.40,
p[.10).
2. An overall test of the total indirect effect on suicide
attempts via mental health disorders was also found to
be statistically significant (B=.48, SE =.23,
p\.05).
The results of these analyses suggest that the associa-
tions between ADHD and suicide attempts were also
mediated by mental health disorders (but not emotion-
focussed coping). Those individuals with higher ADHD
scores were at greater risk of attempting suicide, and this
risk was largely explained by a greater burden of mental
health disorder comorbidity amongst those with higher
ADHD scores.
Supplementary analyses
As noted in ‘‘Methods’’ section, the analyses above were
repeated using the dichotomous ADHD classification in
place of the four-group measure of ADHD. The results of
these analyses were largely congruent with those reported
above, suggesting that the findings were robust to alterna-
tive specifications of the measure of ADHD.
ADHD
Emotion
Coping
Suicidal
Ideation
Mental
Health
Fig. 2 Mediation model showing the single direct pathway between
ADHD and suicidal ideation (a), and two indirect pathways between
ADHD and suicidal ideation, the first via the mental health disorders
variable (b), the second via the emotion-focussed coping variable (c).
The total indirect effect via both mediating factors is also shown (d).
Notes:*p\.05, **p\.01
ADHD Suicide
Attempt
Mental
Health
Fig. 3 Mediation model showing the single direct pathway between
ADHD and suicide attempts (a), and a single indirect pathway
between ADHD and suicide attempts, via the mental health disorders
variable (b). Notes:*p\.05, **p\.01
Attention problems and risk of harm 309
123
Discussion
The present pilot study explored the associations between
ADHD in adults and a number of adverse psychosocial
outcomes, including self-harm; suicidal ideation; and sui-
cide attempt, using data from a case–control study of
ADHD in adulthood. In particular, the study explored
whether the linkages between ADHD and self-harm/sui-
cidal behaviour could be explained by the mediating effects
of both comorbid mental health disorders and emotion-
focussed coping style. This issue is of particular interest
because a number of studies have found an association
between ADHD and the risk factors for suicide and self-
harm behaviours (Hesslinger et al. 2003). In addition,
ADHD symptomatology that persists into adulthood is
associated with a number of other high-risk behaviours
(Barkley and Gordon 2002). The analyses of the present
data revealed the following findings.
First, the findings suggested that ADHD symptomatol-
ogy severity was found to be significantly and indirectly
associated with lifetime histories of self-harm (self-injuri-
ous behaviour performed without the intention of ending
one’s own life); recurrent suicidal ideation (thoughts of
ending one’s life); and previous occurrences of suicide
attempts (act performed with the intention to end one’s life
with some belief in the lethality of the method used). This
is in general agreement with a range of research linking
ADHD and self-harm/suicidal behaviour in both children
and adults (Barkley 2006; Goldston et al. 2009; Haavisto
et al. 2005; Semiz et al. 2008; Westmoreland et al. 2010).
Second, the results of the present study suggested that
the linkages between ADHD symptomatology and self-
destructive behaviours could be explained by intervening
factors, including mental health disorders and emotion-
focussed coping style. Specifically, it was found that life-
time psychiatric comorbidity (mood, anxiety, and sub-
stance abuse disorders) significantly mediated the
relationship between ADHD and all three self-destructive
behaviours. This finding was consistent with previous
studies that have found psychopathology to be the primary
risk factor for suicidal behaviour (Cavanagh et al. 2003)
and self-harm (Nock et al. 2006).
In addition to psychiatric comorbidity, emotion-focussed
coping style was found to be a significant mediator between
ADHD and self-harm/suicidal ideation (but not suicide
attempts). Emotion-focussed coping style reflects trait-like
tendencies in response to stressful or upsetting events, as
measured by Endler and Parker’s Coping Inventory for
Stressful Situations (Endler and Parker 1990). Individuals
who score highly on the Emotion-Oriented Coping Style
Index are more likely to resort to emotional responses (such
as blaming); preoccupation with their feelings, and
increased self-focus; all responses which may inadvertently
increase distress (Endler and Parker 1990). Consistent with
the results from this study, Endler and Parker (1994) and
Billings and Moos (1984) found an association between
emotion-focussed coping styles with high levels of psy-
chopathology, above any association with other styles of
coping. The lack of a significant association between
emotion-focussed coping style and suicide attempts in this
study was inconsistent with a study by Edwards and Holden
(2001) who found that emotion-focussed coping style pre-
dicted suicide attempts in both women and men. A possible
reason for this discrepancy in findings is that the present
study may have lacked adequate power to examine this
pathway due to a relatively small sample size.
The results of the present study highlight the importance
of efforts to identify maladaptive coping styles and pro-
cesses, and diagnosis and treatment of comorbid psycho-
pathology amongst individuals with ADHD (Connor et al.
2003b). Because ADHD is frequently diagnosed at early
ages (McGoey et al. 2002), it is critical that follow-up care
be provided to reduce later psychopathology that may
increase the risk of self-harm and suicidal behaviour in the
longer term. In addition, treatments that focus on fostering
more adaptive coping styles and mechanisms may help to
reduce the risk of self-harm and suicidal behaviour
amongst the population of individuals diagnosed with
ADHD (Davidson 2007).
The limitations of this study are largely related to the
small sample size achieved. As a result, whilst power
levels were adequate for two-step mediation models, more
in-depth exploration was not possible without increasing
type II error levels significantly. Further research is nee-
ded to replicate the present findings using an adequate
sample size. A further limitation of this study is the ret-
rospective design of the data that were utilized to assess
both ADHD symptomatology and the dependent variables
of self-destructive behaviours. An ideal research design
would involve the longitudinal follow-up of individuals
with ADHD. This follow-up would ideally extend past age
30, as many individuals with ADHD who attempt suicide
do so at an older age (James et al. 2004). An additional
limitation of this study is the reliance on self-reported
data. Whilst ADHD diagnosis was conditional on cor-
roborated information, self-destructive behaviours and
psychosocial mediators were based on self-report in this
study, which leads to caution regarding the reliability of
these data. Finally, individuals in this study were found to
achieve higher than average intellectual functioning
scores, as measured by the WASI, which may be due to
selection bias and an overrepresentation (relative to the
population as a whole) of university students and staff
amongst the sample.
Notwithstanding the limitations of the study, the present
analyses found significant associations between ADHD and
310 M. R. Taylor et al.
123
self-harm, suicidal ideation and suicide attempts, which
were mediated by comorbid psychopathology and emotion-
focussed coping style. The results suggest that increased
risks of self-injurious outcomes amongst those exhibiting
ADHD symptomatology may be particularly acute for
individuals with comorbid mental health disorders, and
those who exhibit a reliance on emotion-focussed coping.
These findings may not only help to elucidate the risks
associated with the persistence of ADHD into adulthood,
but they also may provide a useful model for potential
intervention into suicide and self-harm prevention amongst
adults with impulsive or disinhibited traits.
Acknowledgments This research was conducted as partial fulfil-
ment of the requirements for the degree of Doctor of Philosophy at the
University of Canterbury by the first author, under the supervision of
the third author. The first author would like to acknowledge the
ongoing research support from the University of Canterbury. In
addition, the authors would like to thank Jason Brown, Sarah-Eve
Harrow, and Rachel Harrison for their help with data collection. Most
importantly, we would like to thank the participants who shared their
experiences with us.
Conflict of interest The first author is a Registered Clinical Psy-
chologist specializing in ADHD treatment. The authors report no
other potential conflicts of interest. All statistical analyses were
conducted by the second author.
References
Arias AJ, Gelernter J, Chan G, Weiss RD, Brady KT, Farrer L et al
(2008) Correlates of co-occurring ADHD in drug-dependent
subjects: prevalence and features of substance dependence and
psychiatric disorders. Addict Behav 33(9):1199–1207. doi:10.
1016/j.addbeh.2008.05.003
Barkley RA (2006) Attention-deficit hyperactivity disorder: a hand-
book for diagnosis and treatment, 3rd edn. Guilford Press, New
York
Barkley RA, Gordon M (2002) Research on comorbidity, adaptive
functioning, and cognitive impairments in adults with ADHD:
implications for a clinical practice. In: Goldstein S, Ellison AT
(eds) Clinicians’ guide to adult ADHD: assessment and inter-
vention. Academic Press, Amsterdam, pp 43–69
Beautrais AL, Joyce PR, Mulder RT, Fergusson DM, Deavoll BJ,
Nightingale SK (1996) Prevalence and comorbidity of mental
disorders in persons making serious suicide attempts: a case
controlled study. Am J Psychiatry 153(8):1009–1014
Bernstein DP, Fink L (1998) Childhood Trauma Questionnaire: a
retrospective self-report manual. The Psychological Corporation,
San Antonio
Biederman J, Faraone S, Milberger S, Guite J, Mick E, Chen L et al
(1996) A prospective 4-year follow-up study of attention-deficit
hyperactivity and related disorders. Arch Gen Psychiatry
53(5):437–446
Biederman J, Monuteaux MC, Mick E, Spencer T, Wilens TE, Silva
JM et al (2006) Young adult outcome of attention deficit
hyperactivity disorder: a controlled 10-year follow-up study.
Psychol Med 36(2):167–179
Billings AG, Moos RH (1984) Coping, stress, and social resources
among adults with unipolar depression. J Personal Soc Psychol
46(4):877–891
Cavanagh J, Carson A, Sharpe M, Lawrie S (2003) Psychological
autopsy studies of suicide: a systematic review. Psychol Med
33(3):395–405
Chronis-Tuscano A, Molina BS, Pelham WE, Applegate B, Dahlke A,
Overmyer M et al (2010) Very early predictors of adolescent
depression and suicide attempts in children with attention-
deficit/hyperactivity disorder. Arch Gen Psychiatry
67(10):1044–1051. doi:10.1001/archgenpsychiatry.2010.127
Conners CK, Erhardt D, Sparrow E (1999) Conners’ Adult ADHD
Rating Scales (CAARS) technical manual. Multi-Health Systems
Inc, North Tonawanda
Connor DF, Edwards G, Fletcher KE, Baird J, Barkley RA, Steingard
RJ (2003) Correlates of comorbid psychopathology in children
with ADHD. J Am Acad Child Adolesc Psychiatry
42(2):193–200. doi:10.1097/00004583-200302000-00013
Davidson K (2007) Cognitive therapy for personality disorders: a
guide for clinicians. Routledge, New York
Davis P, McLeod K, Ransom M, Ongley P (1997) The New Zealand
Socioeconomic Index of Occupational Status (NZSEI): Research
Report #2. Statistics New Zealand, Wellington
Edwards MJ, Holden RR (2001) Coping, meaning in life, and suicidal
manifestations: examining gender differences. J Clin Psychol
57(12):1517–1534
Endler NS, Parker JDA (1990) Coping inventory for stressful
situations (CISS): manual. Multi-Health Systems, Toronto
Endler NS, Parker JDA (1994) Assessment of multidimensional
coping: task, emotion, and avoidance strategies. Psychol Assess
6(1):50–60. doi:10.1037/1040-3590.6.1.50
Epstein JN, Johnson DE, Conners CK (2001) Conners’ adult ADHD
diagnostic interview for DSM-IV. Multi-Health Systems, North
Tonawanda
First M, Spitzer R, Gibbon M, Williams J (2002) Structured clinical
interview for DSM-IV Axis I disorders, research version, non-
patient edition (SCID-I/NP). New York State Psychiatric Insti-
tute, Biometrics Research, New York
Fischer M, Barkley RA, Smallish L, Fletcher K (2002) Young adult
follow-up of hyperactive children: self-reported psychiatric
disorders, comorbidity, and the role of childhood conduct
problems and teen CD. J Abnorm Child Psychol 30(5):463–475
Galera C, Bouvard MP, Encrenaz G, Messiah A, Fombonne E (2008)
Hyperactivity-inattention symptoms in childhood and suicidal
behaviors in adolescence: the Youth Gazel cohort. Acta Psychiatr
Scand 118(6):480–489. doi:10.1111/j.1600-0447.2008.01262.x
Goldston DB, Daniel SS, Erkanli A, Reboussin BA, Mayfield A,
Frazier PH et al (2009) Psychiatric diagnoses as contemporane-
ous risk factors for suicide attempts among adolescents and
young adults: developmental changes. J Consult Clin Psychol
77(2):281–290. doi:10.1037/a0014732
Gratz KL (2001) Measurement of deliberate self-harm: preliminary
data on the deliberate self-harm inventory. J Psychopathol Behav
Assess 23(4):253–263. doi:10.1023/a:1012779403943
Haavisto A, Sourander A, Multimaki P, Parkkola K, Santalahti P,
Helenius H et al (2005) Factors associated with ideation and acts
of deliberate self-harm among 18-year-old boys. A prospective
10-year follow-up study. Soc Psychiatry Psychiatr Epidemiol
40(11):912–921. doi:10.1007/s00127-005-0966-2
Hampel P, Manhal S, Roos T, Desman C (2008) Interpersonal coping
among boys with ADHD. J Atten Disord 11(4):427–436. doi:10.
1177/1087054707299337
Hayes AF (2009) Beyond Baron and Kenny: statistical mediation
analysis in the new millennium. Commun Monogr 76(4):408–
420
Hesslinger B, Tebartz van Elst L, Mochan F, Ebert D (2003) A
psychopathological study into the relationship between attention
deficit hyperactivity disorder in adult patients and recurrent brief
depression. Acta Psychiatr Scand 107(5):385–389
Attention problems and risk of harm 311
123
Horwitz AG, Hill RM, King CA (2011) Specific coping behaviors in
relation to adolescent depression and suicidal ideation. J Adolesc
34(5):1077–1085. doi:10.1016/j.adolescence.2010.10.004
Impey M, Heun R (2012) Completed suicide, ideation and attempt in
attention deficit hyperactivity disorder. Acta Psychiatr Scand
125(2):93–102. doi:10.1111/j.1600-0447.2011.01798.x
Izutsu T, Shimotsu S, Matsumoto T, Okada T, Kikuchi A, Kojimoto
M et al (2006) Deliberate self-harm and childhood hyperactivity
in junior high school students. Eur Child Adolesc Psychiatry
15(3):172–176. doi:10.1007/s00787-005-0520-5
James A, Lai FH, Dahl C (2004) Attention deficit hyperactivity
disorder and suicide: a review of possible associations. Acta
Psychiatr Scand 110(6):408–415. doi:10.1111/j.1600-0447.2004.
00384.x
Jensen PS, Hinshaw SP, Kraemer HC, Lenora N, Newcorn JH,
Abikoff HB et al (2001) ADHD comorbidity findings from the
MTA study: comparing comorbid subgroups. J Am Acad Child
Adolesc Psychiatry 40(2):147–158
Klein RG, Mannuzza S, Olazagasti MA, Roizen E, Hutchison JA,
Lashua EC et al (2012) Clinical and functional outcome of
childhood attention-deficit/hyperactivity disorder 33 years later.
Arch Gen Psychiatry 69(12):1295–1303. doi:10.1001/archgenp
sychiatry.2012.271
Lam LT (2002) Attention deficit disorder and hospitalization due to
injury among older adolescents in New South Wales, Australia.
J Atten Disord 6(2):77–82
Lovibond SH, Lovibond PF (1995) Manual for the Depression
Anxiety Stress Scales, 2nd edn. Psychology Foundation, Sydney
McGoey KE, Eckert TL, Dupaul GJ (2002) Early intervention for
preschool-age children with ADHD: a literature review. J Emot
Behav Disord 10(1):14–28. doi:10.1177/106342660201000103
Murphy K, Barkley RA (1996) Attention deficit hyperactivity
disorder adults: comorbidities and adaptive impairments. Compr
Psychiatry 37(6):393–401
Muthen LK, Muthen BO (2007) Mplus users’ guide, 5th edn. Muthen
and Muthen, Los Angeles
Nock MK, Joiner TE Jr, Gordon KH, Lloyd-Richardson E, Prinstein
MJ (2006) Non-suicidal self-injury among adolescents: diagnos-
tic correlates and relation to suicide attempts. Psychiatry Res
144(1):65–72. doi:10.1016/j.psychres.2006.05.010
Ramirez Basco M, Bostic JQ, Davies D, Rush AJ, Witte B,
Hendrickse W et al (2000) Methods to improve diagnostic
accuracy in a community mental health setting. Am J Psychiatry
157(10):1599–1605
Ramos Olazagasti MA, Klein RG, Mannuzza S, Belsky ER,
Hutchison JA, Lashua-Shriftman EC et al (2013) Does childhood
attention-deficit/hyperactivity disorder predict risk-taking and
medical illnesses in adulthood? J Am Acad Child Adolesc
Psychiatry 52(2):153–162.e154
Selzer ML (1971) The Michigan alcoholism screening test: the quest
for a new diagnostic instrument. Am J Psychiatry 127(12):
1653–1658
Semiz UB, Basoglu C, Oner O, Munir KM, Ates A, Algul A et al
(2008) Effects of diagnostic comorbidity and dimensional
symptoms of attention-deficit-hyperactivity disorder in men with
antisocial personality disorder. Aust N Z J Psychiatry
42(5):405–413. doi:10.1080/00048670801961099
Skinner HA (1982) The drug abuse screening test. Addict Behav
7(4):363–371
Sobel ME (1982) Asymptotic confidence intervals for indirect effects
in structural equation models. Sociol Methodol 13:290–312
Wechsler D (1999) Wechsler Abbreviated Scales of Intelligence
(WASI). Psychological Corporation, San Antonio
Westmoreland P, Gunter T, Loveless P, Allen J, Sieleni B, Black DW
(2010) Attention deficit hyperactivity disorder in men and
women newly committed to prison: clinical characteristics,
psychiatric comorbidity, and quality of life. Int J Offender Ther
Comp Criminol 54(3):361–377. doi:10.1177/0306624x09332313
312 M. R. Taylor et al.
123
