Prognosis and Psychosocial Outcomes of Attempted Suicide by Early Adolescence A 6-Year Follow-Up of School Students Into Early Adulthood

Abstract

Adulthood psychiatric and psychosocial outcomes of early adolescence suicidal acts were studied. A representative sample of school adolescents (T1, mean age, 13.7 years; n = 2464; 50.8% female; 88.3% participation) was followed up a year later with the same questionnaire (T2). High scorers of depression were matched with low or moderate scorers and interviewed using the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version (mean age, 14.9 years; n = 345; 94% participation). They were reassessed after 5 years (T3, mean age, 20.0 years; n = 242; 73% participation). Those who attempted suicide before the age of 14 years and repeated suicidal acts between ages 14 and 15 years had worser prognostic profiles than incident cases between ages 14 and 15 years. Male attempters had better psychiatric prognosis than female attempters. Attempters were more likely to have contacted child protection services but not mental health services. Clinicians need to be aware of long-term pervasive outcomes of adolescent suicidality.

Full text

Prognosis and Psychosocial Outcomes of Attempted Suicide
by Early Adolescence
A 6-Year Follow-Up of School Students Into Early Adulthood
Latha Nrugham, PhD,*†Are Holen, MD, PhD,‡§ and Anne Mari Sund, MD, PhD†§
Abstract: Adulthood psychiatric and psychosocial outcomes of early adoles-
cence suicidal acts were studied. A representative sample of school adolescents
(T1, mean age, 13.7 years; n= 2464; 50.8% female; 88.3% participation) was
followed up ayear later with the same questionnaire (T2). High scorers of depres-
sion were matched with low or moderate scorers and interviewed using the
Kiddie–Schedule for Affective Disorders and Schizophrenia–Present and Life-
time version (mean age, 14.9 years; n= 345; 94% participation). They were re-
assessed after 5 years (T3, mean age, 20.0 years; n= 242; 73% participation).
Those who attempted suicide before the age of 14 yearsand repeated suicidal acts
between ages 14 and 15 years had worser prognostic profiles than incident cases
between ages 14 and 15 years. Male attempters had better psychiatric prognosis
than female attempters. Attempters were more likely to have contacted child pro-
tection services but not mental health services. Clinicians need to be aware of
long-term pervasive outcomes of adolescent suicidality.
Key Words: Prospective, adolescence, young adult, suicidal acts, attempted suicide
(JNervMentDis2015;203: 294–301)
Research on adolescent suicidal behavior has focused mainly on risk
factors, less on outcomes (Bridgeet al.,2006; Brière et al., 2014) or
prognosis, despite the clear clinical relevance and policy implications
of such knowledge (Fergusson et al., 2005a). The most common cause
of death among adolescent girls aged 15 to 19 years is suicide (Patton
et al., 2009); however, more male adolescents die by suicide (Hawton
et al., 2012). The prevalence of lifetime suicide attempt for girls
(10.1%) is more than double that for boys (3.8%) in community sam-
ples of school-going adolescents (Lewinsohn et al., 1996). It is impor-
tant to note that almost half (44.1%) of all suicide attempts in this
adolescent sample were reported by those between 14 and 15 years.
In this study on adolescent depression, suicidal acts among young
women were predicted by prior suicidal acts during adolescence
(Lewinsohn et al., 2001). However, this longitudinal prospective study
did not particularly assess the outcomes of those age groups with the
highest percentage of suicidal acts: those in early adolescence aged be-
tween 14 and 15 years. When followed up at age 30 years, adolescent
suicide attempts did not predict repeated attempts or major depressive
disorder (MDD; Brière et al., 2014). However, in another longitudinal
prospective study, suicidal ideation and major depression between ages
18 and 25 years were the outcomes associated with suicide attempts
made by age 18 years among community adolescents; for the women,
two additional outcomes were found: substance abuse and later suicidal
acts (Fergusson et al., 2005b).
Other reports on outcomes of adolescent suicidal behavior are
based on inpatients or on suicidal ideation. Inpatients with attempted
suicide are more likely than nonsuicidal inpatients to report later sui-
cidal acts (Pfeffer et al., 1993). Among outpatients followed up from
childhood up to age 17 years, the female sex difference disappeared
when adjusted for depressive disorders (Kovacs et al., 1993). A 9-year
follow-up of 92 teenaged inpatients admitted to a general hospital for
attempted suicide reported that 5 had died, 2 by suicide (Grøholt and
Ekeberg, 2009). Among those alive, available, and consenting (71/92),
79% had at least one psychiatric disorder and repeated suicide attempts
were reported by 44%. In epidemiological samples, the adult outcomes
associated with adolescent suicidal ideation were subsequent suicidal
behavior, substance abuse disorders, depressive disorders, impaired
psychosocial functioning, interpersonal problems, poorer coping, and
lowered cohesion with their social network (Herba et al., 2007;
Reinherz et al., 2006; Steinhausen and Winkler Metzke, 2004).
Adolescent depression in both clinical and community samples
predicted suicide, attempted suicide, depressive disorders, impaired
functioning, and persistent interpersonal difficulties in adulthood
(Fergusson et al., 2005a; Fombonne et al., 2001a; Weissman et al.,
1999). Lower age at the onset of depression predicted higher suicidal
intent in adulthood, especially among women (Thompson, 2008). In
an epidemiological depression study on adolescents reassessed as
young adults, a drop in the rate of suicide attempts, but not in depres-
sion, was noted among the young women after a steep rise in attempted
suicide during adolescence (Lewinsohn et al., 2001). These longitudi-
nal clinical and epidemiological studies indicate that suicidal behavior
and depression are not only bidirectionally linked; they have overlap-
ping outcomes varying with sex and age. However, the studies which
reported on outcomes of suicidal acts during adolescence did not adjust
for important diagnoses such as depressive disorders or depression
scores. Neither have studies on adolescent depressive disorders or de-
pression adjusted for suicidal acts. Therefore, it is not clearly known
from the existing studies whether the outcomes were specific to suicidal
behavior or depression scores or diagnoses of depressive disorders.
The first objective of the present study was to examine the psy-
chiatric and psychosocial aspects 5 to 6 years after the suicidal acts
done by early adolescence (up to age 14 years) of school students living
in the open community. Therefore, the outcomes were explored for their
age period between 15 and 20 years. The second objective was to check
for sex differences in the outcomes. It is well known that psychopa-
thology with earlier onset has a worse prognosis (Caspi et al., 1996;
Lewinsohn et al., 1996; Sourander et al., 2009). Hence, a worse progno-
sis was expected among adolescents who started suicidal behavior ear-
lier than those who started such acts during middle adolescence, that is,
between ages 14 and 15 years. We expected that negative outcomes
would be reflected in both psychiatric and psychosocial aspects of their
lives (Brière et al., 2014). Despite its clear clinical relevance, whether
adolescents who are early starters and repeaters of suicidal acts differ
on outcomes in adulthood from those with a single attempt has not
yet been reported. We expected the single attempters to have a better
*National Centre for Suicide Research and Prevention, Institute for Clinical Medicine, Fac-
ulty of Medicine, University of Oslo, Oslo; †Regional Centre for Child and Youth
Mental Health and Child Welfare–Central Norway,Department of Neuroscience, Fac-
ulty of Medicine, Norwegian University of Science and Technology (NTNU); ‡De-
partment of Neuroscience, Faculty of Medicine, Norwegian University of Science
and Technology (NTNU); and §St. Olav's University Hospital, Trondheim, Norway.
Send reprint requests to Latha Nrugham, PhD, National Centre for Suicide Research
and Prevention, Institute for Clinical Medicine, Faculty of Medicine, University of
Oslo, Oslo, Norway. E-mail: nrugham@gmail.com.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
ISSN: 002 2-3018/15/20304–0294
DOI: 10.1097/NMD.0000000000000281
ORIGINAL ARTICLE
294 www.jonmd.com The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

prognosis as compared with the repeaters. Therefore, the third objective
was to examine the outcomes of adolescents by age and attempt status,
such as early starters, repeaters, and incident cases. The last objective
was to adjust the outcomes of suicidal acts for possible influences of
depressive disorders by early adolescence because the empirical lit-
erature demonstrates overlapping features between depressive disorders
or higher depression scores and suicidality. We expected depressive dis-
orders to explain a considerable part of the outcomes of suicidal acts
in adolescence. To complete these objectives, the outcome profile of
adolescent high scorers on depression was studied.
METHODS
Participants and Design
Cluster sampling resulted in a representative sample of 2792 stu-
dents from 22 schools in two counties of Central Norway. Details of the
sampling procedure and the original sample are available (Sund et al.,
2001, 2003). Informed consent during all assessments was obtained
as per the Norwegian Data Inspectorate's norms. The research project
was approved by the Regional Committee for Medical Research Ethics,
Central Norway.
Time Points
At the first assessment, 2464 students (T1, mean age, 13.7 [SD,
0.5] years; 50.8% female; participation, 88.3%) completed a question-
naire, including a screening measure for depression (Mood and Feel-
ings Questionnaire [MFQ], see below). One year later (T2, mean age,
14.9 [SD, 0.5] years; 50.3% female), 2432 adolescents completed the
same questionnaire, also at school.
Subset Selection
At T2, each pair of high scorers (≥26) on the MFQ (described
below) was randomly matched by age and sex with one of the low or
moderate scorers. Among these 364 adolescents selected, 345 (partici-
pation, 94.8%; 72.5% girls) were diagnostically interviewed face-to-
face at school by one of six trained interviewers. High scorers numbered
220 (64%), and the others totaled 125 (36%). In the present study, a
stricter cutoff score of 26 on the MFQ has been used, although a cutoff
of 25 has also been used.
After 5 years (T3), all adolescents of the subset were alive asper the
Norwegian Population Register. Consenting young adults (n= 303) from
the subset were invited for T3, and, as Figure 1 shows, 265 participants
FIGURE 1. Sample flow of the research project “Youth and Mental Health,”follow-up of mainly depressed adolescents from age 14 to 20 years.
The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015 Attempted Suicide in Adolescence
© 2015 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 295
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

were reassessed. The overall participation rate at T3 was 76.9%. The T2
ratio of 1.8:1 between the interviewed high scorers and the others
remained at T3. The mean age of the participants at T3 was 20.0 (SD,
0.6; range, 18.9–21.4) years; 76.9% were women. Filled-in question-
naires were mailed back by 252 of the participants (73%), whereas
242 (70.1%) were interviewed on telephone. The present analyses use
data from these 242 young adults. In the present study, the three phases
of adolescence have been defined as follows: early, between ages 12
and 14 years; middle, between ages 14 and 15 years; and late, between
ages 15 and 20 years, whereas young adulthood is defined as
being 20 years.
Assessments
Questionnaire
The MFQ, developed for ages 8 to 18 years, addresses depres-
sive symptoms based on Diagnostic and Statistical Manual of Mental
Disorders, Revised Third Edition (DSM-III-R) criteria for major depres-
sion (Angold et al., 1987). Participants' feelings and behaviors in the
past 2 weeks are explored by 34 items and rated on a 0-to-2 scale with
a potential total score between 0 and 68. Psychometric properties in the
original sample from which the subset was selected were satisfactory
(Sund et al., 2001). The mean MFQ score for the full original sample
was 10.6 (SD, 9.5; n= 2464), whereas the present subset had a mean
score of 26.4 (SD, 14.6; n= 242) at T1. Details of MFQ scores in this
subset have been reported (Nrugham et al., 2008a, 2008b).
Interview
The Kiddie–Schedule for Affective Disorders and Schizophrenia–
Present and Lifetime version (K-SADS-PL; Kaufman et al., 1997), a
semistructured diagnostic interview, assesses current and past axis I
psychopathology in adolescents according to the DSM-III-R and
DSM-IV criteria (American Psychiatric Association, 1987, 2000). Each
item is rated from 0 to 3, with 3 representing the clinical threshold.
At T2, all parents of the participant subset were invited to be
interviewed separately; 79.8% of the adolescents thus had at least one
additional informant. The T3 telephone interviews with parents and ad-
olescents explored both their current psychopathology and, retrospec-
tively, their psychopathology between ages 15 and 20 years, that is, for
their past 5 years.
The mean interval between the completion of the questionnaires
and the interviews was 20 days at T2 and 21 days at T3. Interviewers
were experienced clinicians trained in the use of the K-SADS-PL. They
were blind at T2 and T3 to the MFQ score status of the participants. At
T3, all interviewers were blind to the T2 diagnostic status and the MFQ
score status of the participants at T2 and T3. The respondents were ran-
domized to one of three interviewers at T3. Interrater reliability for all
affective symptoms, estimated before interviewing started, was good with
Cohen's κat 0.71 at T2 and 0.70 at T3 with the third author (A. M. S.).
Interview integrity was maintained with a mean κof 0.83 at T2 and
0.80 at T3. At T2, the interviewers were allocated participants accord-
ing to geographical proximity. One of the interviewers at T2 was also
an interviewer at T3. However, weighted randomization ensured that
this interviewer did not get any of her T2 interviewees despite a viola-
tion of randomization in 29.2% of the allocations at T3 which took
place for practical reasons related to the completion of the data collec-
tion. Full details of the T3 interviews have been already provided
(Nrugham et al., 2010).
Study Factors
Suicide Attempt Status
The screening section for depression of the K-SADS-PL used
this probe on suicidal acts: “Have you ever (used at T2) or since the last
interview (used at T3) tried to kill yourself or done something which
could have killed you?”An action assessed to be at or greater than the
clinical threshold was scored as a suicidal act, and details about the most
serious act, including its date, were recorded. The item on attempted sui-
cide used in the questionnaire was “Have you ever tried to kill yourself?”
(“no, never”;“yes, once”;“yes, several times”; Wichstrøm, 2000). If en-
dorsed, the participants were to fill out details of the timing of the last
suicidal act: “How long ago was your last act of attempted suicide?”
(“years”;“months”). Acts of deliberate self-harm without suicidal intent
were excluded, differentiated by the respondent or by the interviewer.
The adolescents were grouped into two main groups according
to the timing of the suicidal act. The first main group included all ado-
lescents with a suicidal act by age 15 years (T2; n= 62; 45 girls), hence-
forth referred to as the attempters; they were further subgrouped. The
early starters were all those with a suicidal act by age 14 years (T1;
n= 35; 28 girls). The repeat ers were a pooled group because it also
included the early starters with repeated suicidal acts between 14 and
15 years old (T1–T2; n= 15; 13 girls). The incident cases were adoles-
cents who had attempted suicide for the first time between 14 and
15 years old (T1–T2; n= 19; 12 girls).
The second main group included all adolescents who had
attempted suicide between ages 15 and 20 years (T2–T3; n= 34; 30 fe-
males), here referred to as late adolescence. Membership in this group
was one of the outcome variables studied.
Female attempters were all female adolescents with suicidal acts
by age 15 years (n= 45), regardless of whether they were early starters,
repeaters, or cases incident between ages 14 and 15 years. Similarly, the
male attempters (n= 17) were all male adolescents with suicidalacts by
age 15 years, regardless of attempt status.
Outcome Variables Between T2 and T3
Psychiatric diagnoses and psychosocial variables (details below)
from the follow-up period, that is, between T2 and T3, were used as out-
come variables or end points.
Psychiatric Diagnoses
Depressive disorders included MDD and dysthymia. A diag-
nosis of depressive disorder not otherwise specified required a mini-
mum of three depressive symptoms to have lasted for at least 2 weeks.
Other psychiatric disorders included posttraumatic stress disorder
(PTSD), conduct disorder, substance use disorders, and alcohol abuse
or dependence, chosen for their documented associations with suicidality.
Current and past psychiatric diagnoses were pooled for the two interview
periods: by age 15 years and between ages 15 and 20 years. Remission
from any of these diagnoses required at least 2 monthswithout symptoms,
as per the DSM-IV-TR criteria (American Psychiatric Association, 2000).
Psychosocial Outcomes
Outcomes Between T2 and T3 (Between Ages 15 and 20 Years)
The participants were asked whether their leisure and education
or work activities had been disrupted at least for a month or more be-
cause of mental health issues (“break in leisure activities,”“break in
work or education related activities”). Contact with child protection ser-
vices (“child protection services”); outpatient mental health care ser-
vices, either at the primary or specialist levels (“psychiatric OPD”);
admission to psychiatric or somatic hospitals because of psychiatric
complaints (“psychiatric admission”); and use of medication because
of psychiatric (“psychotropic medication”) or somatic complaints (“so-
matic medication”) were recorded. All variables mentioned above in
this paragraph were scored in a dichotomous way (yes/no).
Outcome Variables at T3 (Age 20 Years)
The participants' residential, employment, and financial statuses
(“living alone or not,”“working now or not,”“financial problems or
not”) were probed. The fourth variable checked was whether the partic-
ipant had been offered a referral to the mental health care services at
age 20 years (“offered referral or not”). The Children's Global
Nrugham et al. The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015
296 www.jonmd.com © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Assessment of Function (C-GAS; Shaffer et al., 1983) was used to as-
sess the participant's current functioning and symptom severity on a
scale from 1 to 100, with higher scores indicating both better function-
ing and lower symptom levels. The mean C-GAS score at T2 among the
adolescents who fulfilled the diagnostic criteria for any depressive dis-
order by age 15 years was 61.0 (SD, 9.1).
Statistical Analyses
Missing values were less than 5% and imputed with the expectation-
maximization or the regression method, as per the indicators of the
Little's Chi-square test (Tabachnick and Fidell, 2006). Chi-square and
Fisher's exact tests were used to examine associations between categor-
ical variables. Differences between group means of continuous vari-
ables were checked by Student's t-test. Outcome variables that emerged
as significant in the cross-tabulations and the t-tests were entered into
bivariate logistic regression analysis to obtain crude odds ratios (ORs)
for all attempters. Comparisons by attempt status were conducted. Fe-
male and male attempters were compared with the nonattempters within
their sex only and by pooling them regardless of whether they were
early starters, repeaters, or incident cases between ages 14 and 15 years.
The significant bivariate relationships thus obtained were adjusted for
the influence of any depressive disorder by age 15 years as a covariate
in block 2. The standard method of entry was used in the analyses. The
value was p<0.05.
RESULTS
Five-Year Outcomes of Suicide Attempters
Outcomes between ages 15 and 20 years predicted by a suicidal
act before having reached age 15 years included not only all psychiatric
diagnostic variables but also a later suicidal act after age 15 years, that
is, between T2 and T3. In addition, outcomes included breaks in leisure
and educational or work activities as well as contact with the child pro-
tection services and the outpatient mental health care services (Table 1).
At T3, the attempters had lower C-GAS scores than the nonattempters:
73.5 (SD, 11.6) vs. 80.3 (SD, 8.3), t(240) = 5.1, p< 0.0005, Cohen's d=
0.3. The attempters were also more likely to have been offered referrals
to the mental health care services and to report financial problems at
T3 than the nonattempters.
Impaired Functioning by Attempt Status and Sex
Regardless of attempt status, the suicide attempters had consis-
tently lower C-GAS scores than the nonattempters, despite significant
improvement at age 20 years (T3). Improvements were most pronounced
for incidentcases between ages 14 and 15 years; their scores at T3 were
no longer significantly different from the nonattempters (Table 2).
Early Adulthood Outcomes by Attempt Status and Sex
Crude and Adjusted Analyses
When all those who had attempted suicide by age 15 years were
pooled together (n= 62), only 2 of the 16 outcome variables reached in-
significance: psychiatric admission and the use of psychotropic medi-
cation between the ages 15 and 20 years (Table 3). When analyzed
according to the subgroup status for attempters, only one of the out-
come variables reached significance for incident cases (n= 19); they
were more likely to be offered referrals at age 20 years, OR (95% con-
fidence interval [CI]), 2.6 (1.02–6.8). Dysthymia between ages 15 and
20 years was specific only for the female attempters. Similarly, MDD,
use of psychiatric medication between the ages of 15 and 20 years,
TABLE 1. Comparing Early Suicide Attempters and Nonattempters: Frequencies and Percentages of Psychiatric and Psychosocial Outcomes
5 Years After Early Suicide Attempts Made Before Age 15 Years in a Subset of School Adolescents (n=242)
SA (n=62),n(%) Non-SA (n=180),n(%) χ
2
Psychiatric outcomes between ages 15 and 20 yrs
Suicidal acts 21 (33.9) 13 (7.2) 27.1*
MDD 32 (51.6) 46 (25.6) 14.3*
Dysthymia
a
5 (8.1) 3 (1.7) 0.02
PTSD 19 (30.6) 18 (10.0) 15.1*
Conduct disorder 12 (19.4) 15 (8.3) 5.6**
Alcohol disorder 16 (25.8) 19 (10.6) 8.6***
Substance disorder
a
9 (14.5) 6 (3.3) 0.004
Break in leisure activities
a
11 (17.7) 13 (7.2) 5.7**
Break in work/educational activities 16 (25.8) 17 (9.4) 10.4***
Child protection services 9 (25.8) 16 (5.0) 21.5*
Outpatient–mental health services 29 (46.8) 42 (23.3) 12.2*
Psychiatric admission
a
7 (11.3) 9 (5.0) NS
Psychotropic medication
a
6 (9.7) 7 (3.9) NS
Somatic medication 33 (53.2) 94 (52.2) NS
Psychosocial outcomes at age 20 yrs
Living alone 15 (24.2) 28 (15.6) NS
Working 42 (67.7) 111 (61.7) NS
Financial problems 15 (24.2) 13 (7.2) 12.9*
Offered referrals 31 (50) 45 (25.0) 13.3*
Associations measured by the chi-square or Fisher's exact tests. df for the chi-square test was 1.
Non-SA indicates nonattempters; NS, not significant; SA, suicide attempters.
a
Fisher's exact test.
*p<0.001; **p<0.05; ***p<0.01.
The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015 Attempted Suicide in Adolescence
© 2015 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 297
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

and lower global functioning scores on the C-GAS at age 20 years were
significant outcomes only for the early starters but not for the repeaters.
Two outcomes were significant for the high scorers on depression at
age 15 years: contact with the outpatient mental health care services,
OR (95% CI), 2.2 (1.2–4.2), and contact with child protection services,
OR (95% CI), 3.0 (1.01–9.2). One outcome reached significance for
the male attempters; they were more likely to have been offered a re-
ferral for mental health care services at age 20 years, OR (95% CI),
5.1 (1.4–18.0).
Each bivariate estimate of the relationship between the early ad-
olescent suicidal act and its later outcome was adjusted for the influence
of any depressive disorder by age 15 years (Table 3). All outcomes for
the early starters remained significant. For the repeaters, two outcomes
lost their significance when thus controlled: contact with the mental
health care services and psychiatric admission. For the female attempters,
all the bivariate outcomes remained significant in the multivariate
analyses except for conduct disorder, psychiatric admission, and the
use of psychotropic medication. Similarly, the single bivariately signif-
icant outcome for the male attempters remained significant, OR (95%
CI), 3.8 (1.01–14.9). Overall, the multivariate relationships remained
robust, despite a reduction in strength.
DISCUSSION
The present study has four main findings, all new and confirm-
ing our hypotheses. Outcomes of suicidal acts carried out by early
adolescence are expressed both in psychiatric and psychosocial mani-
festations. Those who attempt suicide by early adolescence, including
the repeaters, are more likely to have been in contact with the child pro-
tection services than with the mental health care services. Outcomes of
suicidal acts by early adolescence differ by sex and attempt status. Male
attempters and those with suicidal acts initiated in middle adolescence,
defined as the time between ages 14 and 15 years, have better prognoses
than the early starters and the repeaters. Lastly, depressive disorders by
the age of 15 years are associated with the outcomes of adolescent sui-
cidal acts,and even thosewithout a diagnosis of depressive disorder but
withonlydepressivesymptomsarelikelytohavebeenincontactwith
mental health services and child protection services.
Seven outcomes were common to all groups of attempters. Be-
tween ages 15 and 20 years, attempters are more likely than non-
attempters to report a repeated suicidal act; have PTSD, alcohol abuse,
or other substance disorders; report breaks in their leisure, educational,
or work activities; and also to have been in contact with the child pro-
tection services. At age 20 years, attempters are more likely to report
financial problems than others. However, only early starters are likely
to come in contact with the mental health care services, either as outpa-
tients or inpatients, and also to receive psychotropic medication. When
adjusted for any depressive disorder, the relationship of contact with
mental health services and the repeated suicidal acts during adoles-
cence was no longer significant, indicating that this association is statis-
tically explained by the presence of a depressive disorder. However, the
relationship between these depressed adolescents and the occurrence
of contact with the child protection services continued to remain signif-
icant even after adjusting for any depressive disorder, indicating it to
be a stable relationship. Clients of the child protection services have in-
creased suicidality-related psychiatric morbidity and mortality (Kalland
et al., 2001; Vinnerljung et al., 2006) and should be considered a group
at high risk, especially the girls.
In addition to the later attempted suicide, attempters are more
likely to be diagnosed not only with depressive disorders but also with
PTSD and substance abuse disorders by the time they reach adulthood.
Despite the diversity in sex proportions, age ranges, and follow-up
periods, school-based and clinical samples have reported similar out-
comes for various aspects of suicidality (Fergusson et al., 2005b; Grøholt
and Ekeberg, 2009; Herba et al., 2007; Lewinsohn et al., 2001; Pfeffer
et al., 1993; Reinherz et al., 2006; Steinhausen and Winkler Metzke,
2004) and depression (Fergusson et al., 2005a; Fombonne et al., 2001a,
2001b; Weissman et al., 1999). Although appropriate care was taken to
record remissions and new episodes of depressive disorders, because of
recall bias, it is also possible that a full remission had not happened.
Therefore, this finding of depressive disorders being one of the outcomes
associated with an early adolescent suicidal act needs to be interpreted
with caution. It is also to be noted that the Oregon Adolescent Depression
Project did not find such an outcome when its adolescents were followed
up at age 30 years (Brière et al., 2014).
The present study design has three added benefits as compared
with other studies. It explores in distinct phases the entire time span be-
tween adolescence into early adulthood. Separate outcome profiles by
sex, attempt status, and high scorers on depression at age 15 years are
provided. Thereby, this study can specify some of the outcomes for
the subgroups, for example, PTSD for repeaters,dysthymia for females,
and mental health care services referrals at adulthood for males. Dys-
thymia and substance use disorders have been reported in epidemiolog-
ical samples as outcomes specific for female attempters (Fergusson
et al., 2005b). Lastly, all outcomeprofiles were adjusted for the possible
influence of depressive disorders, revealing their contribution to these
outcomes and also the outcomes specific to only suicidal acts during
adolescence. Although small, these changes in the adjusted outcomes
TABLE 2. Results of Cross-sectional and Longitudinal Comparisons (t-Tests): Mean (Standard Deviation) Scores of the C-GAS by Suicide
Attempt Status and Sex in a Subset of Mainly Depressed School Adolescents (n= 242) Followed Up as Young Adults (Age 20 Years)
Early Starters Repeaters Incident Cases Females Males
SA Non-SA SA Non-SA SA Non-SA SA Non-SA SA Non-SA
n=35 n=207 n=15 n=227 n=19 n=223 n=45 n=141 n=17 n=39
Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD)
Age 15 68.8 (9.7)* 79.3 (11.6) 63.4 (7.2)* 78.7 (11.5) 70.0 (10.4)** 78.5 (11.8) 68.0 (10.2)** 81.0 (11.1) 69.8 (8.4)* 80.8 (10.1)
Age 20 72.6 (11.2)*79.5 (9.0)*72.6 (8.1)*78.9 (9.6)*77.5 (9.1)*78.6 (9.7) 74.0 (10.9)*80.3(7.7)71.8 (13.2)* 80.3 (10.1)
Early starters group comprises adolescents with one or more suicidal acts by age 14 years; repeaters group, those with one or more suicidal acts by age 14 years and
with at least one additional act between ages 14 and 15 years; incident cases group, adolescents with their first suicidal act between ages 14 and 15 years.
Non-SA indicates nonattempters; SA, adolescents with suicidal acts.
Italics indicates significant difference between scores of SA and non-SA at T2.
Bold indicates significant difference between SA at age 15 years and SA at age 20 years (longitudinal).
Bold italics indicates significant difference between scores of SA and non-SA at age 20 years.
*p < 0.001; **p <0.01.
Nrugham et al. The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015
298 www.jonmd.com © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

reveal that the prognosis of adolescent suicide attempters may improve
with timely treatment of depression. Suicidal behavior being a complex
temporal process, early detection and appropriate treatment of psychiat-
ric disorders may lead to a decline in the suicidal occurrences, which
can also eventually subside (Wasserman et al., 2012).
In the present study, the percentage of attempters diagnosed with
MDD between ages 15 and 20 years was the same as that in the 9-year
follow-up study (Grøholt and Ekeberg, 2009) of inpatients: 51%. The
percentages of those who were in contact with the outpatient mental
health care services are also similar in the two samples: 48% vs.
46.8%. This indicates that the prognostic profile of the attempters
drawn from school-going adolescents is comparable with the inpatients
on these two outcomes. The present study confirms the finding of
MDD being related to adolescent suicide attempts (Fergusson et al.,
2005b). Females and early starters, irrespective of sex, are more likely
to have MDD than repeaters, incident cases, and nonattempters. Dys-
thymia is not only a predictor of adolescent suicidal acts (Nrugham
et al., 2008b) but also an outcome of attempted suicide by early
adolescence.
At age 20 years, the attempters did not differ in employment or
residential status from the nonattempters. Despite improvements over
the years, differences in the global functioning reveal that some impair-
ment persists among the attempters. Maladaptive coping with stressful
life events has been linked to suicidal behavior among adolescents, not
only among boys (Lewinsohn et al., 2001) but also among girls (Nrugham
et al., 2012). The prognostic profiles of the early starters and the re-
peaters were more pervasive than that of the incident cases. Follow-
up studies of early and recurrent depression report a similar outcome
profile (Fergusson et al., 2005a; Fombonne et al., 2001a, 2001b). Fe-
male adolescents were more likely to repeat suicidal acts (Fergusson
et al., 2005b; Lewinsohn et al., 2001) and to have multiple psychiatric
diagnoses yet little contact with the mental health care services; they are
more likely to have been in contact with the child protection services. A
longitudinal register study (Vinnerljung et al., 2006) on the clients of
TABLE 3. Outcomes of Suicidal Acts by Ages 14 and 15 Years in a Subset of School Adolescents (n= 242) Followed Up at Age 20 Years and
Adjusted for Influence of Any Depressive Disorder by Age 15 Years (T2)
Early Starters by
Age 14, n=35
Repeaters Between
Ages 14 and 15, n= 15 All SA by Age 15, n=62
Sex: Females by
Age 15, n=45
Outcomes Between Ages 15 and 20 OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Suicidal acts 7.4 (3.2–16.8) 4.7 (1.5–14.3) 6.5 (3.0–14.2) 8.6 (3.6–20.2)
Adjusted 6.7 (2.8–15.9) 4.7 (1.4–15.1) 5.0 (2.2–11.2) 6.70 (1.1–7.0)
MDD 3.4 (1.6–7.1) 3.1 (1.7–5.6) 3.3 (1.6–6.6)
Adjusted 3.0 (1.4–6.5) 2.3 (1.2–4.4) 2.5 (1.2–5.2)
Dysthymia 5.1 (1.1–22.3) 6.7 (1.1–38.5)
Adjusted 6.7 (1.1–41.1)
PTSD 4.5 (2.0–10.0) 7.8 (2.6–23.1) 3.9 (1.9–8.2) 4.8 (2.2–10.8)
Adjusted 4.0 (1.7–9.2) 7.7 (2.5–23.6) 3.3 (1.5–7.1) 4.3 (1.9–9.8)
Conduct disorder 2.9 (1.1–7.3) 9.0 (2.9–27.5) 2.6 (1.1–6.0) 3.0 (1.04–8.9)
Adjusted 2.7 (1.09–7.0) 8.8 (2.8–27.0) 2.5 (1.06–6.0)
Alcohol disorder 3.4 (1.5–8.0) 4.5 (1.5–13.7) 2.9 (1.4–6.1) 3.4 (1.4–8.5)
Adjusted 3.2 (1.3–7.5) 4.3 (1.4–13.3) 2.6 (1.2–5.8) 3.2 (1.2–8.5)
Substance disorder 3.6 (1.2–10.7) 8.9 (2.6–30.6) 4.7 (1.7–13.1) 11.5 (2.9–44.6)
Adjusted 3.1 (1.04–9.3) 8.7 (2.4–30.7) 3.7 (1.2–11.0) 10.3 (2.5–42.5)
Break in work or educational activities 3.8 (1.6–8.9) 3.5 (1.1–11.5) 3.3 (1.5–7.1) 4.5 (1.9–10.4)
Adjusted 3.2 (1.3–7.9) 3.4 (1.01–11.5) 2.2 (1.01–5.1) 3.1 (1.2–7.5)
Break in leisure activities 3.5 (1.3–9.0) 5.4 (1.6–17.6) 2.7 (1.1–6.5) 4.2 (1.6–10.8)
Adjusted 2.8 (1.06–7.7) 5.4 (1.5–19.3) 2.8 (1.06–7.8)
Child protection services 5.1 (2.0–12.6) 7.2 (2.3–22.7) 6.6 (2.7–15.9) 8.3 (3.2–21.3)
Adjusted 4.5 (1.8–11.4) 7.0 (2.2–22.6) 5.8 (2.3–14.6) 7.9 (2.9–21.3)
Outpatient mental health services 2.6 (1.2–5.5) 2.9 (1.03–8.5) 2.8 (1.5–5.2) 3.4 (1.7–6.9)
Adjusted 2.3 (1.09–5.0) 2.2 (1.1–4.2) 2.8 (1.3–5.9)
Psychiatric admission 5.5 (1.8–15.9) 4.1 (1.03–16.4) 3.06 (1.04–8.9)
Adjusted 4.8 (1.6–14.4)
Psychotropic medication 5.9 (1.8–18.8) 3.4 (1.05–11.3)
Adjusted 4.7 (1.4–15.8)
Financial problems 3.4 (1.4–8.3) 4.4 (1.3–14.1) 4.1 (1.8–9.2) 3.9 (1.6–9.4)
Adjusted 2.8 (1.1–7.3) 4.2 (1.2–14.3) 3.0 (1.2–7.2) 2.8 (1.1–7.4)
Psychiatric referrals offered 3.0 (1.6–5.4) 2.5 (1.2–5.1)
Adjusted 2.5 (1.3–4.7) 2.2 (1.09–4.6)
Early starters comprise adolescents with suicidal acts by age 14 years; repeaters, adolescents with suicidal acts by age 14 years and an additional act between ages 14
and 15 years; incident cases, adolescents with their first suicidal act between ages 14 and 15 years.
Results of logistic regression analyses: OR (95% CI).
Adjusted, outcomes of suicidal acts between ages 15 and 20 years (T2–T3) controlled for influence of any depressive disorder by age 15 years.
SA indicates suicidal acts.
The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015 Attempted Suicide in Adolescence
© 2015 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 299
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

the child protection services did not find such sex differences, perhaps
because of a different sex distribution and the exclusion criteria used.
Although the present study is prospective, the diagnostic inter-
view data gathered at T3 were in part retrospective, and a recall bias
is possible in the estimates of the number of depressive symptoms
and in their timing. The attrition rate during follow-up was higher than
in the earlier assessments. However, 70% of the adolescents interviewed
at T2 were reinterviewed 5 years later. Because the total number of boys
was less than that of girls, the risk for type II error cannot be ruled out.
Anxiety disorders, with the exception of PTSD, were not chosen as one
of the possible outcomes because of the absence of associations with
attempted suicide in this sample, in the exploratory analyses done for
previous studies using this data set.
This study has substantial strengths. A subset of early adolescents
with high scores on depression derived from a large and representative
school-based sample was followed up from adolescence to adulthood.
Blind and randomized diagnostic interviews by clinicians with excellent
interview integrity at both interview assessment times minimize the pos-
sibility of systematic errors. Stringent criteria were used to define sui-
cidal acts by excluding acts of self-harm without suicidal intent and
also by including only those acts that were at or greater than the clinical
threshold in the interview data. The sample resembles adolescent outpa-
tients in several ways because of the subset's selection strategy. The current
knowledge about their suicidal behavior is sparse (Bridge et al., 2006).
A broader outcome perspectivewith age and sex variations in re-
lation to attempt status should be used in research and clinical settings.
The present study reveals a wider range of psychiatric diagnoses than
depression as the outcome for suicidal adolescents, in linewiththe find-
ings of another prospective study of adult outcomes of adolescent
suicidality (Brière et al., 2014).
School health personnel and clinicians assessing suicidal risk
among adolescents should be alerted to the importance of a history of
suicidal acts by early adolescence. Our findings indicate that the out-
comes of early starters and repeaters have a prognosis worse than that
of those who attempt suicide for the first time between the ages of 14
and 15 years, and this finding is important in terms of policy planning
for the health services. To increase timely and adequate referrals to the
mental health care services, officers at the child protection services
should be trained to probe for suicidal behavior in their assessments
among youth, especially girls.
DISCLOSURES
The authors declare no conflict of interest.
REFERENCES
American Psychiatric Association (1987) Diagnostic and statistical manual of mental
disorders (3rd ed–rev). Washington, DC: American Psychiatric Association.
American Psychiatric Association (2000) Diagnostic and statistical manual of mental
disorders (4th ed–text rev). Washington, DC: American Psychiatric Association.
Angold A, Costello EJ, Pickles EJ, Winder F (1987) The development of a question-
naire for use in epidemiological studies of depression in children and adolescents.
London: Medical Research Council.
Bridge JA, Goldstein TR, Brent DA (2006) Adolescent suicide and suicidal behaviour.
J Child Psychol Psychiatry.47:372–394.
Brière FN, Rohde P, Seeley JR, Klein D, Lewinsohn PM (2014) Adolescent suicide
attempts and adult adjustment. Depress Anxiety. 2014. doi:10.1002/da.22296.
[Epub ahead of print].
Caspi A, Moffitt TE, Newman DL, Silva PA (1996) Behavioral observations at age
3 years predict adult psychiatric disorders. Longitudinal evidence from a birth co-
hort. Arch Gen Psychiatry.53:1033–1039.
Fergusson DM, Horwood LJ,Ridder EM, BeautraisAL (2005a) Sub-threshold depres-
sion in adolescence and mental health outcomes in adulthood. Arch Gen Psychia-
try.62:66–72.
Fergusson DM, Horwood LJ, Ridder EM, Beautrais AL (2005b) Suicidal behaviour in
adolescence and subsequent mental health outcomes inyoung adulthood. Psychol
Med.35:983–993.
Fombonne E, Wostear G, Cooper V, Harrington R, Rutter M (2001a) The Maudsley
long-term follow-up of child and adolescent depression. 1. Psychiatric outcomes
in adulthood. Br J Psychiatry.179:210–217.
Fombonne E, Wostear G, Cooper V, Harrington R, Rutter M (2001b) The Maudsley
long-term follow-up of child and adolescent depression. 2. Psychiatric outcomes
in adulthood. Br J Psychiatry.179:218–223.
Grøholt B, Ekeberg O (2009) Prognosis after adolescent suicide attempt: Mental
health, psychiatric treatment, and suicide attempts in a nine-year follow-up study.
Suicide Life Threat Behav.39:125–136.
Hawton K, Saunders KE, O'Connors RC (2012) Self-harm and suicide in adolescents.
Lancet.379:2373–2382.
Herba CM, Ferdinand RF, van der Ende J, Verhulst FC (2007) Long-term associa-
tions of childhood suicide ideation. J Am Acad Child Adolesc Psychiatry. 46:
1473–1481.
Kalland M, Pensola TH, Merilainen J, Sikkonen J (2001) Mortality in children regis-
tered in the Finnish child welfare registry: A population based study. BMJ.323:
207–208.
Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P, Williamson D, Ryan N
(1997) Schedule for Affective Disorders and Schizophrenia for School-Age
Children–Present and Lifetime Version (K-SADS-PL): Initial reliability and va-
lidity data. J Am Acad Child Adolesc Psychiatry. 36:980–988.
Kovacs M, Goldston D,Gatsonis C (1993) Suicidal behaviors and childhood-onset de-
pressive disorders:A longitudinal investigation. J Am Acad Child Adolesc Psychi-
atry.32:8–20.
Lewinsohn PM, Rohde P, Seeley JR (1996) Adolescent suicidal ideationand attempts:
Prevalence, risk factors, and clinical implications. Clin Psychol.3:25–46.
Lewinsohn PM, Rohde P, Seeley JR, Baldwin CL (2001)Gender differencesin suicide
attempts from adolescence to young adulthood. J Am Acad Child Adolesc Psychi-
atry.40:427–434.
Nrugham L, Holen A, Sund AM (2010) Associations between attempted suicide,
violent life events, depressive symptoms, and resilience in adolescents and
young adults. JNervMentDis198:131–136. Erratum. JNervMentDis.
198:389.
Nrugham L, Holen A, Sund AM (2012) Suicide attempters and repeaters: Depression
and coping: A prospective study of early adolescents followedup as young adults.
JNervMentDis.200:197–203.
Nrugham L, Larsson B, Sund AM (2008a) Predictors of suicidal acts across ado-
lescence: Influences of familial, peer and individual factors. J Affect Disord.
109:35–45.
Nrugham L, Larsson B, Sund AM (2008b) Specific depressive symptoms and disor-
ders as associates and predictors of suicidal acts across adolescence. J Affect
Disord.111:83–93.
Patton GC, Coffey C, Sawyer SM, Viner RM, Haller DM, Bose K, Vos T, Ferguson J,
Mathers CD (2009) Global patterns of mortality in young people: A systematic
analysis of population health data. Lancet.374:881–892.
Pfeffer CR, Klerman GL, Hurt SW, Kakuma T, Peskin JR, Siefker CA (1993)Suicidal
children grow up: Rates and psychosocial risk factors for suicide attempts during
follow-up. J Am Acad Child Adolesc Psychiatry.32:106–113.
Reinherz HZ, Tanner JL, Berger SR, Beardslee WR, Fitzmau rice GM (2006)
Adolescent suicidal ideation as predictive of psychopathology, suicidal be-
haviour and compromised functioning at age 30. Am J Psychiatry. 163:
1226–3122.
Shaffer D, Gould MS, Brasic J, Ambrosini P, Fisher P, Bird H, Aluwahlia S (1983) A
children's global assessment scale (CGAS). Arch Gen Psychiatry.40:1228–1231.
Sourander A, Klomek AB, Niemelä S, Haavisto A, Gyllenberg D, Helenius H,
Sillanmäki L, Ristkari T, Kumpulainen K, Tamminen T, Moilanen I, Piha J,
Almqvist F, Gould MS (2009) Childhood predictors of completed and severe
suicide attempts: Findings from the Finnish 1981 Birth Cohort Study. Arch
Gen Psychiatry. 66:398–406.
Nrugham et al. The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015
300 www.jonmd.com © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Steinhausen HC, Winkler Metzke CW (2004) The impact of suicidal ideation in pre-
adolescence, adolescence and young adulthood on psychosocial functioning and
psychopathology in young adulthood. Acta Psychiatr Scand.110:438–445.
Sund AM, Larsson B, Wichstrøm L (2001) Depressive symptoms among young Nor-
wegian adolescents as measured by the Mood and Feelings Questionnaire (MFQ).
Eur Child Adolesc Psychiatry.10:222–229.
Sund AM, Larsson B, Wichstrøm L (2003) Psychosocial correlates of depressive symptoms
among 12-14-year-old Norwegian adolescents. J Child Psychol Psychiatry.44:588–597.
Tabachnick BG, Fidell L (2006) Using multivariate statistics (5th ed). Boston: Allyn
&Bacon.
Thompson AH (2008) Younger onset of depression is associated with greater suicidal
intent. Soc Psychiatry Psychiatr Epidemiol.43:538–544.
Vinnerljung B, Hjern A, Lindblad F (2006) Suicide attempts and severe psychiatric
morbidity among former child welfare clients—A national cohort study. JChild
Psychol Psychiatr.47:723–733.
Wasserman D, Rihmer Z, Rujescu D, Sarchiapone M, Sokolowski M, Titelman D,
Zalsman G, Zemishlany Z, Carli V (2012) The European Psychiatric Associa-
tion (EPA) guidance on suicide treatment and prevention. Eur Psychiatry. 27:
129–141.
Weissman MM, Wolk S, Goldstein RB, Moreau D, Adams P, Greenwald S, Klier CM,
Ryan ND, Dahl RE, Wickramaratne P (1999) Depressed adolescents grown up.
JAMA.281:1707–1713.
Wichstrøm L (2000) Predictors of adolescent suicide attempts: A nationally represen-
tativelongitudinal study of Norwegianadolescents. J Am Acad Child Adolesc Psy-
chiatry.39:603–610.
The Journal of Nervous and Mental Disease •Volume 203, Number 4, April 2015 Attempted Suicide in Adolescence
© 2015 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 301
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.