Risk for Suicide Attempts
Among Adolescents Who
Engage in Non-Suicidal
Self-Injury

Jennifer J. Muehlenkamp and Peter M. Gutierrez

The current study examined whether common indicators of suicide risk differ between
adolescents engaging in non-suicidal self-injury (NSSI) who have and have not
attempted suicide in an effort to enhance clinicians’ ability to evaluate risk for suicide
within this group. Data were collected from 540 high school students in the Midwest
who completed the RADS, RFL-A, SIQ, and SHBQ as part of a larger adolescent
risk project. Results suggest that adolescents engaging in NSSI who also attempt
suicide can be differentiated from adolescents who only engage in NSSI on measures
of suicidal ideation, reasons for living, and depression. Clinical implications of the
findings are discussed.

Keywords

adolescents, deliberate self-harm, non-suicidal self-injury, suicide

There is growing interest in understanding
the psychological correlates and psychiatric
morbidity of non-suicidal self-injury (NSSI)
in adolescents. Prevalence rates of NSSI
among adolescent samples range from
14% to 40% in the community (Muehlen-
kamp & Gutierrez, 2004; Ross & Heath,
2002) and from 40% to 61% in inpatient
samples

(Darche,

1990;

DiClemente,

Ponton, & Hartley, 1991). There is some
evidence to suggest that the incidence of
NSSI may be increasing (Hawton, Fagg,
Simkin et al., 1997; Olfson, Gameroff,
Marcus et al., 2005) within this age group.
The increasing rate of NSSI is of particular
concern because some individuals with his-
tories of NSSI are at greater risk for suicide
(Dulit, Fryer, Leon et al., 1994; Zlotnick,
Donaldson, Spirito et al., 1997), although
not all self-injurers will attempt suicide

(Kessler, Borges, & Walters, 1999). Research
to date has not progressed far beyond docu-
menting an association between increased
risk for suicide and NSSI, limiting clinicians’
ability to evaluate risk for suicide attempts
within

this

at-risk

group.

Therefore,

additional studies examining psychological
correlates that may differentiate adolescents
engaged in NSSI who attempt suicide from
those who do not attempt suicide are needed
to better inform current knowledge and
clinical practice.

Non-suicidal self-injury is conceptua-

lized as behavior existing along a continuum
of self-harm on which suicide is the final and
most severe endpoint (O’Carroll, Berman,
Marris et al., 1996; Stanley, Winchel, Mol-
cho et al., 1992). Consequently, there is a
potential for shared psychiatric morbidity
and other risk factors underlying both NSSI

Archives of Suicide Research, 11:69–82, 2007
Copyright # International Academy for Suicide Research
ISSN: 1381-1118 print/1543-6136 online
DOI: 10.1080/13811110600992902

69

and suicide attempts. For example, research
has documented that childhood sexual
abuse, depression, interpersonal or family
conflict, isolation or loneliness, impulsivity,
borderline personality disorder, and psychi-
atric illness are among the risk factors for
both NSSI and suicidal behavior (Brent,
Perper, Mortiz et al., 1993b; Garrison,
Addy, McKeown et al., 1993; Maris, Ber-
man, & Silverman, 2000; Walsh, 2005).
These shared risk factors make it difficult
to clarify potential suicide risk among indi-
viduals engaged in NSSI, and it is unknown
whether there are gradations of severity on
these variables between those with a history
of NSSI and=or suicide attempts. Identifying
potential differences in symptom severity
among individuals who engage in NSSI
and=or suicide attempts is important to
improving risk assessment, but there is little
research in this area.

We were able to identify only two stu-

dies (Guerton, Lloyd-Richardson, Spirito
et al., 2001; Stanley, Gameroff, Michalsen
et al., 2001) that have examined potential
differences

in

psychosocial

correlates

between

NSSI

and

suicide

attempts.

Stanley and colleagues (2001) examined
potential psychosocial differences in 53
adult women with borderline personality
disorder who presented to a psychiatric
treatment

facility

following

a

suicide

attempt. Individuals with both NSSI and
suicide attempts were found to have signifi-
cantly higher levels of depression, hope-
lessness, aggression, anxiety, impulsivity,
and suicidal ideation. Guertin et al. (2001)
conducted a similar study in a sample of
95 adolescents presenting to a child psychi-
atric clinic following a suicide attempt, of
which 54.7% had a history of at least one
act of NSSI. Results indicated that indivi-
duals who had engaged in both NSSI
and had made a suicide attempt were more
likely

to

be

diagnosed

with

major

depression, dysthymia, or oppositional
defiant disorder. Additionally, the NSSI
and suicide attempt group had significantly

higher levels of depression, loneliness,
anger, and risk taking behaviors than the
suicide attempt only group.

These findings enhance our under-

standing of suicide attempters, confirming
ideas that individuals who engage in a range
of self-harming behaviors (e.g., NSSI, risk
taking, suicide attempts) are likely to
experience increased psychological impair-
ment. However, neither study provided
data that would improve understanding of
psychiatric differences between individuals
engaged in NSSI who do and do not
attempt suicide. Furthermore, these studies
did not explore what aspects of the psychi-
atric disorders or psychosocial correlates
may have contributed to increased risk
for a suicide attempt. The findings are also
restricted to clinical populations who were
accessing treatment, making it difficult to
determine whether similar patterns of
results would generalize beyond such sam-
ples. Given the increased prevalence of
NSSI within community samples of adoles-
cents, it is imperative to expand the find-
ings by Guertin et al. (2001) and Stanley
et al. (2001) to non-clinical samples, as well
as to determine if similar patterns of suicide
risk exist for adolescents who have a his-
tory NSSI.

Another reason to further examine dif-

ferences in risk factors such as depression
and suicidal ideation within samples of
individuals who engage in NSSI, is that
NSSI is identified as a potential precursor
to suicide attempts. Joiner (2005) has pro-
posed a model of suicide in which he posits
that participation in self-injurious behaviors
over time desensitizes an individual to self-
harm. This desensitization to self-harm is
one mechanism which increases risk for
suicide attempts because the person has
habituated to fears and physical pain asso-
ciated with self-injury. Consequently, a
person with a history of NSSI more readily
acquires the capacity to engage in lethal
acts of self-harm. This understanding of
NSSI’s potential relationship to suicide

Risk For Suicide

70

VOLUME 11 NUMBER 1 2007

supports the need to further clarify indica-
tors of risk that may signal a shift from
NSSI to a suicide attempt.

Examining potential differences in

levels of depression and suicidal ideation
between

self-injuring

adolescents

who

attempt suicide and those who do not can
also inform clinical practice. However,
understanding of NSSI and risk for suicide
within this group may be improved by
examining additional factors that could
shed light on suicide risk. Reasons for liv-
ing have been identified as important pro-
tective

factors

against

suicide-related

behavior in both adolescent and adult sam-
ples (e.g., Gutierrez, Osman, Kopper, et al.,
2000; Osman, Downs, Kopper et al., 1998).
Reasons for living have also been shown to
accurately differentiate adolescents at high
and low risk for suicide (Gutierrez, Osman,
Kopper et al., 2000); however there are no
known studies that have examined how
reasons for living may affect suicide risk
in self-injuring samples. From a theoretical
viewpoint, NSSI is often conceptualized as
an emotion regulation strategy (Nock &
Prinstein, 2004), suggesting that individuals
engaged in NSSI are motivated to live but
have difficulty coping with distress. Con-
sistent with this hypothesis, Muehlenkamp
and Gutierrez (2004) found that adoles-
cents who had a history of NSSI had more
positive attitudes towards life than adoles-
cents who attempted suicide. It is possible
that individuals engaged in NSSI would
identify greater numbers of reasons for liv-
ing than those who attempt suicide because
they have a desire to live. Therefore, rea-
sons for living may also be able to indicate
risk for (or protection against) suicide
attempts within self-injuring adolescents.
To date, the relationship between reasons
for living and NSSI has not been examined,
nor is it known whether reasons for living
could act as an indicator of risk for suicide
attempts within a sample of self-injurers.

Current

understandings

of

NSSI

and suicide suggest that there should be

gradations in the severity of pathology
associated with each behavior. Differences
are expected to exist on factors such as sui-
cide ideation, depression, and reasons for
living; yet, there are few studies evaluating
this assumption. In addition to variations
in severity, there may also be quantitative
differences within specific elements of
these

psychosocial

correlates

between

NSSI and suicide, but we are unaware of
any research examining this idea. The clo-
sest approximation is research exploring
specific traits associated with NSSI such
as

increased

aggression=hostility

and

impulsivity (Brent, Johnson, Bartle et al.,
1993a; Kumar, Pope, & Steer, 2004; Ross
& Heath, 2003). Results from studies in
this area suggest that adopting a symp-
tom-based

approach

to

understanding

NSSI may be warranted over approaches
that attempt to understand the behavior
within the context of specific disorders. A
study by Muehlenkamp, Jacobson, and
Miller (2005) further supports efforts to
understand NSSI from a symptom-based
perspective, finding that specific symptoms
of impulsivity, anger, and chronic emptiness
predicted NSSI in an outpatient sample of
adolescents better than axis I diagnoses
alone. Examining differences in the types
of depressive symptoms and reasons for
living endorsed by adolescents engaged in
NSSI who have and have not attempted
suicide would likely enhance our under-
standing of risk for these behaviors.

In

summary,

researchers

have

attempted

to

establish

a

theoretical

differentiation between NSSI and suicide
attempts on a number of psychosocial cor-
relates to better understand these behaviors.
Many of these studies have been limited in
their ability to identify factors that would
indicate heightened risk for a suicide
attempt within groups who engage in NSSI,
which is important due to the documented
risk for suicide associated with NSSI beha-
viors. The current study expands upon
research conducted by Guertin et al. (2001)

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

71

and Stanley et al. (2001) with the goal of
identifying group differences that can
inform assessments of risk for suicide
within a community sample of adolescents
who engage in NSSI.

We hypothesized that adolescents who

have a history of NSSI would report lower
levels of depressive symptoms and suicidal
ideation than individuals with a history
of both NSSI and a suicide attempt. In
addition, it was hypothesized that adoles-
cents with both NSSI and a suicide attempt
would report fewer reasons for living than
adolescents who engaged in only NSSI. We
also explored potential differences in spe-
cific aspects of depressive symptoms and
reasons for living between the adolescent
groups to expand our understanding of risk
associated with the behaviors. Due to the
exploratory nature of the secondary analy-
ses, no hypotheses were specified.

METHODS

Participants and Procedures

Participants were recruited as part of an
ongoing suicide screening project at an
urban high school in the Midwest. The cur-
rent data were collected over the course of
three academic years from 2001–2004. Data
were collected from 540 adolescents of
whom 62.3% were female. The mean age
of the participants was 15.53 (SD ¼ 1.42).
Among the participants, 48.3% identified
themselves as Caucasian, 28.4% as African
American, 8.1% as Hispanic, 11.8% as
Multi-ethnic, and 2.0% as Asian. Seven
participants (1.4%) did not report their
ethnicity.

Participant

recruitment

occurred

through announcements and parent letters
handed out in English, Math, Social
Studies, Psychology, and Drama classes.
Because we do not know the total enroll-
ment of classes in which recruitment
occurred, and due to variations in the way

consent data were tracked across the three
years of the study, participation rates are
difficult to characterize. However, infor-
mation obtained from the school adminis-
tration suggests that the selected classes
contained the majority of freshman and
seniors enrolled in any given year, with
lower representation of sophomores and
juniors. We estimate that the parents of
1200 students were contacted regarding
the study and that we averaged a 60%
return rate, with the majority giving con-
sent. Across the three years, on average,
86.9% of students with parental consent
were able to be invited to participate in
the study. The remaining 13.1% were
either absent on the day of data collection,
were unable to participate due to classroom
activities, or had moved. While these
participation estimates are not as high as
one would like, and may not have resulted
in a purely representative sample, they
seem to be comparable to other studies
relying on school-based recruitment and
data collection. Students from whom we
obtained parental consent to participate,
and who gave their assent (95.8%), were
excused from their classrooms to complete
a packet of questionnaires within small
groups in a semi-private room of the
school library. All questionnaires were
counterbalanced with the exception of the
demographic form, which always occurred
first. It took approximately 50 minutes to
complete the packet. Upon completion,
students’ responses were examined for
indicators of depression or suicide risk.
Participants endorsing critical items were
spoken with privately by the examiner
about their safety and were given an appro-
priate referral to the school psychologist
who followed up the student’s parents if
the student was currently suicidal.

Measures

Reynolds

Adolescent

Depression

Scale

(RADS;

Reynolds, 1987).

This 30-item self-report

Risk For Suicide

72

VOLUME 11 NUMBER 1 2007

instrument is designed to measure depress-
ive symptoms in adolescents ages 13 to 18.
During the course of the current study, the
second edition of the RADS (RADS-2;
Reynolds, 2002) was released. The content
of items and scoring did not change, but
updated norms and interpretation infor-
mation became available which were used
for all analyses reported here. The RADS
breaks down depressive symptoms into
four subscales: dysphoric mood, anhedo-
nia, negative self-evaluation, and somatic
complaints. Each item is scored on a 4-
point Likert scale ranging from 1 (almost
never) to 4 (most of the time). Scores are
obtained by summing responses and range
from 30–120 points, with higher scores
representing a greater level of depressive
symptoms. The RADS has demonstrated
strong reliability and validity (Reynolds,
1987; 2002). The internal consistency of
the RADS in the current study was a ¼ .94.

Suicidal Ideation Questionnaire (SIQ; Reynolds,
1988).

The SIQ is a 30-item self-report

measure of current suicidal ideation for
adolescents in grades 10 through 12. Parti-
cipants respond to items on a 7-point scale
indicating the frequency of their thoughts
during

the

last

month.

Frequency

responses range from 6 (almost every
day) to 0 (I never had the thought), creating
a range of scores from 0 to 180, with
higher scores representing higher levels of
suicidal ideation. The SIQ has demon-
strated good internal consistency, and has
adequate concurrent and construct validity
(Reynolds, 1988). The SIQ-JR is a 15-item
version of the SIQ that was administered to
freshman participants since it is appropriate
for use with adolescents in grades 7
through 9. The SIQ-JR has also demon-
strated

good

psychometric

properties

(Range & Knott, 1997; Reynolds, 1988).
In the current sample, an internal consist-
ency of a ¼ .94 was obtained. Based on
tables provided in the manual, total scores
on the SIQ and SIQ-JR were transformed

into percentile scores so that data from
the two versions could be combined during
analyses.

Reasons for Living Inventory for Adolescents (RFL-A;
Osman, Downs, Kopper et al., 1998).

The RFL-A

consists of 32 items and five subscales
measuring reasons adolescents give for not
committing suicide. Items are rated on a
6-point scale ranging from 1 (not at all impor-
tant) to 6 (extremely important). Scores are
obtained by averaging responses on each sub-
scale, with low scores being indicative of
fewer reasons for living, and higher suicide
risk. Total scores are calculated by averaging
the subscale scores. Good internal consist-
ency of the RFL-A has been demonstrated
(Osman, Downs, Kopper et al., 1998), and
an alpha of .95 was obtained in the current
sample. Adequate concurrent validity has
been shown through significant negative cor-
relations between RFL-A scores and scores
on measures of suicidal behaviors, hopeless-
ness, and depression (Gutierrez, Osman,
Kopper et al., 2000; Osman, Downs,
Kopper et al., 1998). The RFL-A has also
demonstrated discriminant validity between
psychiatric suicidal adolescents and norma-
tive high school students (Gutierrez, Osman,
Kopper et al., 2000).

Self-Harm Behavior Questionnaire (SHBQ; Gutierrez,
Osman, Barrios et al. 2001).

The SHBQ is a

self-report measure consisting of forced-
choice and free-response items assessing
the degree to which participants have
engaged in self-harmful activities. The
SHBQ consists of four distinct sections
that inquire about the incidence and fre-
quency of non-suicidal self-harm, suicide
attempts,

suicide

threats,

and

suicide

ideation. Each section contains follow-up
questions about features of the target
behavior, such as the methods used and
frequency of the behavior. The SHBQ
defines NSSI behaviors as acts of inten-
tional self-harm without intent to die and

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

73

defines suicidal behavior as an act of self-
harm with any intent to die. The SHBQ
has shown good internal consistency with
alpha estimates from .89 to .96 among
the four sections (Gutierrez, Osman,
Barrios et al., 2001). A total score alpha
of .93 was obtained in the current sample.
Tests

of

convergent validity

revealed

SHBQ total scores were correlated with
existing measures of suicidal behavior
(r ¼ .34 to .70; Gutierrez, Osman, Barrios
et al., 2001).

Based on their responses to the SHBQ,

participants were divided into four cate-
gories of no self-harm (NoSH), non-
suicidal self-injurious behavior only (NSSI),
suicide attempts only (SA), and both self-
injury and suicide attempts (NSSI þ SA).
Responses describing the method of self-
harm (e.g., burn, cut) were coded for
descriptive data. It should be noted that
the section of the SHBQ assessing suicidal
ideation taps lifetime thoughts of suicide,
as distinct from current thoughts assessed
by the SIQ. A participant could indicate
that all of their thoughts about suicide
occurred within the same time frame as
assessed by the SIQ, in which case their
scores on that subscale would overlap with
their SIQ total score. Because the SHBQ
was used to classify students based on the
content of their responses, rather than
from the calculated subscale scores, we
do not believe this potential overlap poses
a problem for analyses.

RESULTS

Incidence and Description of Self-Harm

Non-suicidal self-injury was reported by
23.2% (n ¼ 125) of the sample, and
8.9% (n ¼ 48) reported making a suicide
attempt. After dividing participants into
the four self-harm categories, 75.2%
(n ¼ 406) fell into the NoSH group,

16.1% (n ¼ 87) fell into the NSSI only
group, 1.9% (n ¼ 10) comprised the SA
group, and 7.0% (n ¼ 38) comprised the
NSSI þ SA group. There were no signifi-
cant sex differences across three of the
self-harm groups, but a significant sex dif-
ference was found within the NSSI þ SA
group, v

2

(1) ¼ 4.63, p < .05. Females were

more likely to report both engaging in
NSSI and having attempted suicide than
were males. A significant difference was
also found for ethnicity across the self-
harm groups, v

2

(3) ¼ 9.53, p < .05, with

Caucasians reporting significantly more
self-harm than non-Caucasians, v

2

(1) ¼

8.83, p < .01. There were no significant dif-
ferences in self-harm status within the eth-
nic minority groups, v

2

(3) ¼ 8.83, p < .01.

The ethnic composition of each self-harm
group is reported in Table 1.

Many participants reported that the

NSSI began at age 13 (15.1%) or 14
(28.4%). Among those reporting NSSI
behavior, 32 (25.6%) reported one inci-
dent, 43 (34.4%) reported between two
and three incidents, and 27 (21.6%)
reported four or more incidents (23 did
not report the frequency of their NSSI).
Seventy four (59.2%) participants reported
at least one act of NSSI in the past year. Of
participants reporting a suicide attempt, 21
(43.8%) reported one attempt, another 15
(31.2%) reported two or three attempts,
and 6 (12.5%) reported attempting suicide
four or more times (6 participants did not
report a frequency). Sixteen participants
(33.3%) reported making a suicide attempt
within the past year. There was not a
significant difference in the frequency
of NSSI acts between the NSSI and
NSSI þ SA groups, t(121) ¼ .594, p > .05,
suggesting that the groups were compara-
ble in the number of times they had self-
injured without suicidal intent. Descriptive
data regarding the specific methods of
self-harm and number of methods used
are reported in Table 1.

Risk For Suicide

74

VOLUME 11 NUMBER 1 2007

Group Differences

A series of ANCOVAs with sex and time
since last act of self-harm as covariates
were used to assess differences among the
NSSI, SA, NSSI þ SA, and NoSH groups
on levels of depression, suicidal ideation,
and reasons for living. The means, standard
deviations, and reliability of each measure
are presented in Table 2. Given the number
of planned contrasts conducted, the risk of
type I error was reduced by adopting a con-
servative p-value based on a Bonferroni
correction (.05=6; p < .001) to determine
statistical significance. Significant between
groups differences were found on the
RADS,

F

(3,532) ¼ 37.44,

p <

.000,

g

2

¼ .176;

SIQ,

F

(3,531) ¼ 57.28,

p <

.000, g

2

¼ .244; and the RFL-A, F

(3,530) ¼ 35.28, p < .000, g

2

¼ .166. The

covariate of time since last act of self-harm
did not have a significant effect on any of
the dependent variables (p > .20 for all),
so it was dropped in subsequent analyses.

Planned follow-up pairwise compari-

sons revealed significant differences between
the NoSH group and the NSSI and
NSSI þ SA groups on all dependent vari-
ables (p ¼ .000 for each). The only signifi-
cant difference between the NoSH and SA
group was on the SIQ (p ¼ .000). Parti-
cipants reporting some type of self-harm
had greater levels of depression and suicidal
ideation, as well as fewer reasons for living
than participants who did not report a
history of self-harm. When comparing the
NSSI, SA, and NSSI þ SA groups, signifi-
cant differences emerged between the NSSI

TABLE 1.

Descriptive Data for the Self-Harm Categories

Feature

NoSH (n ¼ 406)

NSSI (n ¼ 87)

SA (n ¼ 10)

NSSI þ SA (n ¼ 38)

Gender

Female

245

52

8

30

Ethnicity

Caucasian

182

50

5

25

African American

133

13

2

6

Hispanic

30

9

2

3

Multi-Ethnic

48

12

1

3

Other Race=Ethnicity

10

2

0

1

Method of Self-Harm

Cut

-

48

3

14

Scratch

-

36

0

0

Burn

-

5

0

0

Self-Hit

-

15

0

0

Punch=Kick

-

9

0

0

Banging

-

3

0

0

Other Method

-

16

8

4

Use of 1 Method

-

61

9

16

Use of 2 or More Methods

-

17

1

20

The values represent n-size for each feature. There are different amounts of missing data for each of the cate-
gories so the numbers reported will not add up to the expected totals. NoSH ¼ No Self-Harm group,
NSSI ¼ Non-suicidal self-injury only group, SA ¼ Suicide Attempt only group, NSSI þ SA ¼ both Non-suici-
dal self-injury and Suicide Attempt group.

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

75

and NSSI þ SA groups on the SIQ,
F

(2,129) ¼ 10.96,

p <

.001,

g

2

¼ .145,

and RFL-A, F(2,129) ¼ 7.18, p < .001,
g

2

¼ .101, with the NSSI þ SA group

reporting greater suicidal ideation and fewer
reasons for living than the NSSI group.
Differences between the NSSI and NSSI þ
SA groups on the RADS suggest that the
NSSI þ SA group may have slightly higher
levels of depressive symptoms, F(2,129) ¼
3.46, p ¼ .03, g

2

¼ .045, but this needs to

be interpreted with caution given the signifi-
cance value and small effect size. The SA
group did not significantly differ from either
the NSSI or NSSI þ SA group on any of the
dependent variables. This lack of significance
is likely the result of the small n-size leading
to underpowered analyses (power ¼ .27).
Consequently, the SA group was excluded
from further analyses and interpretations
regarding this group will not be made.

Follow-up a priori planned contrasts

were used to explore differences in the

specific symptoms of depression and com-
ponents of reasons for living between the
NSSI and NSSI þ SA groups. We included
depressive

symptoms

in

the

analyses

because the omnibus test demonstrated a
possible difference between the groups,
suggesting there may be a unique disparity
on one or more symptom areas of
depression measured by the RADS. Results
are presented in Table 3. There were
significant differences between groups on
the RADS subscales of anhedonia and
negative self-evaluation (score excluded
the suicide item to prevent an inflated
relationship), with the NSSI þ SA group
reporting higher levels of each symptom
cluster. On the RFL-A, significant differ-
ences were found on the future orientation,
suicide related concern, family alliance, and
self-acceptance subscales (see Table 3),
with the NSSI group reporting a greater
number of reasons for living associated
with each category.

TABLE 2.

Means, Standard Deviations, and Reliability of Variables Across Self-Harm Groups

Variable

NoSH (n ¼ 406)

NSSI (n ¼ 87)

SA (n ¼ 10)

NSSI þ SA (n ¼ 38)

a

RADS

.94

Mean

56.28

70.75

63.40

77.24

(SD)

(14.68)

(16.40)

(18.17)

(16.41)

Range

0–104

30–99

41–91

45–111

% Over cut-off

12.32

43.68

30.00

52.63

SIQ

a

.94

Mean

45.77

70.61

72.90

86.27

(SD)

(24.48)

(18.30)

(23.30)

(11.19)

Range

9–98

21–99

23–94

53–99

% Over cut-off

6.12

27.59

40.00

60.53

RFL-A

.95

Mean

5.04

4.47

4.49

3.80

(SD)

(.809)

(.853)

(.802)

(1.03)

Range

.14–6.0

1.89–5.82

3.60–6.0

1.09–6.0

a

SIQ scores are percentile scores. NoSH ¼ No Self-Harm group, NSSI ¼ Non-Suicidal Self-Injury Only group,
SA ¼ Suicide Attempt only group, NSSI þ SA ¼ both Non-Suicidal Self-Injurious and Suicide Attempt group.
RADS ¼ Reynolds Adolescent Depression Scale, SIQ ¼ Suicidal Ideation Questionnaire, RFL-A ¼ Reasons
for Living Inventory for Adolescents. ‘‘% Over cut-off’’ represents participants with scores at or above the
clinically significant cut-off scores (84th percentile on SIQ; Score of 76 or greater on RADS).

Risk For Suicide

76

VOLUME 11 NUMBER 1 2007

DISCUSSION

The present study identified significant dif-
ferences in the level of suicidal ideation,
reasons for living, and depressive symp-
toms between adolescents with a history
of NSSI who had and had not attempted
suicide. Specifically, self-injuring adoles-
cents who attempted suicide reported
greater levels of suicidal ideation, signifi-
cantly fewer reasons for living, and may
be more likely to experience higher levels
of depressive symptoms. Both groups
reported significantly greater depressive
symptoms, suicidal ideation, and fewer rea-
sons for living than the group without any
self-harm. These findings offer empirical
data to suggest that NSSI may be concep-
tualized along a continuum of self-harm,
placing NSSI below suicide attempts in
terms of severity of the behavior and corre-
sponding psychological correlates. Addition-
ally, the results from this study appear
consistent with the theoretical understanding
of NSSI as a dysfunctional coping strategy.
Individuals engaged in NSSI were able to
identify a number of reasons to keep living,
suggesting that they are motivated to live,

which is antithetical to the motivations
underlying suicidal behavior (e.g., desire not
to live). This interpretation is congruent with
Muehlenkamp and Gutierrez’s (2004) find-
ing that self-injuring adolescents have a
greater attraction to life than those who
attempt suicide, indicating adolescents’ per-
spectives on life are important variables to
consider during treatment, and when asses-
sing risk for suicide. However, our inability
to make comparisons with the group of ado-
lescent who had only made suicide attempts
(due to the small n) limits the scope of this
interpretation.

Our finding that adolescents engaged

in NSSI who attempt suicide have higher
levels of suicidal ideation than those who
do not attempt suicide appears intuitive
and replicates numerous studies document-
ing that thoughts about suicide are associa-
ted with greater risk (e.g., Beautrais, Joyce,
& Mulder, 1996; Brent, Perper, Moritz et al.,
1993; De Man & Leduc, 1995; King,
Hovey, Brand et al., 1997; Kovacs, Gold-
ston, & Gatsonis, 1993). However, suicidal
ideation is often overlooked as a risk factor
within studies of NSSI because NSSI is
defined as a behavior void of any thoughts

TABLE 3.

Specific Content Differences between NSSI and NSSI þ SA Groups

Dependent Variable

NSSI Mean (SD)

NSSI þ SA Mean SD

df

t-value

p-level

RADS

Dysphoric Mood

2.60 (.644)

2.71 (.745)

535

.884

.377

Anhedonia

1.97 (.478)

2.19 (.561)

535

2.43

.015

Negative Self-Evaluation

2.23 (.736)

2.54 (.733)

535

2.34

.020

Somatic Complaints

2.65 (.644)

2.86 (.680)

535

1.65

.100

RFL-A

Future Orientation

4.94 (.937)

4.37 (1.37)

534

3.17

.002

Suicide Related Concerns

3.96 (1.50)

2.86 (1.60)

534

4.06

.000

Family Alliance

4.30 (1.20)

3.66 (1.38)

534

2.97

.003

Peer Acceptance & Support

4.65 (1.21)

4.25 (1.33)

533

1.92

.055

Self-Acceptance

4.49 (1.15)

3.88 (1.22)

534

2.93

.004

Significant values are in bold. RADS ¼ Reynolds Adolescent Depression Scale; RFL-A ¼ Reasons for Living
scale, Adolescent version; NSSI ¼ non-suicidal self-injury only group, NSSI þ SA ¼ group with both non-sui-
cidal self-injury and a suicide attempt.

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

77

of death. The results we found provide
empirical evidence supporting the impor-
tance of monitoring suicidal ideation when
working with adolescents who are engaging
in NSSI in order to identify heightened
risk for suicide, and potentially to prevent
suicide attempts within this at-risk group.

Contrary to our hypotheses, level of

depressive symptoms did not significantly
differ between the NSSI and NSSI þ SA
groups (p ¼ .03); however, results suggest
a potential for the NSSI þ SA group to
have more severe levels of depressive
symptoms. Future studies in this area will
have to clarify our findings. We did find
that adolescents engaged in NSSI reported
significantly greater depressive symptoms
than those who did not have a history of
NSSI, which is consistent with previous
studies using community samples (e.g.,
Ross & Heath, 2002). What these results
appear to demonstrate is that adolescents
who engage in NSSI are experiencing a
significant level of distress relative to ado-
lescents without a history of self-harm.
Therefore, depression can be viewed as a
correlate of NSSI. Severity of depressive
symptoms may be an element that increases
risk for making a suicide attempt within sub-
groups of NSSI adolescents, but the findings
are inconclusive at this point. Other
researchers have noted that depressive
symptoms alone may not be strong predic-
tors of suicide (Mann, Waternaux, Haas
et al., 1999). Instead, it may be that other fac-
tors interact with a high level of depression
to increase risk for attempting suicide, or as
Joiner (2005) suggests, that high levels of
emotional distress coupled with repeated
acts of NSSI combine to propel an individual
to make a suicide attempt. The data is incom-
plete at this time and future research should
continue to evaluate these ideas.

In addition to documenting a signifi-

cant difference between NSSI adolescents
who had and had not attempted suicide
on our variables of interest; we examined
specific content differences within the

reasons for living reported and depressive
symptoms.

Some

consistent

findings

emerged across the two domains, indicating
robust results. The self-injuring adolescents
who attempted suicide reported signifi-
cantly higher levels of anhedonia on the
RADS and less of a future orientation on
the RFL-A than the self-injuring adoles-
cents

who

did

not

attempt

suicide.

Together, these complementary findings
suggest that feelings of apathy, lack of
motivation, and an inability to identify
something to work towards in the future
warrant increased attention by clinicians
to evaluate possible suicide risk. Self-
injuring adolescents who have a more nega-
tive outlook on life or who struggle to
identify future positive goals appear to be
at greater risk for attempting suicide. The
results are comparable to findings that feel-
ings of hopelessness significantly elevates
risk for suicide in adolescents (e.g., Maris,
1991; Maris, Berman, Maltsberger et al.,
1992; Mazza & Reynolds, 1998; Rich, Kirk-
patrick-Smith, Bonner et al., 1992), some-
times being a stronger predictor of suicide
attempts than global scores of depression.
Our findings indicate that a similar process
of risk may occur for self-injuring adoles-
cents. Lack of a future orientation and high
levels of anhedonia may contribute to a
pessimistic outlook on life, reducing motiv-
ation to struggle through difficult times,
and increasing the desire to attempt suicide.
Therefore, in order to prevent suicide
attempts within this at-risk group it
becomes important clinically to help self-
injuring adolescents develop long-term
goals, identify positive expectations about
the future, and broaden their perspective
beyond their immediate distress (see Berk,
Henriques, Warman et al., 2004; Brown,
Have, Henriques et al., 2005).

We also found that self-injuring adoles-

cents who did not attempt suicide reported
greater self-acceptance on the RFL-A, and less
negative self-evaluation on the RADS; how-
ever, their self-acceptance=self-evaluation

Risk For Suicide

78

VOLUME 11 NUMBER 1 2007

was still significantly lower than adolescents
who never engaged in NSSI. The fact that
the NSSI þ SA group reported more negative
self-evaluation=self-acceptance is consistent
with research documenting an association
between low self-esteem and suicide risk
(e.g., Overholser, Adams, Lehnert et al.,
1995) in non-self-injuring adolescents. What
our results suggest is that the NSSI only
adolescents were better able to identify some
positive attributes that likely contribute to a
more optimistic outlook and stronger sense
of self, reducing a wish to die. Therefore,
strengthening an adolescent’s self-acceptance
and challenging distorted self-evaluative judg-
ments may enhance feelings of self-worth as
well as increase self-efficacy to cope with
negative life events, adding to resilience
against suicide. Additional studies are needed
to further understand how an adolescent’s
self-perception may impact risk for NSSI
along with risk for suicidal behavior within
this at-risk population.

Two other differences were found

between our groups. The NSSI only group
reported higher levels of family alliance and
suicide

related

concerns

than

the

NSSI þ SA group. The result that adoles-
cents who engaged in only NSSI reported
higher family alliance than those who
attempted suicide suggests that NSSI only
adolescents were still able to identify and
draw upon external support from others
in their life to avoid attempting suicide.
This finding implies that fostering close,
supportive

relationships

is

potentially

important to reducing suicide risk. This
idea is further supported by the lack of sig-
nificant differences between our groups on
peer acceptance and support. Acceptance
by one’s peers and support from friends
is often identified as a protective factor
against suicide (Kandel, Raveis, & Davies,
1991; King, Segal, Naylor et al., 1993;
Rudd, 1990). However, within our sample,
peer support did not differ between the
groups. Therefore, having support and
feeling as though one can turn to a family

member for help may be a more important
protective factor against suicide attempts
within a self-injuring population of adoles-
cents than turning to one’s peers. Focusing
upon family interventions that enhance
supportive, positive connections between
adolescents and a family member may be
particularly warranted.

With regard to suicide related con-

cerns, the NSSI only group reported a
greater number of concerns indicating
greater aversion to suicide. Items from this
subscale

represent

fears

of

suicidal

thoughts, fears of pain associated with a
suicide attempt, and fears of suicidal beha-
vior. Within our sample, adolescents who
engaged in NSSI but did not make a suicide
attempt scored as being more fearful of sui-
cide. This may have implications for under-
standing suicide risk within a NSSI group
using Joiner’s (2005) model. Joiner pro-
poses that individuals who attempt suicide
are able to do so, in part, because they have
habituated to pain or fears associated with
self-destructive acts through exposure to
acts of self-harm. Our finding is consistent
with this perspective because those who
attempted suicide reported significantly
fewer fears of suicide. However, those
who attempted suicide did not differ in
the mean frequency of their NSSI acts
from those who did not attempt suicide,
which

conflicts

with

Joiner’s

thesis.

Additional research is needed to evaluate
the application of Joiner’s model within a
self-injuring sample. What can be taken
from our findings is that adolescents who
demonstrate behaviorally, or verbally, little
fear about severely injuring themselves are
probably at a high level of suicide risk.

A few limitations must be considered

when interpreting our findings. All of the
data were collected through self-report
measures and are retrospective in nature;
although a majority of participants had
self-injured within the past year. The fact
that some participants who reported self-
harm were not currently engaging in NSSI

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

79

means that some scores on the dependent
measures did not reflect the adolescents’
psychological state in close proximity to
the self-harm. However, this allows for a
conservative

estimate

of

differences,

assuming that psychological distress would
be lower during time periods in which the
adolescent is not actively engaging in self-
injury. Another limitation is that data were
drawn from a non-clinical sample of high
school students who endorsed symptoms
of depression and suicidal ideation well
within the normative range. On the other
hand, much of the research on NSSI has
focused upon clinical samples of adoles-
cents, so our sample provides valuable data
about the nature of self-injury in the
broader community of adolescents. The
self-report nature of the measures reduced
the follow-up data that could be collected
regarding self-injury, which prevented clari-
fication of ambiguous responses to the
Self-Harm Behavior Questionnaire (Gutier-
rez, Osman, Barries et al., 2001). Data
regarding the function of SIB, other affect-
ive experiences (e.g., hostility), and family
functioning were not assessed but may also
shed light on potential differences between
self-injurers who do and do not attempt
suicide. Additionally, we did not collect
information on whether or not the parti-
cipants were currently, or had previously,
received therapeutic services, which could
affect scores on the measures we used.
Since our sample drew from a non-clinical
population the generalizability of our find-
ings to inpatient populations or to adoles-
cents with clinically significant psychiatric
disorders may be limited.

In summary, our results provide data

documenting clinically relevant differences
within traditional psychosocial correlates
of suicide risk between self-injuring adoles-
cents who attempt suicide and those who
do not. Additionally, our findings that
self-injuring

adolescents

who

endorse

symptoms of anhedonia, pessimistic future
perspectives, low self-acceptance, poor

family connections, and few fears about
self-harm appear to be at the greatest risk
for suicide add to the field’s understanding
of this at-risk population. Our results also
contribute evidence for arguments that a
symptom based approach should be used
when trying to comprehensively under-
stand NSSI and the associated risk for sui-
cide within this subgroup of adolescents.
Global diagnoses such as depression may
not capture the unique elements contribu-
ting to the destructive behavior(s). Overall,
the clinical implications of our findings are
that careful monitoring of suicidal ideation
and interventions targeting specific symp-
tom clusters underlying the NSSI may help
to prevent a potential suicide attempt or
death. Strengthening reasons for living,
improving family connections, and building
a positive future orientation would most
likely build strong internal resilience against
suicide for these vulnerable adolescents.

AUTHOR NOTE

Jennifer J. Muehlenkamp, University of
North Dakota

Peter M. Gutierrez, Northern Illinois

University

Correspondence regarding this article

should be addressed to Jennifer J. Muehlen-
kamp, Ph.D., Department of Psychology,
319 Harvard St Stop 8380, University of North
Dakota, Grand Forks, ND 58202. E-mail:
[e-mail]

REFERENCES

Beautrais, A. L., Joyce, P. R., & Mulder, R. T. (1996).

Risk factors for serious suicide attempts among
youths aged 13 through 24 years. Journal of the
American Academy of Child and Adolescent Psychiatry

,

35

, 1174–1182.

Berk, M. S., Henriques, G. R., Warman, D. M., et al.

(2004). A cognitive therapy intervention for sui-
cide attempters: An overview of the treatment

Risk For Suicide

80

VOLUME 11 NUMBER 1 2007

and case examples. Cognitive and Behavioral Practice,
11

, 265–277.

Brent, D. A., Johnson, B., Bartle, S., et al. (1993a).

Personality disorder, tendency to impulsive viol-
ence, and suicidal behavior in adolescents. Journal
of the American Academy of Child and Adolescent Psy-
chiatry

, 32, 69–75.

Brent, D. A., Perper, J., Moritz, G., et al. (1993b).

Suicide in adolescents with no apparent psycho-
pathology. Journal of the American Academy of Child
and Adolescent Psychiatry

, 32, 494–500.

Brown, G. K., Have, T. T., Henriques, G. R., et al.

(2005). Cognitive therapy for the prevention of
suicide attempts: A randomized controlled trial.
JAMA: Journal of the American Medical Association

,

294

, 563–570.

Darche, M. A. (1990). Psychological factors differen-

tiating self-mutilating and non-self-mutilating ado-
lescent inpatient females. The Psychiatric Hospital,
21

, 31–35.

De Man, A. F. & Leduc, C. P. (1995). Suicidal idea-

tion in high school students: Depression, and
other correlates. Journal of Clinical Psychology, 51,
173–180.

DiClemente, R. J., Ponton, L. E., & Hartley, D.

(1991). Prevalence and correlates of cutting beha-
vior: Risk for HIV transmission. Journal of the
American Academy of Child & Adolescent Psychiatry

,

30

, 735–739.

Dulit, R. A., Fryer, M. R., Leon, A. C., et al. (1994).

Clinical correlates of self-mutilation in borderline
personality disorder. American Journal of Psychiatry,
151

, 1305–1311.

Garrison, C. Z., Addy, C. L., McKeown, R. E., et al.

(1993). Nonsuicidal physically self-damaging acts
in adolescents. Journal of Child & Family Studies,
2

, 339–352.

Guertin, T., Lloyd-Richardson, E., Spirito, A., et al.

(2001). Self-mutilative behavior in adolescents
who attempt suicide by overdose. Journal of the
American Academy of Child and Adolescent Psychiatry

,

40

, 1062–1069.

Gutierrez, P. M., Osman, A., Barrios, F. X., et al.

(2001). Development and initial validation of the
self-harm

behavior

questionnaire.

Journal

of

Personality Assessment

, 77, 475–490.

Gutierrez, P. M. Osman, A., Kopper, B. A., et al.

(2000). Why young people do not kill themselves:
The reasons for living inventory for adolescents.
Journal of Clinical Child Psychology

, 29, 177–187.

Hawton, K., Fagg, J., Simkin, S., et al. (1997). Trends

in deliberate self-harm in Oxford, 1985–1995:

Implications for clinical services and the preven-
tion of suicide. British Journal of Psychiatry, 171,
556–560.

Joiner, T. E. (2005). Why people die by suicide. Cam-

bridge, MA: Harvard University Press.

Kandel, D. B., Raveis, V. H., & Davies, M. (1991).

Suicidal ideation in adolescence: Depression,
substance use, and other risk factors. Journal of
Youth and Adolescence

, 20, 289–309.

Kessler, R. C., Borges, G., & Walters, E. E. (1999).

Prevalence of and risk factors for lifetime suicide
attempts in the National Comorbidity Survey.
Archives of General Psychiatry

, 56, 617–626.

King, C. A., Hovey, J. D., Brand, E., et al. (1997).

Prediction of positive outcomes for adolescent
psychiatric inpatients. Journal of the American
Academy of Child and Adolescent Psychiatry

, 36,

1434–1442.

King, C. A., Segal, H., Naylor, M., et al. (1993). Fam-

ily functioning and suicidal behavior in adolescent
inpatients with mood disorders. Journal of the
American Academy of Child and Adolescent Psychiatry

,

32

, 1198–1206.

Kovacs, M., Goldston, D., & Gatsonis, C. (1993).

Suicidal behaviors and childhood-onset depressive
disorders: A longitudinal investigation. Journal of
the American Academy of Child and Adolescent Psy-
chiatry

, 32, 8–20.

Kumar, G., Pope, D., & Steer, R. A. (2004). Ado-

lescent psychiatric inpatients’ self-reported reasons
for cutting themselves. The Journal of Nervous and
Mental Disease

, 192, 830–836.

Mann, J. J., Waternaux, C., Haas, G. L., et al. (1999).

Toward a clinical model of suicidal behavior in
psychiatric patients. The American Journal of Psy-
chiatry

, 156, 181–189.

Maris, R. W. (1991). Special issue: Assessment and

prediction of suicide: Introduction. Suicide and
Life-Threatening Behavior

, 21, 1–17.

Maris, R. W., Berman, A. L., Maltsberger, J. T. H,

et al. (Eds.). (1992). Assessment and prediction of sui-
cide

. New York: Guilford.

Maris, R. W., Berman, A. L., & Silverman, M. M.

(2000). Comprehensive textbook of suicidology. New
York: Guilford.

Mazza, J. J. & Reynolds, W. M. (1998). A longitudinal

investigation of depression, hopelessness, social
support, and major and minor life events and their
relation to suicidal ideation in adolescents. Suicide
and Life-Threatening Behavior

, 28, 359–374.

Muehlenkamp, J. J. & Gutierrez, P. M. (2004). An

investigation of differences between self-injurious

J. Muehelenkamp and P. Gutierrez

ARCHIVES OF SUICIDE RESEARCH

81

behavior and suicide attempts in a sample of
adolescents. Suicide and Life-Threatening Behavior,
34

, 12–23.

Muehlenkamp, J. J., Jacobson, C. M., et al. (2005).

Borderline symptoms and adolescent parasuicide.
Poster presented at the annual American Associ-
ation of Suicidology conference, Broomfield, CO.

Nock, M. K. & Prinstein, M. J. (2004). A functional

approach to the assessment of self-mutilative
behavior. Journal of Consulting and Clinical Psychology,
72

, 885–890.

O’Carroll, P.W., Berman, A.L., Marris, R.W., et al.

(1996). Beyond the Tower of Babel: A nomencla-
ture for suicidology. Suicide and Life-Threatening
Behavior

, 26, 237–52.

Olfson, M., Gameroff, M. J., Marcus, S. C., et al.

(2005). National trends in hospitalization of youth
with intentional self-inflicted injuries. The American
Journal of Psychiatry

, 162, 1328–1335.

Osman, A., Downs, W. R., Kopper, B. A., et al. (1998).

The reasons for living inventory for adolescents
(RFL-A): Development and psychometric proper-
ties. Journal of Clinical Psychology, 54, 1063–1078.

Overholser, J. C., Adams, D. M., Lehnert, K. L., et al.

(1995). Self-esteem deficits and suicidal tendencies
among adolescents. Journal of the American Academy
of Child and Adolescent Psychiatry

, 34, 919–928.

Range, L. M. & Knott, E. C. (1997). Twenty suicide

assessment instruments: Evaluation and recom-
mendations. Death Studies, 21, 25–58.

Reynolds, W. M. (1987). Reynolds adolescent depression

scale: Professional manual

. Lutz, FL: Psychological

Assessment Resources, Inc.

Reynolds, W. M. (1988). SIQ professional manual. Odessa,

FL: Psychological Assessment Resources, Inc.

Reynolds, W. M. (2002). Reynolds adolescent depression

scale–2nd edition: Professional manual

. Lutz, FL:

Psychological Assessment Resources, Inc.

Rich, A. R., Kirkpatrick-Smith, J., Bonner, R. L., et al.

(1992). Gender differences in the psychosocial
correlates of suicidal ideation among adolescents.
Suicide and Life-Threatening Behavior

, 22, 364–373.

Ross, S. & Heath, N. (2002). A study of the fre-

quency of self-mutilation in a community sample
of adolescents. Journal of Youth and Adolescence, 31,
67–77.

Ross, S. & Heath, N. (2003). Two models of ado-

lescent self-mutilation. Suicide and Life-Threatening
Behavior

, 33, 277–287.

Rudd, M. D. (1990). An integrative model of suicidal

ideation. Suicide and Life-Threatening Behavior, 20,
16–31.

Stanley, B., Gameroff, M. J., Michalsen, V., et al.

(2001). Are suicide attempters who self-mutilate
a unique population? The American Journal of Psy-
chiatry

, 158, 427–432.

Stanley, B., Winchel, R., Molcho, A., et al. (1992).

Suicide and the self-harm continuum: Phenom-
enological and biological evidence. International
Review of Psychiatry

, 4, 149–155.

Walsh, B.W. (2005). Treating self-injury: A practical guide.

New York, NY: Guilford Press.

Zlotnick, C., Donaldson, D., Spirito, A., et al. (1997).

Affect regulation and suicide attempts in ado-
lescent inpatients. Journal of the American Academy
of Child and Adolescent Psychiatry

, 36, 793–798.

Risk For Suicide

82

VOLUME 11 NUMBER 1 2007