You are on page 1of 8

Journal of Autism and Developmental Disorders

https://doi.org/10.1007/s10803-022-05575-2

BRIEF REPORT

Brief Report: Exploratory Evaluation of Clinical Features Associated


with Suicidal Ideation in Youth with Autism Spectrum Disorder
Kimberly S. Ellison1,2 · Elzbieta Jarzabek1 · Scott L. J. Jackson1,3 · Adam Naples1 · James C. McPartland1 

Accepted: 13 April 2022


© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022

Abstract
There has been a heightened awareness of an increased risk of suicidality among individuals with autism spectrum disorder
(ASD) due to high rates of suicidal ideation (SI) in this population (11–66%). The current study investigated the rate of
parent-endorsed SI and associated clinical features in 48 youths with ASD (Age; M: 12.97 years, SD: 2.33). SI was endorsed
in 18.75% of participants. Youth with SI exhibited significantly higher levels of affective problems, externalizing problems,
feelings of humiliation and rejection, and symptoms related to perfectionism. Results indicate that co-occurring mental
health problems are associated with suicidal ideation and provide relevant targets for psychotherapeutic intervention. This
preliminary study in a modest sample suggests the value of further research in larger samples to replicate and generalize
these findings.

Keywords  Autism spectrum disorder · Suicidal ideation · Suicide risk · Risk factors · Youth

Autism spectrum disorder (ASD) is a heterogeneous neu- 100,000 individuals) in the last 17 years (National Institute
rodevelopmental condition that manifests in childhood and of Mental Health (NIMH), 2018). In 2016, suicide was the
is characterized by a dyad of core features: (1) challenges second leading cause of death in youth between the ages of
with social emotional reciprocity and social communica- 10–24 (Curtin et al., 2016; NIMH: Suicide, 2018). Parent
tion and (2) restricted interests, unusual sensory responses, report indicates that frequency of SI or suicide attempts is
and repetitive behaviors (American Psychiatric Association, significantly greater for individuals with psychiatric con-
2013). Co-occurring psychiatric difficulties, such as depres- ditions (Mayes et al., 2015) and that individuals with SI
sion and anxiety (Mannion et al., 2014; Mukaddes & Fateh, exhibit significantly elevated rates of co-occurring psycho-
2010) are common in ASD and are predisposing factors for pathology (e.g., mood, anxiety, and disruptive disorders)
suicide (Gould et al., 1996; Kashani et al., 1989). Despite its compared to nonsuicidal youths (Gould et al., 1996; Kashani
critical relevance to ASD, suicidality is poorly understood et al., 1989; Mayes et al., 2015). Self-report reveals similar
in this population. relationships, with child-reported SI and co-occurring psy-
Suicidal ideation (SI) is recognized as an emergent con- chiatric symptoms associated with depression and anxiety
cern in the general population and in psychiatric conditions (Kashani et al., 1989; Pezzimenti et al., 2019; Wijnhoven
other than autism. The age-adjusted suicide rate between et al., 2019) as well as aggressiveness and social difficulties
1999 and 2016 has increased 28% (from 10.5 to 13.4 per (Gould et al., 1996). Social communication difficulties are
associated with higher risk for suicidal thoughts and suicidal
plans (Culpin et al., 2018). Consequently, youth who have
* James C. McPartland
james.mcpartland@yale.edu social challenges by virtue of having ASD, in addition to
internalizing and/or externalizing problems, may be even at
1
Child Study Center, School of Medicine, Yale University, 40 higher risk for developing SI.
Temple Street, Suite 6A2, New Haven, CT 06520, USA The increased risk in ASD for co-occurring psychi-
2
Department of Psychology, Louisiana State University, 236 atric conditions may indicate increased risk for SI and
Audubon Hall, Baton Rouge, LA 70803, USA related behaviors (Hess et al., 2010). Nevertheless, the
3
Office of Assessment and Analytics, Southern Connecticut body of research studying SI in ASD is limited and has
State University, 501 Crescent Street, New Haven, CT 06515, yielded inconsistent findings and utilized inconsistent
USA

13
Vol.:(0123456789)
Journal of Autism and Developmental Disorders

methodologies (Cha et al., 2018; Gould et al., 1996; Mayes Even fewer studies have used clinician-administered inter-
et al., 2015). Estimates of SI in ASD vary drastically, with views or formal suicide risk assessments to confirm the
rates ranging from 9.6 to 66% (Hedley & Uljarević, 2018; existence of SI (Hedley & Uljarević, 2018).
Hunsche et al., 2020; La Buissonnière Ariza et al., 2021). The purpose of this study was to advance understanding
Contributing to this wide range in estimates may be the of SI in youth with ASD by addressing several of these
large age span across studies, with youth ranging from 6 limitations of previous research. Rather than relying on
to 24 years old (Cassidy et al., 2014; Hedley & Uljarević, caregiver or self-report measures alone, we integrated both
2018; Hunsche et al., 2020; Mukaddes & Fateh, 2010; to determine both parent and child experience of emo-
Storch et al., 2013). tional, behavioral, and anxiety-related problems. Addition-
As in typical development, being male and an older ally, we integrated clinician-administered comprehensive
adolescent constitute risk factors associated with SI in risk assessments with both the child and caregiver to con-
ASD (Brent et  al., 1999; Cassidy et  al., 2014; Mayes firm SI. We ensured that all youth met rigorous research
et al., 2013; McDonnell et al., 2020). Likewise, non-demo- diagnostic criteria for ASD. Aligning with prevailing theo-
graphic risk factors established in typically developing ries of SI in typical development, (the cognitive model and
individuals apply to youth with ASD including: the pres- the integrated motivational-volitional model (O’Connor &
ence of comorbid psychopathology (Cassidy et al., 2014; Kirtley, 2018; Wenzel & Beck, 2008), we anticipated that
Hirvikoski et al., 2020; Mayes et al., 2013; McDonnell both dispositional vulnerability factors (e.g., depression,
et al., 2020; Mukaddes & Fateh, 2010), impulsive/aggres- behavioral rigidity) and cognitive processes (e.g., anxious
sive tendencies (Kanne & Mazurek, 2011), social isolation coping) would contribute to SI in youths with ASD.
and bullying (Hedley et al., 2018; van Roekel et al., 2010;
Zeedyk et al., 2014), and genetic factors (Hirvikoski et al.,
2020). In a recent examination of risk factors for suicide
in ASD, females with ASD without comorbid intellectual Methods
disability were found to be at the highest risk for suicide,
whereas both genders had the highest risk when presenting Participants
with comorbid Attention-Deficit/Hyperactivity Disorder
(Hirvikoski et al., 2020). The relationship of autism symp- The sample for this exploratory study was comprised of
tom severity and SI has not been consistent across studies 48 youths with ASD participating in ongoing research at
(Mayes et al., 2013; McDonnell et al., 2020; Storch et al., the Yale Child Study Center in New Haven, Connecticut.
2013); although children having fewer ASD symptoms has Exclusionary criteria included current use of benzodiaz-
been found to be associated with less SI (Hunsche et al., epine or anticonvulsant medication; a history of seizures
2020). Nevertheless, a diagnosis of ASD increases risk of or head injuries; primary psychiatric diagnosis that was
suicide attempts regardless of the presence of comorbid not ASD; full scale IQ < 70, as measured by the Differen-
psychopathology in adolescents and young adults (Chen tial Ability Scales, 2nd Edition (DAS-II; Elliott, 2007) or
et al., 2017; Moses, 2017). Additionally, relatives of indi- other factors that would preclude successful data collec-
viduals with ASD were found to have higher risk for sui- tion. Participants with ASD were diagnosed using gold
cidal behavior than relatives of controls (Hirvikoski et al., standard research tools, including the Autism Diagnostic
2020); further demonstrating that impact of having ASD Observation Schedule 2nd Edition (ADOS-2; Lord et al.,
on not only the individual themselves but also the extent 2002)) and Autism Diagnostic Interview Revised (ADI-R;
on close family members. Rutter et al., 2003), administered by a research-reliable
Only 1 of 13 studies included in a recent review of administrator; all participants received a DSM-5 diagnosis
suicidality and ASD used “gold standard” diagnostic by a licensed clinician. ADOS-2 Comparison Scores (CS)
tools (i.e., Autism Diagnostic Observation Schedule, Sec- for those individuals who received an ADOS-2, Module 4
ond Edition (ADOS-2;Lord et al., 2002), and the Autism were calculated using the revised ADOS-2 DSM-5-based
Diagnostic Interview, Revised (ADI-R; Rutter et al., 2003)) algorithm (Hus & Lord, 2014). Demographic informa-
(Hedley et al., 2018; Storch et al., 2013). Further, only tion (e.g., age, sex, ethnicity), the SRS-2, CBCL/6–18,
three recent studies published after this review utilized and MASC-P were completed by a primary caregiver. The
the ADOS-2 for characterization (Hunsche et al., 2020; La participant completed the MASC-C questionnaire. Study
Buissonnière Ariza et al., 2021; McDonnell et al., 2020). procedures were reviewed and approved by the university
The vast majority of studies rely on caregiver report, with institutional review board. All parents and youth who took
few integrating caregiver and child report to comprehen- part in the study provided written informed consent or
sively evaluate the presence of SI (Storch et al., 2013). assent (respectively) prior to participation.

13
Journal of Autism and Developmental Disorders

Determination of SI Child Behavior Checklist for Ages 6–18 (CBCL/6–18;


Achenbach & Rescorla, 2001)
Presence of SI was determined by caregiver-report of
whether their child “talks about killing self” over the The CBCL/6–18 is a caregiver report instrument com-
past 6 months, as reported on item 91 of the CBCL/6–18. prised of 113 statements (e.g., “Can’t sit still, restless,
Suicidal ideation was considered present if the caregivers or hyperactive.”/“Sudden changes in mood of feelings.”)
responded to this item with either “Somewhat or Some- that caregivers’ rate on how well the statement describes
times True” or “Very True or Often True.” If endorsed, the behavior of their child over the past 6 months on a
a clinician verified the presence of SI by conducting a 3-point Likert-type scale ranging from 0 (Not True) to
thorough risk assessment with both the caregiver(s) and 2 (Very True or Often True). Items on the CBCL/6–18
the child. The risk assessment procedure utilized was a were developed to provide assessment of specific syn-
combination of (a) the Columbia Suicide Severity Rat- drome scales and DSM-oriented scales. Higher scores on
ing Scale (CSSR-S; Posner et al., 2008) a standardized the CBCL/6–18 scales represent the presence of increased
interview that assessed the individual’s current level symptoms related to that specific emotional or behavioral
and intensity of suicidal ideation, intent to harm oneself problem. Scores above 70 are considered to be clinically
(e.g., an active or passive plan, access to means), and sui- significant. Psychometric properties for the CBCL/6–18
cidal behavior (i.e., previous attempts) and (b) follow-up have been established in previous studies (Achenbach &
questions pertaining to related risk factors (e.g., stress- Rescorla, 2001; Achenbach et al., 2003). The analyses in
ful life events, amount of social support). The CSSR-S this study focus on the DSM-5 scales (i.e., Affective Prob-
was applied as a clinical information-gathering tool to lems, Attention-Deficit/Hyperactivity Problems, Anxiety
assist clinicians in assessing the level of risk the par- Problems, Conduct Problems, Oppositional Defiant Prob-
ticipant posed at that current time; quantitative data from lems, and Somatic Problems), which we felt were most
the CSSR-S were not used in the present analysis. The conceptually relevant as well as relate to the dispositional
risk assessments were conducted by either licensed clini- vulnerabilities associated with the cognitive model of
cal psychologists or by a masters-level clinician under suicide.
the direct supervision of a licensed clinical psychologist
who consulted and reviewed the risk assessments to make
the final determination. In all cases, presence of SI was Multidimensional Anxiety Scale for Children, Child‑Report
confirmed with both the caregiver(s) and the child based and Parent‑Report (MASC‑C and MASC‑P; March et al.,
upon the risk assessment described above. 1997)

Both the parent-report and the child-report versions of


Materials the MASC were utilized to assess symptoms of anxiety.
Both versions of the MASC are comprised of 39 statements
Social Responsiveness Scale, 2nd Edition, School‑Age Form (e.g., “I (My child) feel(s) tense or uptight.”/“The idea
(SRS‑2; Constantino & Gruber, 2012) of going away to camp would scare me (him/her)”) that
the informant will rate on how often the statement is true
This caregiver-report measure is comprised of 65 state- based upon recent feelings on a 4-point Likert-type scale
ments (e.g., “Is able to communicate his or her feeling ranging from 0 (Never True) to 3 (Often True). Items on
to others.”/“Seems overly sensitive to sounds, textures, the MASC were developed to assess overall anxiety as well
or smells.”) that caregivers rate on how well the state- as components of Physical Symptoms (i.e., Tense/Rest-
ment describes the behavior of their child over the past less, Somatic/Autonomic), Social Anxiety (i.e., Humilia-
6 months on a 4-point Likert-type scale ranging from 1 tion/Rejection, Performance Fears), and Harm Avoidance
(Not True) to 4 (Almost Always True). Items on the SRS-2 (i.e., Perfectionism, Anxious Coping). Higher scores for
were developed to assess ASD-related symptoms associ- the MASC subscales and total scores for each domain rep-
ated with social communication and interaction abilities, resent the presence of increased levels of symptoms of
as well as restricted interests and repetitive behaviors. anxiety. It is important to note that MASC-C T-scores and
Total scores on the SRS-2 are translated into T-scores MASC-P raw scores were utilized for analyses since there
(μ = 50, σ = 10), with higher scores representing the pres- are no established norms for the MASC-P. Psychometric
ence of increased levels of ASD-related symptoms. Psy- properties for the parent- and child-report versions of the
chometric properties for the SRS-2 have been established MASC have been established in previous studies (March
in previous studies (Constantino & Gruber, 2012; Reszka et al., 1997; Rynn et al., 2006).
et al., 2014).

13
Journal of Autism and Developmental Disorders

Statistical Analysis standard score of 70, the lowest obtained Full Scale-IQ for
an included participant was a standard score of 84. As shown
All statistical analyses were conducted using R. Given the in Table 1, youth with SI and those without SI did not differ
small sample size, bootstrap analysis was performed to avoid on IQ, age, sex, SRS-2 or ADOS-2 CS.
the assumption of normality in the sample distribution. This
nonparametric statistical approach uses multiple small sam- Clinical Features of Suicidal Ideation in ASD
ples drawn from the original sample to estimate the sam-
pling distribution of the statistic through simulations, based Participants with ASD with reported SI were compared to
on sampling with replacement (Davison & Hinkley, 1997; those with no reported SI across the DSM-oriented scales
Efron & Tibshirani, 1993; Wright et al., 2011). Due the limi- of the CBCL/6–18, the parent- and child-report versions of
tation of the sample size, bootstrapping was utilized instead the MASC (MASC-P and MASC-C, respectively). Looking
of parametric statistical methods to ensure that a) there specifically at the CBCL/6–18 DSM-oriented scales, youth
was no violation of normality and b) the robustness of the with SI obtained significantly higher scores on the Affec-
results (Carleton et al., 2012; Walters & Campbell, 2004). tive Problems, and Oppositional Defiant Problems. On the
Bootstrapping with 9999 sample replications was used to MASC-C youth with SI had higher scores on the Perfec-
calculate nonparametric 95% confidence intervals (CI) for tionism subscale. According to parent report, individuals
the difference in means for each scale (bias-corrected and with SI obtained higher scores on the Humiliation/Rejection
accelerated’ correction (BCa) was used on each confidence subscale (Table 2).
interval). We report BCa means and corresponding stand-
ard errors along with mean of the bootstrapped values. We
adjusted for multiple comparisons using the Benjamíni- Discussion
Hochberg method; the False Discovery Rate utilized was
10% (Thissen et al., 2002). This study examined frequency of endorsement of SI in a
sample of youth with ASD and explored associated clinical
features. Advancing from prior research on this topic, we
Results carefully characterized our sample using “gold-standard”
research tools (i.e., ADI-R and ADOS-2), integrated both
SI was endorsed for 18.75% (n = 9; 7 male; Mean parent and child report, and confirmed SI by direct clini-
age = 12.97, SD = 2.33) of participants; correspondingly, 39 cian assessment. Results align with prior studies indicating
were reported not to have SI (16 male; Mean age = 13.77, that SI is overrepresented in ASD compared to the general
SD = 2.84). Ethnicity/race distributions for the ASD with population (Curtin et al., 2016; Hedley & Uljarević, 2018;
SI and ASD without SI groups are as follows: ASD with Mukaddes & Fateh, 2010; Shtayermman, 2008; Storch et al.,
SI (88.9% White/Caucasian, 11.1% Asian) and ASD with- 2013). In our study, rates of suicidal ideation were signifi-
out SI (79.5% White/Caucasian, 10.3% Hispanic, 7.7% cantly higher in youth with ASD (18.75%) than in popula-
Black, 2.6% Asian); these groups did not differ on ethnic- tion estimates of typically developing youth (4.27%; crude
ity/race (p = 0.29). Additionally, although the Full Scale- rate of suicide injury deaths per 100,000, ages 8–17 from
IQ exclusion criteria for this study extended to as low as a the 2017 national mortality statistics from the Centers for

Table 1  Demographic comparisons
SI (n = 9) No SI (n = 39) Mean group Benjamini–Hoch- Confidence intervals*
difference berg p-value

Gender
 Male (Female) 7(2) 33 (6) – – –
Age range 8–17 years old 8–17 years old
Mean, SD M = 12.97, SD = 2.33 M = 13.77, SD = 2.84 0.80 0.83 –
DAS-2 verbal M = 107.11, SD = 19.88 M = 104.94, SD = 13.04 2.10 0.83 7.99, 19.68
DAS-2 NVR M = 108.67, SD = 27.13 M = 104.49, SD = 14.72 4.22 0.83 11.35, 23.42
DAS-2 spatial M = 102.11, SD = 21.40 M = 100.15, SD = 14.49 1.94 0.83 9.87, 17.73
DAS-2 GCA​ M = 107.56, SD = 26.06 M = 103.79, SD = 13.08 3.76 0.83 9.23, 26.41

SI suicidal ideation, DAS-II differential ability scales (2nd edition), NVR nonverbal reasoning, GCA​ general conceptual ability
*Based on Bootstrapping (BCa)

13
Journal of Autism and Developmental Disorders

Table 2  Clinical features
SI (n = 9) No SI (n = 39) Mean group diff Benjamini– Confidence intervals*
Hochberg
p-value

CBCL oppositional defiant probs M = 67.67, SD = 7.83 M = 56.82, SD = 7.20 10.85 0.03± 5.29, 16.15
CBCL affective probs M = 71.22, SD = 7.93 M = 61.51, SD = 7.79 9.71 0.04± 3.57, 14.67
MASC-P humiliation/rejection M = 11.13, SD = 2.90 M = 7.33, SD = 4.43 3.79 0.04± 1.30, 6.00
MASC-C perfectionism M = 65.75, SD = 6.32 M = 56.26, SD = 13.10 9.49 0.05± 2.72, 14.47
MASC-P tense/restless M = 9.13, SD = 4.52 M = 5.85, SD = 3.27 3.28 0.27 .03, 6.14
MASC-C public performance fears M = 48.25, SD = 10.82 M = 54.55, SD = 11.26 − 6.30 0.43 − 14.86, .75
MASC-P somatic/automatic M = 4.63, SD = 3.20 M = 2.67, SD = 3.51 1.96 0.50 .62, 4.08
CBCL anxiety probs M = 68.22, SD = 7.68 M = 63.64, SD = 8.70 4.58 0.50 1.53, 9.51
MASC-P anxious coping M = 10.13, SD = 1.36 M = 9.46, SD = 3.01 0.66 0.83 .61, 2.02
MASC-C tense/restless M = 57.63, SD = 7.58 M = 55.05, SD = 11.03 2.57 0.83 3.54, 8.62
MASC-P perfectionism M = 8.63, SD = 1.85 M = 9.13, SD = 1.70 − 0.50 0.83 − 1.91, .71
CBCL conduct probs M = 56.78, SD = 6.67 M = 55.59, SD = 6.65 1.19 0.83 2.69, 6.83
MASC-C somatic/automatic M = 48.00, SD = 6.28 M = 49.11, SD = 9.14 − 1.11 0.83 − 3.38, 6.41
MASC-C humiliation/rejection M = 49.63, SD = 13.32 M = 51.61, SD = 10.11 − 1.98 0.83 − 6.40, 12.29
MASC-C anxious coping M = 55.00, SD = 15.26 M = 57.00, SD = 9.91 − 2.00 0.83 − 11.72, 8.67
MASC-P public performance fears M = 7.25, SD = 2.71 M = 6.95, SD = 2.87 0.30 0.83 − 1.56, 2.38
CBCL ADHD probs M = 61.44, SD = 10.13 M = 60.72, SD = 7.61 0.72 0.83 5.74, 7.65
CBCL somatic probs M = 61.22, SD = 6.59 M = 57.26, SD = 8.80 3.97 0.83 1.44, 8.36
SRS-2 awareness M = 68.11, SD = 13.11 M = 67.28, SD = 12.04 0.78 0.83 7.42, 10.10
SRS-2 comm M = 75.67, SD = 14.71 M = 72.00, SD = 11.56 3.68 0.83 5.44, 13.73
SRS-2 cognition M = 69.56, SD = 12.28 M = 68.44, SD = 11.32 1.10 0.83 8.18, 8.76
SRS-2 motivation M = 72.22, SD = 10.08 M = 69.00, SD = 11.68 3.20 0.83 3.33, 11.94
SRS-2 RIRB M = 75.33, SD = 16.49 M = 71.21, SD = 14.32 4.08 0.83 7.17, 14.98
ADOS-2 CS M = 7.44, SD = 1.33 M = 7.31, SD = 1.91 0.13 0.83 0.92, 1.09

SI suicidal ideation, Diff difference, CBCL childhood behavior checklist, Probs problems, MASC-C/P, multidimensional anxiety scale for chil-
dren, child-report and parent-report, SRS-2 social responsiveness scale, 2nd Edition, Comm communication, RIRB restricted interests and repeti-
tive behavior, ADOS-2 CS autism diagnostic observation schedule, 2nd edition comparison scores
*Based on Bootstrapping (BCa)
±Significant at a False Discovery Rate of 10%

Disease Control and Prevention; (CDC, 2017; Mayes et al., Kashani et al., 1989; Mayes et al., 2015). More specifically,
2013)). Endorsement of SI in our sample was consistent with parents of individuals with SI reported significantly higher
previous research (Hedley et al., 2018) relying solely on car- levels of both internalizing and externalizing problems in
egiver report/with community diagnosis. Though previous their children. With regard to internalizing problems, those
studies indicated higher SI among males and older adoles- with SI were reported to display significantly greater lev-
cents, (Brent et al., 1999; Mayes et al., 2013) there were els of affective problems (e.g., crying, feeling worthless,
no significant differences between individuals with SI and being nervous, worrying, being self-conscious) aligning
without SI with regard to demographic characteristics, such with depression, a recognized correlate of SI in the general
as age, sex, and ethnicity in our sample. Autism symptom population (Shtayermman, 2008).
severity, based on both parent and clinician report, was com- Though anxiety symptoms were clinically elevated in the
parable among those with and without SI; this contrasts with ASD group, no significant differences were found in individ-
prior findings of less severe symptomatology being related uals with and without SI. However, there were discrepancies
to SI (Brent et al., 1999; Mayes et al., 2013). This may be between self- and parent- report on the MASC-2. Signifi-
due to our sample generally having homogenous levels of cantly higher levels of humiliation and feelings of rejection
autism symptoms with limited range. were reported by parents of individuals with ASD with SI.
With regard to clinical features of SI, our results support In contrast to parental perception, youths with SI did not
previous findings that co-occurring mental health problems self-report feeling humiliated, rejected, but they reported
are associated with suicidal ideation (Gould et al., 1996; significantly higher levels of perfectionistic behaviors (e.g.,

13
Journal of Autism and Developmental Disorders

trying hard to obey others, trying to do things other people was assessed via both caregiver and self-report, emotional
would like, and trying to do things exactly right). and behavioral functioning was based solely on caregiver
With respect to theoretical accounts of the development report. Future research should include self-report of emo-
of suicidal ideation, our results align with both the cognitive tional and behavior problems such as the Youth Self-Report
model of suicide (Wenzel & Beck, 2008) and the Integrated (YSR; Achenbach & Rescorla, 2001) to further understand
Motivational-Volitional model of suicidal behavior (IMV; the emotional experiences of individuals with ASD. Fifth,
O’Connor & Kirtley, 2018), providing limited discriminant our results are based on a small predominately white sample.
evidence for one of the other. The cognitive model of suicide Though we were able to detect significant results, it will be
posits that suicidal behavior is based on dispositional vul- important to verify our findings in larger and more diverse
nerabilities (trait-like psychological variables (e.g., impul- samples. Furthermore, our study did not formally assess
sivity, neuroticism, and problem-solving impairments)) for co-occurring conditions and their possible impact on SI
and cognitive processes (distorted perceptions of external in ASD, an important objective for future research. Lastly,
or internal events or stimuli which distorts and their impact we solely explored the clinical features of SI in a sample
on experience) associated with both psychiatric disturbance which had an overall IQ and verbal IQ over 80 in order to
and suicidal behavior (Wenzel & Beck, 2008). The IMV ensure that participants would understand the questions on
model delineates a motivation phase related to SI in contrast the self-report measures of anxiety (MASC-C). Increasing
to a volitional phase related to actual suicidal behavior; we the IQ range for future studies is necessary to address SI in
focused on the former as most relevant to this study. In the individuals with ASD and lower cognitive ability.
motivational phase, predispositional factors and life expe- This study provides information important for both clini-
riences (e.g., feelings of defeat, rejection and humiliation cal practice and future research. Findings add to a growing
engender feelings of entrapment) contribute to a baseline body of literature demonstrating increased SI in youths with
suicide risk of SI (O’Connor & Kirtley, 2018). In ASD, ASD and highlight the relevance of co-occurring emotional
commonly experienced social isolation and peer rejection and behavioral challenges to the presence of SI. Our integra-
(Hedley et al., 2018; van Roekel et al., 2010), as well as tion of caregiver-report and self-report measures indicated
cognitive rigidity may contribute to this process. In keeping discrepancies in several areas and highlight the import of
with the cognitive model of suicide (Wenzel & Beck, 2008), integrating data across multiple informants in investigating
challenges associated with ASD spectrum may confer risk and assessing SI. In addition to these clinical metrics, future
for experiencing suicidal thoughts both through dispositional studies should investigate potentially informative biologi-
vulnerability factors associated with ASD (e.g., insistence cal differences between youth with ASD with SI and youth
on sameness) as well as cognitive processes associated with with ASD without SI. Examining biological markers is an
psychiatric distress (e.g., co-occurring anxiety and depres- important adjunct to parent/self-report in tracking SI risk
sion). As described by the IMV model (O’Connor & Kirt- clinical factors/symptoms in ASD over time to determine if
ley, 2018) we observed that experiences of humiliation and particular interventions decrease SI in the ASD population.
rejection are relevant to the experience SI in youth with Furthermore, recognizing that parents of youth with ASD
ASD. Additional research in larger samples is warranted to may be noticing their child’s affective difficulties as well as
better elucidate theoretical models of development of SI in their overt expression of feelings related to rejection may be
youth with ASD. key in terms of identifying these youth who are at greater
There are several study limitations that should be risk for developing SI. Taken together, this study provides
acknowledged. First, this study is exploratory and descrip- important insight into the clinical features associated with
tive and lacks a typically developing comparison group (the SI in ASD.
larger study from which this data was obtained excluded typ-
ically developing participants with any psychiatric difficul-
ties). Similarly, co-occurring psychiatric diagnoses for the Author Contributions  Formal analysis was done by KSE, SLJ and AN.
Funding acquisition was done by SLJ and JCM. Investigation was done
participants with ASD were not formally assessed. Second, by KSE and EJ. JCM done project administration. Resources, Supervi-
suicidal ideation was operationalized based upon endorse- sion, Validation was done by JCM and AN. Visualization was done by
ment of a single item on a caregiver-report scale and sub- KSE and AN. Original draft preparation was done by KSE. Review &
sequent verification by a clinician. It is possible that youth editing was done by KSE, AN and JCM. All authors contributed to the
study's conception and methodology.
with SI did not express suicidality to their caregivers, mak-
ing them unaware of the presence of SI when completing the Funding  Funding for this manuscript was provided by the National
CBCL/6–18, could be miscategorized as into the no suicidal Institute of Mental Health (R01 MH100173, U19 MH108206, R01
ideation group. Future studies can address this weakness MH107426). Scott Jackson received funding through the National Insti-
by utilizing multiple measures across multiple informants tute of Mental Health (T32 MH18268).
to assess the presence suicidal ideation. Third, anxiety

13
Journal of Autism and Developmental Disorders

Declarations  American Academy of Child & Adolescent Psychiatry, 57(5), 313-


320.e6.
Curtin, S. C., Warner, M., & Hedegaard, H. (2016). Suicide rates for
Conflict of interest  James C. McPartland consults with Customer Val-
females and males by race and ethnicity: United States, 1999 and
ue Partners, Bridgebio, Determined Health, and BlackThorn Thera-
2014
peutics, has received research funding from Janssen Research and De-
Davison, A., & Hinkley, D. (1997). Bootstrap methods and their appli-
velopment, serves on the Scientific Advisory Boards of Pastorus and
cation. Journal of the American Statistical Association. https://​
Modern Clinics, and receives royalties from Guilford Press, Lambert,
doi.​org/​10.​1017/​CBO97​80511​802843
and Springer.
Efron, B., & Tibshirani, R. (1993). An introduction to the bootstrap.
Chapman & Hall.
Ethical Approval  All procedures performed in studies involving human
Elliott, C. D. (2007). The differential abilities scales, (DAS-II). Pear-
participants were in accordance with the ethical standards of the insti-
son Education Inc.
tutional and/or national research committee and with the 1964 Helsinki
Gould, M. S., Fisher, P., Parides, M., Flory, M., & Shaffer, D. (1996).
declaration and its later amendments or comparable ethical standard.
Psychosocial risk factors of child and adolescent completed sui-
cide. Archives of General Psychiatry, 53(12), 1155–1162.
Informed Consent  Informed consent was obtained from all individual
Hedley, D., & Uljarević, M. (2018). Systematic review of suicide
participants included in the study.
in autism spectrum disorder: Current trends and implications.
Current Developmental Disorders Reports, 5(1), 65–76.
Hedley, D., Uljarević, M., Foley, K.-R., Richdale, A., & Trollor, J.
(2018). Risk and protective factors underlying depression and
References suicidal ideation in autism spectrum disorder. Depression and
Anxiety, 35(7), 648–657.
Hess, J. A., Matson, J. L., & Dixon, D. R. (2010). Psychiatric symptom
Achenbach, T. M., & Rescorla, L. (2001). Manual for the ASEBA endorsements in children and adolescents diagnosed with autism
school-age forms & profiles: An integrated system of multi-inform- spectrum disorders: A comparison to typically developing chil-
ant assessment. Burlington: Aseba Burlington. dren and adolescents. Journal of Developmental and Physical
Achenbach, T. M., Dumenci, L., & Rescorla, L. A. (2003). DSM-ori- Disabilities, 22(5), 485–496.
ented and empirically based approaches to constructing scales Hirvikoski, T., Boman, M., Chen, Q., D’Onofrio, B. M., Mittendor-
from the same item pools. Journal of Clinical Child and Adoles- fer-Rutz, E., Lichtenstein, P., Bölte, S., & Larsson, H. (2020).
cent Psychology, 32(3), 328–340. Individual risk and familial liability for suicide attempt and
Association, A. P. (2013). Diagnostic and Statistical Manual of Mental suicide in autism: A population-based study. Psychological
Disorders (DSM-5®). Washington: American Psychiatric Pub. Medicine, 50(9), 1463–1474.
Brent, D. A., Baugher, M., Bridge, J., Chen, T., & Chiappetta, L. Hunsche, M. C., Saqui, S., Mirenda, P., Zaidman-Zait, A., Ben-
(1999). Age-and sex-related risk factors for adolescent suicide. nett, T., Duku, E., Elsabbagh, M., Georgiades, S., Smith, I.
Journal of the American Academy of Child & Adolescent Psy- M., Szatmari, P., Ungar, W. J., Vaillancourt, T., Waddell, C.,
chiatry, 38(12), 1497–1505. Zwaigenbaum, L., & Kerns, C. M. (2020). Parent-reported rates
Carleton, R. N., Mulvogue, M. K., Thibodeau, M. A., McCabe, R. E., and clinical correlates of suicidality in children with autism
Antony, M. M., & Asmundson, G. J. G. (2012). Increasingly cer- spectrum disorder: A longitudinal study. Journal of Autism and
tain about uncertainty: Intolerance of uncertainty across anxiety Developmental Disorders, 50(10), 3496–3509.
and depression. Journal of Anxiety Disorders, 26(3), 468–479. Hus, V., & Lord, C. (2014). The autism diagnostic observation
https://​doi.​org/​10.​1016/j.​janxd​is.​2012.​01.​011 schedule, module 4: Revised algorithm and standardized sever-
Cassidy, S., Bradley, P., Robinson, J., Allison, C., McHugh, M., & ity scores. Journal of Autism and Developmental Disorders,
Baron-Cohen, S. (2014). Suicidal ideation and suicide plans or 44(8), 1996–2012.
attempts in adults with Asperger’s syndrome attending a specialist Kanne, S. M., & Mazurek, M. O. (2011). Aggression in children and
diagnostic clinic: A clinical cohort study. The Lancet Psychiatry, adolescents with ASD: Prevalence and risk factors. Journal of
1(2), 142–147. Autism and Developmental Disorders, 41(7), 926–937.
CDC. (2017). WISQARS (Web-based Injury Statistics Query and Kashani, J. H., Goddard, P., & Reid, J. C. (1989). Correlates of sui-
Reporting System) Injury Center|CDC. Retrieved April 2017 from cidal ideation in a community sample of children and adoles-
https://​www.​cdc.​gov/​injury/​wisqa​rs/​index.​html cents. Journal of the American Academy of Child & Adolescent
Cha, C. B., Franz, P. J., Guzmán, E. M., Glenn, C. R., Kleiman, E. M., Psychiatry, 28(6), 912–917.
& Nock, M. K. (2018). Annual research review: suicide among La Buissonnière Ariza, V., Schneider, S. C., Cepeda, S. L., Wood, J.
youth – epidemiology, (potential) etiology, and treatment. Journal J., Kendall, P. C., Small, B. J., Wood, K. S., Kerns, C., Saxena,
of Child Psychology and Psychiatry, 59(4), 460–482. K., & Storch, E. A. (2021). Predictors of suicidal thoughts in
Chen, M.-H., Pan, T.-L., Lan, W.-H., Hsu, J.-W., Huang, K.-L., Su, children with autism spectrum disorder and anxiety or obses-
T.-P., Li, C.-T., Lin, W.-C., Wei, H.-T., & Chen, T.-J. (2017). Risk sive-compulsive disorder: The unique contribution of exter-
of suicide attempts among adolescents and young adults with nalizing behaviors. Child Psychiatry & Human Development.
autism spectrum disorder: A nationwide longitudinal follow-up https://​doi.​org/​10.​1007/​s10578-​020-​01114-1
study. The Journal of Clinical Psychiatry, 78(9), e1174–e1179. Lord, C., Rutter, M., DiLavore, P. C., Risi, S., & Katherine, & Somer
Constantino, J., & Gruber, C. (2012). Social Responsiveness Scale L. (2002). Autism diagnostic observation schedule. Los Ange-
Second Edition (SRS-2) Manual. Western Psychological Services les: Western Psychological Services, 29, 30.
(WPS) Mannion, A., Brahm, M., & Leader, G. (2014). Comorbid psychopa-
Culpin, I., Mars, B., Pearson, R. M., Golding, J., Heron, J., Bubak, thology in autism spectrum disorder. Review Journal of Autism
I., Carpenter, P., Magnusson, C., Gunnell, D., & Rai, D. (2018). and Developmental Disorders, 1(2), 124–134.
Autistic traits and suicidal thoughts, plans, and self-harm in March, J. S., Parker, J. D. A., Sullivan, K., Stallings, P., & Conners,
late adolescence: Population-based cohort study. Journal of the C. K. (1997). The multidimensional anxiety scale for children
(MASC): Factor structure, reliability, and validity. Journal of

13
Journal of Autism and Developmental Disorders

the American Academy of Child & Adolescent Psychiatry, 36(4), Shtayermman, O. (2008). Suicidal ideation and comorbid disorders
554–565. in adolescents and young adults diagnosed with Asperger’s syn-
Mayes, C., & S. L., Baweja, R., & Mahr, F. (2015). Suicide ideation drome: A population at risk. Journal of Human Behavior in the
and attempts in children with psychiatric disorders and typical Social Environment, 18(3), 301–328.
development. Crisis, 36(1), 55–60. Storch, E. A., Sulkowski, M. L., Nadeau, J., Lewin, A. B., Arnold,
Mayes, G., & A. A., Hillwig-Garcia, J., & Syed, E. (2013). Sui- E. B., Mutch, P. J., Jones, A. M., & Murphy, T. K. (2013). The
cide ideation and attempts in children with autism. Research in phenomenology and clinical correlates of suicidal thoughts and
Autism Spectrum Disorders, 7(1), 109–119. behaviours in youth with autism spectrum disorders. Journal of
McDonnell, C. G., DeLucia, E. A., Hayden, E. P., Anagnostou, E., Autism and Developmental Disorders, 43(10), 2450–2459.
Nicolson, R., Kelley, E., Georgiades, S., Liu, X., & Stevenson, Thissen, D., Steinberg, L., & Kuang, D. (2002). Quick and easy imple-
R. A. (2020). An exploratory analysis of predictors of youth mentation of the Benjamini-Hochberg procedure for controlling
suicide-related behaviours in autism spectrum disorder: Impli- the false positive rate in multiple comparisons. Journal of Edu-
cations for prevention science. Journal of Autism and Develop- cational and Behavioral Statistics, 27(1), 77–83.
mental Disorders, 50(10), 3531–3544. van Roekel, E., Scholte, R. H. J., & Didden, R. (2010). Bullying among
Moses, T. (2017). Suicide attempts among adolescents with self- adolescents with autism spectrum disorders: Prevalence and per-
reported disabilities. Child Psychiatry & Human Development. ception. Journal of Autism and Developmental Disorders, 40(1),
https://​doi.​org/​10.​1007/​s10578-​017-​0761-9 63–73.
Mukaddes, N., & Fateh, R. (2010). High rates of psychiatric co-mor- Walters, S. J., & Campbell, M. J. (2004). The use of bootstrap methods
bidity in individuals with Asperger’s disorder. The World Journal for analysing health-related quality of life outcomes (particularly
of Biological Psychiatry, 11(2), 486–492. the SF-36). Health and Quality of Life Outcomes, 2(1), 70.
NIMH: Suicide. (2018). Retrieved May 2018 from https://​www.​nimh.​ Wenzel, A., & Beck, A. T. (2008). A cognitive model of suicidal behav-
nih.​gov/​health/​stati​stics/​suici​de.​shtml#​part_​154969 ior: Theory and treatment. Applied and Preventive Psychology,
O’Connor, R. C., & Kirtley, O. J. (2018). The integrated motivational– 12(4), 189–201.
volitional model of suicidal behaviour. Philosophical Transactions Wijnhoven, L. A., Niels-Kessels, H., Creemers, D. H., Vermulst, A. A.,
of the Royal Society B: Biological Sciences, 373(1754), 20170268. Otten, R., & Engels, R. C. (2019). Prevalence of comorbid depres-
Pezzimenti, F., Han, G. T., Vasa, R. A., & Gotham, K. (2019). Depres- sive symptoms and suicidal ideation in children with autism spec-
sion in youth with autism spectrum disorder. Child and Adolescent trum disorder and elevated anxiety symptoms. Journal of Child &
Psychiatric Clinics of North America, 28(3), 397–409. Adolescent Mental Health, 31(1), 77–84.
Posner, K., Brent, D., Lucas, C., Gould, M., Stanley, B., Brown, G., Wright, D. B., London, K., & Field, A. P. (2011). Using bootstrap
Fisher, P., Zelazny, J., Burke, A., & Oquendo, M. (2008). Colum- estimation and the plug-in principle for clinical psychology data.
bia-suicide severity rating scale (C-SSRS) (p. 10). Columbia Uni- Journal of Experimental Psychopathology, 2(2), 252–270.
versity Medical Center. Zeedyk, S. M., Rodriguez, G., Tipton, L. A., Baker, B. L., & Blacher, J.
Reszka, S. S., Boyd, B. A., McBee, M., Hume, K. A., & Odom, S. (2014). Bullying of youth with autism spectrum disorder, intellec-
L. (2014). Brief Report: Concurrent validity of autism symptom tual disability, or typical development: Victim and parent perspec-
severity measures. Journal of Autism and Developmental Disor- tives. Research in Autism Spectrum Disorders, 8(9), 1173–1183.
ders, 44(2), 466–470.
Rutter, M., Le Couteur, A., & Lord, C. (2003). Autism diagnostic inter- Publisher's Note Springer Nature remains neutral with regard to
view-revised. Los Angeles, CA: Western Psychological Services, jurisdictional claims in published maps and institutional affiliations.
29(2003), 30.
Rynn, M. A., Barber, J. P., Khalid-Khan, S., Siqueland, L., Dembiski,
M., McCarthy, K. S., & Gallop, R. (2006). The psychometric
properties of the MASC in a pediatric psychiatric sample. Journal
of Anxiety Disorders, 20(2), 139–157.

13

You might also like