Online Self-Injury Activities among Psychiatrically Hospitalized Adolescents: Prevalence, Functions, and Perceived Consequences

Abstract

The majority of adolescents with psychiatric disorders use social media, engaging in a range of online activities that may confer both risks and benefits. Very little work, however, has examined engagement in online activities related to self-injury among these youth, such as posting about self-injury, viewing self-injury related content, or messaging about self-injury with online or offline friends. This study examined the frequency and types of online self-injury activities in which adolescents engage, perceived functions that these activities serve, and associated risk for self-injurious thoughts and behaviors (SITBs). Participants were 589 psychiatrically-hospitalized adolescents (Mage = 14.88), who completed self-report measures assessing online self-injury activities, perceived functions and consequences of these activities, and SITBs. Results indicated that 43.3% of the sample had engaged in online self-injury activities, that the majority (74.8%) used social networking sites (e.g., Snapchat, Instagram) to do so, and that these activities were significantly more common among sexual and gender minority youth. Adolescents who talked about self-injury with friends met online were more likely to report a history of suicide attempt(s). A latent profile analysis revealed three distinct subgroups of youth based on their perceived functions of engaging in online self-injury activities. Subgroups reporting higher levels of engagement for purposes of identity exploration, self-expression, and aiding recovery were at heightened risk for negative perceived consequences of these activities and reported greater suicidal ideation severity. Findings offer new insights for identifying youth who may be at heightened risk for SITBs in the context of social media use.

Full text

Vol.:(0123456789)
1 3
Research on Child and Adolescent Psychopathology
https://doi.org/10.1007/s10802-020-00734-4
Online Self‑Injury Activities amongPsychiatrically Hospitalized
Adolescents: Prevalence, Functions, andPerceived Consequences
JacquelineNesi1,2 · TaylorA.Burke1,2· HannahR.Lawrence3,4· HeatherA.MacPherson1,5· AnthonySpirito1·
JenniferC.Wol1,2
Accepted: 1 November 2020
© Springer Science+Business Media, LLC, part of Springer Nature 2021
Abstract
The majority of adolescents with psychiatric disorders use social media, engaging in a range of online activities that may
confer both risks and benefits. Very little work, however, has examined engagement in online activities related to self-injury
among these youth, such as posting about self-injury, viewing self-injury related content, or messaging about self-injury
with online or offline friends. This study examined the frequency and types of online self-injury activities in which adoles-
cents engage, perceived functions that these activities serve, and associated risk for self-injurious thoughts and behaviors
(SITBs). Participants were 589 psychiatrically-hospitalized adolescents (Mage = 14.88), who completed self-report measures
assessing online self-injury activities, perceived functions and consequences of these activities, and SITBs. Results indicated
that 43.3% of the sample had engaged in online self-injury activities, that the majority (74.8%) used social networking sites
(e.g., Snapchat, Instagram) to do so, and that these activities were significantly more common among sexual and gender
minority youth. Adolescents who talked about self-injury with friends met online were more likely to report a history of
suicide attempt(s). A latent profile analysis revealed three distinct subgroups of youth based on their perceived functions of
engaging in online self-injury activities. Subgroups reporting higher levels of engagement for purposes of identity explora-
tion, self-expression, and aiding recovery were at heightened risk for negative perceived consequences of these activities
and reported greater suicidal ideation severity. Findings offer new insights for identifying youth who may be at heightened
risk for SITBs in the context of social media use.
Keywords Adolescents· Self-injury· Suicide· Social media· Online
Introduction
Social media plays a central role in the lives of most ado-
lescents. Nearly 97% of youth ages 13–17 use some form of
social media (Anderson and Jiang2018), broadly defined as
any digital tool that allows for social interaction (Moreno
and Kota2013), including social networking sites, messag-
ing apps, and video sharing sites. The diverse landscape
of social media tools gives rise to a number of possible
online activities, some of which may confer benefits for
youth, including opportunities for self-expression, creativ-
ity, and social support (Anderson and Jiang2018; Rideout
and Robb2018). However, a number of risks are present as
well, including possibilities for cybervictimization, social
comparison, and exposure to risky or inappropriate content
(John etal. 2018; Nesi and Prinstein2015; Rideout and
Robb2018). Although youth with mental health concerns
may stand to benefit from opportunities for social support
Electronic Supplementary Material The online version of this
article (https ://doi.org/10.1007/s1080 2-020-00734 -4) contains
supplementary material, which is available to authorized users.
* Jacqueline Nesi
jacqueline_nesi@brown.edu
1 Department ofPsychiatry & Human Behavior, Warren
Alpert Medical School ofBrown University, RI, Providence,
USA
2 Bradley Hasbro Research Center, Rhode Island Hospital,
Providence, RI, USA
3 McLean Hospital, Belmont, MA, USA
4 Harvard Medical School, Cambridge, MA, USA
5 Emma Pendleton Bradley Hospital, Riverside, RI, USA

Research on Child and Adolescent Psychopathology
1 3
and psychoeducation online, they may also be especially
vulnerable to potential risks associated with social media.
For example, a range of possible activities related to self-
injurious thoughts and behaviors (SITBs) may be readily
available online, such as posting about self-injury, viewing
self-injury related content, or messaging about self-injury
with online or offline friends. However, very little work has
explored patterns of engagement in these activities and their
associated risks, particularly among youth with acute mental
health concerns. Research is needed to examine the types of
online self-injury activities in which adolescents engage, as
well as the range of functions these activities may serve for
youth with psychiatric symptoms.
Self-injurious behaviors, or deliberate acts of self-harm
performed with or without suicidal intent, represent a sig-
nificant public health concern among young people. Preva-
lence rates, based on a systematic review, estimate that
16% of adolescents have engaged in deliberate self-harm
(Muehlenkamp etal.2012). Rates of suicide, self-injury, and
depressive symptoms in youth have risen substantially over
the past decade, coinciding with the widespread adoption
of social media (Twenge etal.2018). Although the ques-
tion remains whether social media has contributed to ris-
ing rates of depression and SITBs, it is critical to identify
specific online behaviors that may exacerbate or ameliorate
symptoms. Further, it is necessary to examine individual
characteristics of youth who are at particularly high risk for
negative effects of social media, such as those related to
self-injury.
Interpersonal Theories ofSuicide Risk
The potential risks and benefits of adolescents’ engagement
in online self-injury activities may be understood within
the context of interpersonally-focused theories of suicidal
behavior. In particular, the Interpersonal Theory of Suicide
(Joiner2005; Van Orden etal.2010) posits that feelings of
perceived burdensomeness and thwarted belongingness con-
tribute to suicidal ideation, and that these factors, in com-
bination with an acquired capability for suicide, increase
risk for suicidal behavior. On the one hand, engagement in
online self-injury activities may serve to reduce thwarted
belongingness, as online social support and media-based
psychoeducation may protect against loneliness and isola-
tion (De Choudhury andKiciman2017; Niederkrotenthaler
etal.2014). Indeed, prior systematic reviewsofonline
activities related to deliberate self-harm (Dyson etal.2016)
and nonsuicidal self-injury (NSSI; Lewis and Seko2016)
highlight social benefits such as developing community,
receiving messages and advice encouraging recovery, and
engaging in emotional self-disclosure. On the other hand,
exposure to images, videos, and other content related to self-
injury may decrease fear of death through repeated exposure,
resulting in greater acquired capability for suicide (Smith
etal.2010). In line with this hypothesis, preliminary evi-
dence suggests that exposure to such content may increase
suicide risk (Arendt etal.2019).
Interpersonal factors influencing suicide risk may be
especially relevant for adolescents, given their increased
reliance on peer feedback and heightened biological sensi-
tivity to social information and rewards during this devel-
opmental period (Foulkes and Blakemore2016). In line
with social cognitive theories (Bandura2001), which sug-
gest that individuals’ behavior is informed and influenced
through observation of others’ behavior, adolescents have
been shown to be at heightened risk for suicide following
exposure to news or information about other youths’ suicidal
behavior (Hawton etal.2020). In addition, a large body of
literature has highlighted the role of both media and peer
factors in adolescent self-injury and suicide (King and Mer-
chant2008; Niederkrotenthaler and Stack2017). For exam-
ple, media depictions of suicide have been found to influ-
ence youth suicide risk (Gould etal.2003). Moreover, peer
contagion effects, through both selection and socialization,
play an important role in SITBs through reinforcement and
normalization of self-injurious behavior (Heilbron and Prin-
stein2008; Insel and Gould2008). Thus, exposure to online
self-injury content may not only increase acquired capability
for suicide, but may also trigger urges to self-injure and lead
to the reinforcement and normalization of SITBs (Dyson
etal.2016; Lewis and Seko2016).
The Role ofSocial Media inOnline Self‑Injury
Activities
Social media may represent a confluence of peer and media
influences on self-injury, as it provides new opportunities for
adolescents to engage with other youth who self-injureand
increases the breadth and depth of adolescents’ exposure
to the wider media landscape. Indeed, theoretical work on
social media has noted the potential for these platforms to
amplify both helpful and harmful interpersonal influences
due to their constant accessibility, public audiences, and
potential for anonymity (Nesi etal.2018). Examining how
these effects are related to online self-injury activities on
social media sites is crucial. Themajority of prior studies
have examined online self-injury activities as they occur in
discussion forums and chatrooms designed specifically for
individuals who self-injure, and studies have only recently
begun to explore individuals’ online self-injury activities on
modern social networking sites, such as Instagram (Arendt
2019; Brown etal.2018; Carlyle etal.2018). Yet the social
media landscape is now larger and more diverse than ever.
This creates a plethora of new online self-injury activities in
which youth can engage, with possibilities for sharing and
viewing multiple types of self-injury content (e.g., videos,

Research on Child and Adolescent Psychopathology
1 3
messages, posts, photos) with various audiences (i.e., in per-
son and online ‘friends’) across numerous platforms (e.g.,
social networking sites, apps). As the possibilities for online
self-injury activities expand, it becomes even more critical
to identify specific types of activities that may put youth at
greater risk for SITBs.
Online Activities amongSexual andGender
Minority Youth
The Differential Susceptibility Model of Media Effects high-
lights the importance of examining the specific media behav-
iors in which youth engage, as well as individual differences
that may make certain youth more vulnerable, or resilient,
to media effects (Valkenburg and Peter2013). In the con-
text of online self-injury activities, it is critical to examine
whether certain youth may be especially susceptible to nega-
tive and/or positive effects. Youth identifying as sexual and/
or gender minorities (SGM) may be more likely to experi-
ence both risks and benefits of online self-injury activities.
Prior work suggests that SGM youth are at heightened risk
for SITBs, compared to non-sexual minority and cisgender
youth (Johns etal.2019; Peters etal.2020). SGM youth
are also more likely to have online friends whom they have
never met in person (Ybarra etal.2015), and for SGM youth
with histories of self-injury, it is possible that engagement
with these online-only friends may involve discussions of
SITBs. Notably, sexual minority youth have been found to be
more likely to seek support for self-injury online (Frost and
Casey2016). Thus, more work is needed examining whether
SGM youth differ from non-SGM youth in their online self-
injury activities, and whether this may have implications
for SITB risk.
Online Self‑Injury Activities andSITBs
Prior systematic reviews of the literature on online activi-
ties related to deliberate self-harm (Dyson etal.2016) and
NSSI (Lewis and Seko2016) highlight a range of possible
risks and benefits of these activities. However, the major-
ity of studies in this area have used qualitative methods
and most involve direct observation and thematic coding
of online self-injury content (Dyson etal.2016; Lewis and
Seko2016). Only a small number of studies have directly
surveyed youth regarding motivations for engaging in these
activities, and even fewer studies have examined associations
among specific online self-injury activities, motivations, and
SITBs. This limited research is problematic, given initial
evidence that individuals who engage in online self-injury
activities are a clinically high-risk group. Recent studies
suggest that exposure to online content related to self-injury
is associated with greater likelihood of self-injurious behav-
iors. In a cross-sectional study of over 400 young adults,
those who self-reported online exposure to risky behavior
were more likely to report self-harm and other offline risky
behaviors (i.e., drug use, excessive alcohol use,disordered
eating; Branley and Covey2017). A longitudinal study of
over 700 young adults also found that exposure to self-harm
content on Instagram was associated with increases in self-
harming behavior, suicidal ideation, suicide risk, and hope-
lessness one month later (Arendt etal.2019).
Furthermore, recent evidence indicates that young peo-
ple who use the Internet to communicate about suicide or
search for suicide-related information may be at higher
risk for suicide than those who do not (Bell etal.2018;
Mok etal.2016; Niederkrotenthaler etal.2017). Among
suicidal young adults, those who had engaged in suicide-
related social media use were more socially anxious (Bell
etal.2018; Mok etal.2016), more depressed (Niederk-
rotenthaler etal.2017), and at higher risk for suicide (Bell
etal.2018; Mok etal.2016; Niederkrotenthaler etal.2017)
than those individuals who had not. In addition, in a sample
of over 600 college students with a history of self-injury,
those who sought help from peers online for self-injury
reported greater distress, suicidality, and more frequent
self-injury than those who did not (Frost and Casey2016).
Despite this important initial work, significant gaps in
the literature remain. Quantitative studies with large sam-
ples of adolescents are needed to explore the prevalence
and associated risks of online self-injury activities. Studies
focusing on youth with psychiatric concerns and those who
identify as SGM are also important, given these youth are at
high risk for future SITBs and may be more susceptible to
potential negative effects of social media. Furthermore, there
are likely a multitude of reasons why adolescents engage
in activities related to self-injury on social media. Among
those youth who do engage in online self-injury activities,
identifying potential subgroups based on their associated
motivations may provide valuable information regarding
individual risk profiles. It may be that youth who endorse
certain patterns of perceived functions (e.g., supporting
a self-injury based personal identity, regulating negative
emotions) versus others (e.g., seeking support and recovery
encouragement) may differ in their SITB risk. Identifying
these potential subgroups is important to inform targeted
assessment, prevention, and intervention efforts.
The Current Study
The current study explores patterns of online self-injury
activities in a large, diverse sample of adolescents who
were psychiatrically hospitalized due to risk of harm to self
or others. We examine the proportion of youth ever having
engaged in four online activities: viewing self-injury con-
tent, sharing self-injury content, talking about self-injury
with peers known only online, and talking about self-injury

Research on Child and Adolescent Psychopathology
1 3
(using technology) with peers known offline. We also exam-
ine the digital tools used to engage in these activities, and
associated risks of suicidal ideation (SI), suicide attempts,
and NSSI. Finally, to identify those youth who may be at
highest risk based on their online self-injury activities, we
classify subgroups of adolescents based on the perceived
functions that these activities serve for them. We examine
whether these subgroups differ in demographics, the type of
online activities in which they engage, their perceived con-
sequences of these activities, and their clinical risk profiles,
including SI, attempts, and NSSI.
Methods
Participants
Participants in the current study were 589 psychiatrically
hospitalized adolescents ages 11–18 (M = 14.88, SD = 1.83).
A total of 55.7% identified as female, 35.1% as male, 4.3% as
transgender, 3.7% in another way (e.g., gender nonconform-
ing, other, not sure), and 1.2% declined to state. In regard
to sexuality, 48.6% identified as heterosexual or straight,
25.7% as bisexual or pansexual, 10.6% in another way (e.g.,
asexual, other, not sure), 6.5% as gay or lesbian, and 8.7%
declined to state. Participants were 68.0% White, 14.0%
Black, 1.0% Asian, 0.9% Native Hawaiian/Pacific Islander,
0.7% American Indian/Alaskan Native, 15.0% other races,
and 0.3% declined to state; 24.7% of the sample was His-
panic/Latinx. The hospital Institutional Review Board (IRB)
approved this research. The measures administered and ana-
lyzed as part of this study were part of a standardized admis-
sion process for all adolescents admitted to the inpatient
unit. This assessment battery was used to provide informa-
tion on clinical presentations and enhance unit program-
ming. Assessment results were entered in patient charts.
This study consists of a retrospective chart review; the IRB
approved a waiver of informed consent.
General clinical characteristics. The modal length of
stay on the inpatient unit was 9days. Approximately 53%
of teens are insured by Medicaid. Chart data indicate that
the vast majority (88%) of adolescents admitted to the unit
met criteria for DSM-5 current Major Depressive Episode.
Adolescents also present with a wide range of other inter-
nalizing and externalizing psychiatric disorders, with the
most common diagnoses being Oppositional Defiant Disor-
der (43%), Generalized Anxiety Disorder (37%), Attention
Deficit/Hyperactivity Disorder (34.5%), Conduct Disorder
(28%), Post Traumatic Stress Disorder (19%), and Substance
Use Disorder (17%). These clinical characteristics are rep-
resentative of the full inpatient unit and include information
about patients not represented in the final study sample.
Measures
Suicidal Ideation. The Suicide Ideation Questionnaire
– Junior (SIQ-Jr; Reynolds and Mazza1999) is a 15-item
self-report measure employed to assess severity of suicidal
ideation over the past 30days. Items are rated on a seven-
point Likert scale, from 0 (I’ve never had this thought)
to 6 (Almost every day). Items are summed, with higher
scores reflecting greater severity of SI. The SIQ-Jr has
good psychometric properties (Reynolds and Mazza1999),
and evidence suggests it prospectively predicts suicidal
ideation among psychiatrically hospitalized adolescents
(King etal. 1997). In this sample, internal consistency was
excellent (alpha = 0.96).
Suicide Attempt. A single self-report item was adapted
from the Self-Injurious Thoughts and Behaviors Interview
(SITBI; Nock etal.2007) to assess lifetime history of sui-
cide attempts. Participants were asked to respond yes or
no the question, “Have you ever made an actual suicide
attempt, where you were trying to kill yourself, even just a
little?” Responses were coded 0 for no and 1 for yes.
Nonsuicidal Self-Injury (NSSI). To assess lifetime
history of engagement in NSSI, a single self-report item
was adapted from the SITBI (Nock etal.2007), asking
“Have you ever in your life done anything to purposefully
hurt yourself without trying to die (for example, cutting
or burning your skin)?” Responses were coded 0 for no
and 1 for yes.
Online Self-Injury Activities. A questionnaire assess-
ing the presence of a range of self-injury related social
media experiences, as well as the functions and conse-
quences of these experiences, was developed to improve
clinical care on the inpatient unit (see Online Supple-
ment 1 for the full measure). The information collected
was used to inform topics covered in group therapy, and
to guide individual therapy as part of the standard intake
process. The measure was developed through review of
prior literature on potential benefits and consequences of
online self-injury activities (e.g., Dyson etal.2016; Lewis
and Seko2016), and a process of iterative feedback in
consultation with senior investigators, research staff, and
clinicians with expertise in working with clinically acute
adolescents. Participants were asked whether they engaged
in four online activities: (1) viewing content related to
self-injury, (2) sharing content related to self-injury, (3)
talking (using technology) to others known offline (i.e., in
“real life”) about self-injury, and (4) talking (using tech-
nology) to others known only online (i.e., never met in
person) about self-injury. Responses were coded as either
0 = Never did this or 1 = Did this at least once in lifetime.
Only participants who responded positively to at least one
of these items were asked the questions described below.

Research on Child and Adolescent Psychopathology
1 3
Websites/Apps Used. Participants were asked “Which of
the following websites or apps did you use to do these activi-
ties (i.e., talking about, sharing, or looking at posts/photos
related to injuring oneself)?” They were asked to check all
response options that applied including: (1) social network-
ing site (e.g., Snapchat, Instagram, Twitter, Facebook), (2)
video sharing site (e.g., YouTube), (3) chat or discussion
forum/website specifically for people who injure themselves,
(4) text messaging or messaging apps (e.g., WhatsApp), and/
or (5) other websites or apps.
Functions. Participants rated 9 items on a Likert scale
from 0 (strongly disagree) to 4 (strongly agree) assessing
the extent to which they identified with a range of func-
tions of their own self-injury related online experiences.
Specifically, we examined seven functions with one or two
items per function. (1) Negative affect regulation: “I am
more likely to do these activities when I am feeling a nega-
tive emotion (e.g., upset, sad, angry)”, “I am more likely to
do these online activities when having thoughts of injur-
ing myself”, (2) Boredom reduction: “I am more likely to
do these activities when I am feeling neutral or bored”, (3)
Positive affect enhancement: “I am more likely to do these
online activities when I am feeling a positive emotion (e.g.,
happy, excited)”. (4) Reduced isolation: “I do these online
activities in order to feel less isolated or alone”. (5) Self-
expression: “I do these online activities to “vent” or express
how I feel”. (6) Recovery: “I do these online activities to
help me try and get better”. (7) Identity exploration: “Doing
these online activities helps me feel like I’m part of a group
of other people like me”, and “Doing these online activities
allows me to express who I really am”. Means ranged from
1.39 (SD = 1.28) to 2.17 (SD = 1.30), Skewness from -0.39
to 0.45, and Kurtosis from -1.27 to -0.79 (see Table S1).
Perceived consequences. Participants were adminis-
tered a total of ten items, rated on a Likert scale from 0
(strongly disagree) to 4 (strongly agree) assessing the extent
to which they experienced negative consequences of self-
injury related online experiences. All items were analyzed
individually. Participants indicated, using single items,
whether online self-injury activities led to the (1) normali-
zation of self-injury (“Doing these activities makes me
think that injuring myself is something I could do in my
situation”), the experience of (2) thwarted recovery (“Doing
these activities makes me think that I am unlikely to ever
stop having thoughts of injuring myself”), (3) behavioral
triggering (“Doing these activities makes me more likely to
act on thoughts of injuring myself”), (4) social comparison
(“Doing these activities causes me to compare my experi-
ence of self-injury with others’ experiences of self-injury”)
and (5) discovery of new methods (“Doing these activities
has introduced me to new methods for injuring myself”).
These perceived consequences items had a range of 0 to 4,
means ranged from 1.17 (SD = 1.24) to 1.40 (SD = 1.18),
Skewness from 0.22 to 0.76, and Kurtosis from -0.53 to
-1.18 (see Table S1). Five administered items were excluded
from analyses: four items were excluded because they asked
about consequences of specific online activities that were
not relevant to all participants (e.g., “I have been harassed or
teased online after posting about injuring myself”; “When I
am having thoughts of hurting myself, viewing photographs
or videos about injuring oneself helps reduce this urge”).
One item was excluded because it was redundant with an
item already analyzed (i.e., “Doing these online activities
makes me less likely to act on thoughts of injuring myself”).
Analytic Approach
First, we conducted descriptive statistics within the full sam-
ple to determine the proportion of participants who had ever
engaged in the four online self-injury activities (i.e., viewing
self-injury content, sharing self-injury content, talking to “real
life” friends about self-injury, and talking to online friends
about self-injury) and compared these proportions across
participants of different genders and sexual identities. We
also explored which websites and apps were most commonly
used to engage in these activities. Next, we conducted bivari-
ate logistic and linear regression analyses in SPSS 22.0 to
simultaneously examine associations between each of the four
online self-injury activities and SITBs (i.e., suicidal ideation,
history of NSSI, and history of one or more suicide attempts),
controlling for gender, sexuality, age, race, and ethnicity. All
regression analyses were also repeated with the other SITBs
included as covariates (i.e., for the model predicting suicide
attempt history, we controlled for NSSI and SI; for the SI
model, we controlled for NSSI and suicide attempt history;
and for the NSSI model, we controlled for SI and suicide
attempt history).
The sample was then limited only to those who had
engaged in at least one of the four online self-injury activi-
ties (n = 254), as these were the only participants who com-
pleted follow up measures assessing the functions of these
activities. We conducted latent profile analysis (LPA) using
Mplus 8.2 (Muthén and Muthén 1998–2018) within this
subsample to identify profiles based on adolescents’ self-
reported functions of engaging in these activities. Seven
standardized indicators representing various functions
of online self-injury activities were employed to estimate
classes, with full information maximum likelihood used to
handle missing data. Two items were excluded from analy-
ses, as each was redundant with another item: specifically,
one of the items assessing negative affect regulation (i.e.,
engaging in activities when having thoughts of self-injury)
was redundant with the other (r = 0.77), and the item assess-
ing positive affect enhancement was redundant with the
boredom reduction item (r = 0.64). Classes were compared

Research on Child and Adolescent Psychopathology
1 3
empirically by examining the Akaike Information Criteria
(AIC), Bayesian Information Criteria (BIC), sample-size
adjusted BIC (aBIC), and the parametric bootstrap likeli-
hood ratio test (BLRT) (Nylund etal.2007). The optimal
model was determined empirically based on the BIC, which
is considered to be the most reliable information criteria
and is thus recommended to weight most strongly (Nylund
etal.2007). Interpretability of each class solution was also
considered when selecting the optimal number of classes.
We examined differences among classes in terms of
engagement in each of the four online self-injury activities,
perceived consequences of engaging in these activities,
SITBs, and demographic variables (age, gender identity,
and sexual orientation). Comparisons were conducted with
chi-square tests using Lanza’s method (Lanza etal.2013) for
inclusion of distal outcomes (i.e., online self-injury activi-
ties, perceived consequences, and SITBs) within the latent
class model (Asparouhov and Muthén2014). This model-
based approach uses Bayes theorem to derive the joint dis-
tribution of the latent class and distal variables, and was
estimated using the DCAT (for categorical variables) and
DCON (for continuous variables) specifications for auxil-
iary variables in MPlus 8.0. If omnibus differences in distal
outcomes across classes emerged, we examined pairwise
comparisons between classes on mean parameters (for con-
tinuous variables) or probabilities (for categorial variables).
Results
Descriptive Statistics andRegression Analyses
Of the total sample of 589 adolescents, 254 (43.3%) reported
having ever engaged in any online activities related to self-
injury. The most common activity (31.8% of participants) was
using technology to talk about self-injury with people known
in “real life” (i.e., not exclusively online), followed by viewing
content online (i.e., others’ posts, messages, and videos) related
to self-injury (26.5% of participants). Using technology to talk
about self-injury with people known only online (16.8% of par-
ticipants) and sharing or posting one’s own content related to
self-injury (14.5% of participants) were relatively less common.
In general, gender minority adolescents (i.e., those who do not
identify as male or female) were most likely to have engaged
in online self-injury activities, followed by female adolescents,
and then male adolescents (see Table1). In addition, sexual
minority youth were more likely than non-sexual minority
youth to have engaged in all online self-injury activities.
The 254 adolescents who had engaged in online activities
related to self-injury used a variety of apps and websites to do
so (see Table2). The majority of adolescents (n = 190, 74.8%)
used a social networking site like Snapchat or Instagram. About
one-third of youth used text messaging or messaging apps (n
= 86, 33.9%). Less commonly, adolescents engaged in online
Table 1 Engagement in online self-injury activities, with comparisons by gender and sexuality
Superscript letters denote column proportions differ from each other at the p < .05 level. Note that 2 participants were missing data on gender
and sexuality, 2 were missing data on the item assessing sharing content related to self-injury, and 1 was missing data on the item assessing talk-
ing about self-injury with people known in person. Given small group sizes, youth who did not identify as male or female were combined into a
single “gender minority” group, for comparison across gender groups. Those who did not identify as heterosexual/straight were combined into a
single “sexual minority” group. Comparisons were repeated excluding those who declined to state their gender or sexual identity, and the pattern
of results remained the same. Total N = 589
* p < .05; **p < .01; ***p < .001
Full Sample
(N = 589)
Male
(n = 206)
Female
(n = 327)
Gender Minority
(n = 54)
Non-Sexual Minority
(n = 285)
Sexual Minority
(n = 300)
n (%) n (%) n (%) n (%) χ2n (%) n (%) χ2
Shared content related to
self-injury
85 (14.5) 20 (9.8)a49 (15.0)a16 (30.2)b14.28** 31 (10.9) 54 (18.0) 5.97*
Viewed content related
to self-injury
156 (26.5) 38 (18.4)a98 (30.0)b20 (37.0)b11.94** 60 (21.1) 96 (31.8) 8.66**
Talked about self-injury
with people known
only online
99 (16.8) 23 (11.2)a53 (16.2)a23 (42.6)b30.37*** 35 (12.3) 64 (21.2) 8.31**
Talked about self-injury
with people known in
person
187 (31.8) 40 (19.5)a120 (36.7)b26 (48.1)b24.57*** 78 (27.5) 108 (35.8) 4.65*
Any online self-injury
activity
254 (43.3) 58 (28.3)a160 (48.9)b35 (66.0)c34.20*** 106 (37.3) 147 (48.8) 7.89**

Research on Child and Adolescent Psychopathology
1 3
self-injury activities using video sharing sites like YouTube (n =
46, 18.1%) or websites specifically for people who injure them-
selves (n = 25, 9.8%).
Controlling for demographic variables (gender, sexuality,
age, race, and ethnicity) and for all online self-injury activities,
two activities were significantly associated with increased sui-
cidal ideation severity: viewing content related to self-injury and
using technology to talk to “real life” friends about self-injury
(Table3). In addition, sharing content related to self-injury,
viewing content related to self-injury, and using technology to
talk about self-injury with “real life” friends were significantly
associated with history of NSSI. Finally, talking about self-injury
with people known only online was significantly associated with
a history of one or more suicide attempts. Those identifying as
female and gender minorities were significantly more likely to
report all SITBs; sexual minority status was significantly associ-
ated with NSSI only. White participants, compared to non-White
participants, were also more likely to report a history of NSSI.
In addition, as described in the data analysis section, all models
were re-run controlling for the two other SITBs examined. The
pattern of results remained the same in all three analyses.
Latent Profile Analysis
An LPA was conducted with all participants who endorsed
having used social media to view or share self-injury related
content (n = 254), and five classes were estimated. As can
be seen in Table4, all information criteria indices became
smaller as the number of classes increased. The BLRT indi-
cated significance for all−class solutions, and therefore did
not aid in class selection. The optimal model determined
empirically was the four-class model. However, upon
inspecting classes and weighting interpretability, it was
determined that a three-class solution was superior to a four-
class solution. Indeed, the four-class solution subdivided one
of the classes in the three-class solution; this subdivision
created somewhat redundant classes and was less interpret-
able. Thus, when considering empirical comparison and
when weighing interpretability, a three-class solution was
chosen (see Fig.1 for class profiles).
Class 1. The first class may be regarded as a “Low Func-
tion Endorsement Class” (n = 84, 33.2% of the sample).
This class reported moderate to low engagement in online
self-injury activities for affect regulation and boredom
Table 2 Websites and apps used to engage in online self-injury activ-
ities
Sample is limited to the 254 adolescents who reported engaging in
any online self-injury activities (i.e., talking about, sharing, or look-
ing at posts/photos related injuring oneself)
n%
Social networking site (e.g., Snapchat, Instagram,
Twitter, Facebook)
190 74.8
Video sharing site (e.g., YouTube) 46 18.1
Chat or discussion forum/website specifically for
people who injure themselves
25 9.8
Text messaging or messaging apps (e.g., WhatsApp) 86 33.9
Other websites or apps 23 9.1
Table 3 Linear and bivariate logistic regression models predicting self-injurious ideation and behaviors from online self-injury activities
NSSI nonsuicidal self-injury. For linear regression predicting suicidal ideation, total R2 = 0.19, F(582) = 17.07, p < .001. Reference groups for
categorical demographic variables were: boys, non-sexual minority youth, other races (i.e., non-White, non-Black), and non-Hispanic. Total N
for all models = 582. Number of participants who endorsed a history of NSSI was 345 (58.7%); number of participants who endorsed a history
of suicide attempt(s) was 311 (52.9%). SIQ: M(SD) = 34.59 (27.2), Skewness = 0.22, Kurtosis = -1.33, Range 0–90. Note that each model was re-
run controlling for the other two SITBs examined, and the pattern of results remained the same
* p < .05; **p < .01; ***p < .001
Suicidal Ideation (SIQ) History of NSSI History of Suicide Attempt
B (95% CI)OR (95% CI)OR (95% CI)
Covariates
Age 0.27 (-0.83, 1.38) 1.04 (0.94, 1.16) 1.05 (0.95, 1.15)
Female 12.92 (8.30, 17.53)*** 4.92 (3.19, 7.58)*** 2.69 (1.81, 3.98)***
Gender Minority 19.18 (10.96, 27.40)*** 3.37 (1.52, 7.47)** 3.22 (1.45, 6.64)**
Sexual Minority 3.62 (-0.80, 8.03) 1.56 (1.03, 2.39)*1.09 (0.75, 1.57)
White 2.33 (-3.25, 7.91) 1.92 (1.23, 3.27)* 0.97 (0.60, 1.57)
Black -1.72 (-8.85, 5.40) 0.78 (0.40, 1.52) 1.09 (0.60, 1.99)
Hispanic -4.82 (-9.78, 0.13) 1.01 (0.63, 1.60) 1.03 (0.67, 1.56)
Online Self-Injury Activities
Shared content related to self-injury 2.98 (-3.65, 9.61) 2.49 (1.17, 5.31)*0.79 (0.44, 1.41)
Viewed content related to self-injury 7.10 (1.56, 12.64)*1.76 (1.00, 3.09) 1.24 (0.77, 2.01)
Talked about self-injury with people known only online 4.61 (-1.94, 11.16) 0.87 (0.44, 1.72) 2.04 (1.14, 3.67)*
Talked about self-injury with people known in person 7.73 (2.47, 12.98)** 2.63 (1.54, 4.48)*** 1.25 (0.79, 1.96)

Research on Child and Adolescent Psychopathology
1 3
reduction purposes (0.50 – 0.62 SD below in the mean).
They also reported well below average identification with all
other functions of self-injury activities, including isolation
reduction (0.93 SD below the mean), self-expression (0.83
SD below the mean), recovery (0.98 SD below the mean),
and identity exploration (0.86 – 0.92 SD below the mean).
Class 2. The second class (n = 127, 50.2% of the sample)
reported average likelihood of engagement in online self-injury
activities for purposes of affect regulation and boredom reduction
(0.11 – 0.22 SD above the mean). They also reported average to
moderately above average levels of endorsement of isolation reduc-
tion (0.31 SD above the mean), self-expression (0.23 SD above the
mean), recovery (0.21 SD above the mean), and identity explora-
tion (0.08 – 0.09 SD above the mean) functions. This class may be
regarded as a “Moderate Function Endorsement Class.”
Class 3. The third class (n = 42, 16.6% of the sample)
included participants who endorsed engaging in online self-
injury activities for affect regulation and boredom reduction at
slightly above average levels (0.51 – 0.61 SD above the mean).
They identified with isolation reduction (0.85 SD above the
mean) and self-expression (0.88 SD above the mean) functions
at levels moderately above the mean. However, this group is
noteworthy for very high endorsement of functions of online
self-injury activities related to recovery (1.24 SD above the
mean) and, even more so, identity exploration (1.37 – 1.45
SD above the mean). Thus, this group may be regarded as the
“Identity and Recovery Functions Endorsement Class.”
Table 4 Model fit statistics for
latent class models estimated up
to 5 classes
LLLog Likelihood, AICAkaike Information Criteria, BIC Bayesian Information Criterion, aBICsample
size adjusted BIC, BLRTBootstrap Likelihood Ratio Test. Note that for the 5-class solution, the number
of random starts was increased to 200 to allow for model convergence (replication of the best loglikeli-
hood value); convergence difficulties are sometimes an indication that the data do not support this number
of classes. One participant was missing data on all 7 items, and thus the total sample size for latent profile
analyses was n = 253
Class LL AIC BIC aBIC BLRT Entropy
2 -2507.26 4587.05 4664.78 4595.05 0.000 0.848
3 -2271.52 4438.50 4544.51 4449.40 0.000 0.851
4 -2189.25 4379.09 4513.36 4392.90 0.000 0.898
5 -2147.35 4304.21 4466.75 4320.92 0.000 0.971
-1.5
-1
-0.5
0
0.5
1
1.5
2
Item 1Item 2Item 3Item 4Item 5Item 6Item 7
Standardized Means
Online Self-Injury Functions Indicators
Class 1
Class 2
Class 3
Fig. 1 Class profiles for the final 3-class model (total n = 253). Indi-
cators represent the following self-reported functions of online self-
injury activities: Negative Affect Regulation: Item 1: I am more likely
to do these activities when I am feeling a negative emotion (e.g.,
upset, sad, angry). Boredom Reduction: Item 2: I am more likely to
do these activities when I am feeling neutral or bored. Reduced Isola-
tion: Item 3: I do these activities in order to feel less isolated or alone.
Self-Expression: Item 4: I do these activities to “vent” or express
how I feel. Recovery: Item 5: I do these activities to help me try and
get better. Identity Exploration: Item 6: Doing these activities helps
me feel like I’m part of a group of other people like me and Item 7:
Doing these activities allows me to express who I really am

Research on Child and Adolescent Psychopathology
1 3
Comparisons Across Classes: Social Media Activities,
Perceived Outcomes, andSITBs
When considering the likelihood of engagement in specific
social media activities, youth in Class 3, as compared to
Class 1, were more likely to report having talked about self-
injury with others met online and to have viewed others’
content related to self-injury online (see Table5). In addi-
tion, compared to youth in Class 1, those in Classes 2 and 3
were more likely to report having shared content (e.g., posts,
messages, comments, photos, videos) related to self-injury
online. No significant class differences emerged regarding
participants’ likelihood of using technology to talk about
self-injury with “real life” friends.
In terms of perceived consequences, compared to Class
1, youth in Classes 2 and 3 indicated higher levels of agree-
ment with statements indicating that these online self-injury
activities led to negative consequences. In particular, par-
ticipants in Classes 2 and 3 were significantly more likely to
report that engaging in these activities led to the normaliza-
tion of self-injury, feelings of thwarted recovery, self-injury
related social comparison, and triggering of self-injurious
behaviors. Classes 2 and 3 were also significantly more
likely to report that engaging in these activities resulted in
them discovering new methods of self-injury.
Classes were also compared in regard to participants’ life-
time history of SITBs. Youth in Class 3 were significantly
more likely to have a history of NSSI compared to those in
Class 1. Compared to those in Class 1, youth in Classes 2
and 3 reported higher suicidal ideation severity. However,
no differences emerged between classes in terms of likeli-
hood of one or more prior suicide attempts. In addition, no
differences emerged between classes by age, gender identity,
or sexual orientation.
Discussion
Very little work has examined online experiences that
specifically involve self-injury, such as posting about self-
injury, viewing self-injury related content, or communicat-
ing online about self-injury. Almost no studies have explored
these experiences among youth with clinically severe mental
illness (i.e., those who have been admitted to a psychiatric
inpatient unit). This represents a critical gap in the literature,
given that youth with psychiatric disorders may be more
Table 5 Online self-injury activities, perceived consequences, and SITBs across classes
Prob. = conditional probability of engagement in this behavior or belonging in this group, given latent class membership; NSSI nonsuicidal self-
injury. χ2 represent Wald Chi-Square values for overall tests of association. A single test was conducted to compare probabilities for all three
gender groups (female, male, gender minority) across three classes. Superscript letters denote column proportions differ from each other at the
p < .05 level. Total n = 253. ; **p < .01; ***p < .001; ****p < .07
Class 1 Class 2 Class 3
Online Self-Injury Activities Prob. (SE) /
M (SE) Prob. (SE) /
M (SE) Prob. (SE)) /
M (SE) χ2
Shared content related to self-injury 0.16 (0.05)a0.41 (0.09)b0.47 (0.11)b16.26***
Viewed content related to self-injury 0.51 (0.06)a0.63 (0.09)a,b 0.76 (0.11)b7.11*
Talked about self-injury with online friends 0.28 (0.05)a0.42 (0.06)a,b 0.53 (0.08)b7.48*
Talked about self-injury with “real life” friends 0.64 (0.06) 0.78 (0.04) 0.81 (0.07) 4.93
Perceived Consequences of Online Self-Injury Activities
Behavioral triggering 1.09 (0.12)a1.65 (0.11)b1.67 (0.19)b13.04**
Normalization 0.88 (0.11)a1.77 (0.11)b1.74 (0.19)b38.87***
Thwarted recovery 0.88 (0.11)a1.60 (0.10)b1.80 (0.18)b30.80***
Social comparison 0.81 (0.11)a1.84 (0.11)b2.24 (0.18)b61.89***
Discovery of new methods 0.56 (0.09)a1.50 (0.12)b1.42 (0.20)b45.86***
Self-Injurious Thoughts and Behaviors
History of NSSI 0.72 (0.05)a0.76 (0.05)a,b 0.89 (0.05)b5.46****
Suicidal Ideation 39.21 (2.75)a48.42 (2.09)b49.48 (3.52)b8.43*
History of Suicide Attempt 0.66 (0.06) 0.67 (0.04) 0.54 (0.09) 1.77
Demographics
Age 15.33 (0.18) 15.13 (0.15) 14.69 (0.26) 0.42
Female 0.71 (0.05) 0.60 (0.05) 0.59 (0.08) 0.53
Gender Minority 0.10 (0.04) 0.14 (0.04) 0.19 (0.07) 0.53
Sexual Minority 0.53 (0.06) 0.53 (0.05) 0.62 (0.09) 0.83

Research on Child and Adolescent Psychopathology
1 3
likely to engage in, and be susceptible to, negative or risky
social media experiences, but also may be poised to benefit
most from supportive online interactions. Results show that
there is a range in the types of online self-injury activities
in which youth engage, as well as in the functions that these
activities serve. Each of these factors may be critical for
evaluating which youth are most at risk in regard to their
online behavior, as well as who might benefit from interven-
tion focused on improving online interactions.
Almost half of the psychiatrically hospitalized adoles-
cents surveyed had engaged in some kind of online self-
injury activity. Thus, a meaningful proportion of adolescents
in this high-risk sample are engaging in these activities. The
majority of adolescents reported engaging in online self-
injury activities using social networking sites like Snap-
chat and Instagram. Very few endorsed engaging in these
activities using chat websites or forums dedicated to persons
who self-injure. Prior work has often focused on qualitative
analysis of these forums (see Lewis and Seko2016 for a
review). Our results underscore the need to examine self-
injury related activities in the online spaces most relevant
to adolescents, as some studies of Instagram have begun to
do (e.g., Arendt etal.2019).
Online Self‑Injury Activities amongSexual
andGender Minority Youth
Online self-injury activities were especially common in
SGM adolescents. Among gender minority youth, 30.2% had
shared their own content related to self-injury, and almost
half (42.6%) had talked about self-injury with individuals
known only online. Among sexual minority youth, nearly
half (48.8%) had engaged in a least some online self-injury
activities. It may be that online environments provide a
low risk context for SGM adolescents to seek support from
peers. Prior work suggests that SGM youth are more likely
to have online-only friendships than cisgender and hetero-
sexual youth (Ybarra etal.2015), likely in part due to in-
person social contexts that may be unsupportive or discrimi-
natory. By engaging online, youth may reduce demands to
respond immediately and minimize their exposure to poten-
tial negative response to disclosures of self-injury from
in-person friends or family. Alternatively, in line with the
social compensation hypothesis (Selfhout etal.2009), it is
possible that SGM adolescents who self-injure may have
poorer quality friendships than those who do not self-injure,
leading to fewer opportunities for in-person interactions and
necessitating online interaction. These findings highlight the
unique needs and possible mechanisms underlying the well-
established heightened risk for SITBs in SGM populations
(O’Brien etal.2016).
Interpersonal Factors inOnline Activities andSITB
Risk
Talking about self-injury online with ‘real life’ friends
and viewing content related to self-injury were significant
predictors of suicidal ideation and history of NSSI in this
sample of high-risk adolescents, even after controlling for
gender, sexuality, and age. These findings are consistent with
the Interpersonal Theory of Suicide, i.e., those who attempt
suicide may have an acquired capability over time (Joiner
2005; Van Orden etal2010). From a theoretical perspec-
tive, these online activities may serve as another means of
gaining exposure to suicidal behaviors and reducing fear of
self-harm. Indeed, recent longitudinal evidence suggests
that exposure to self-injury content on social media predicts
increased risk for SITBs (Arendt etal.2019). Findings also
suggest that sharing self-injury content was specifically
associated with history of NSSI. It has been found that not
only does sharing wound images on social media generate
twice as many comments from other users than non-wound
images, but also that as wound severity increases, so do the
number of comments (Brown etal.2018). It is therefore pos-
sible that sharing self-injury related content may reinforce
self-injury engagement, which may account for the current
findings. However, given that the directionality of effects
is not known, it is also possible that youth with histories of
NSSI and suicidal ideation are simply more likely to engage
in each of these activities. Further work will be needed to
explore these possibilities, and to investigate the relation-
ship between sharing content related to self-injury and the
frequency and severity of future NSSI engagement.
Interestingly, the only activity associated with a history of
suicide attempts was talking about self-injury with ‘online
friends’, highlighting a potential unique risk associated
with this activity. Interpersonal theories of suicidal behav-
ior highlight the role of thwarted belongingness, including
social isolation and loneliness, in increasing risk for sui-
cide (Joiner 2005; Van Orden etal2010). Adolescents who
have made a suicide attempt may have fewer offline friend-
ships (i.e., thwarted belongingness), leading them to rely on
online friends for support around self-injury. Given that both
selection and socialization effects have been documented in
regard to adolescents’ self-injurious behavior (e.g., Prinstein
etal.2010), these online friends may also be more likely to
engage in self-injurious and suicidal behavior themselves,
as well as to reinforce such behavior in their friends. How-
ever, longitudinal studies testing mechanisms of contagion
in online adolescent friendships would be needed to evaluate
these hypotheses. Regardless, these findings suggest that it
is important to assess specific social media behaviors, rather
than overall ‘social media use’, to gain a more nuanced
understanding of risk, and ultimately to be able to inform
clinical efforts.

Research on Child and Adolescent Psychopathology
1 3
Adolescent Identity Development andOnline
Activities
Three distinct subgroups emerged in regard to the reported
functions of engaging in online self-injury activities. One
group reported low levels of endorsement of all functions,
one reported moderate levels of endorsement, and one high
levels of endorsement. Notably, however, the high function
endorsement class reported significantly higher levels of two
functions in particular: recovery (or trying to get better) and
identity exploration (being part of a group of others like me
and expressing who I really am). There may be a subset of
youth for whom posting about, viewing, and discussing self-
injury online represents a facet of identity and a means of
connecting with similar others and/or seeking resources and
support. Adolescents in this class were at heightened risk
for NSSI and suicidal ideation, and endorsed high levels of
potential negative consequences of online self-injury activi-
ties, including thwarting recovery, normalizing self-injury,
social comparison, learning new self-injury methods, and
triggering engagement in self-injury. However, this class
is notable due to the fact that despite these negative con-
sequences, they reported significant positive beliefs about
these activities – that they were aiding their recovery and
serving an important component of their identities.
These findings may be understood within the context of
developmental models of adolescent identity development
(e.g., Christie and Viner2005; Erikson1968, 1980). Adoles-
cence represents a critical period for the formation of a cohe-
sive self-identity (Erikson1968). This process is informed
by the navigation of other developmental tasks, including
the establishment of intimate peer relationships and build-
ing a sense of autonomy (e.g., Christie and Viner2005). On
social media, the ability to explore various self-presentations
and to connect with similar others may be heightened (Nesi
etal.2018). However, for some youth, the online environ-
ment may create unique challenges in regard to these tra-
ditional developmental processes. When vulnerable youth,
such as those at high risk for SITBs, seek to connect with
others and explore burgeoning identities online, they are
often doing so without clinical guidance or supervision.
When these online processes involve self-injury activities,
risk for SITBs may increase and recovery efforts may be
hindered. This has implications for intervention with this
group, including the importance of supporting other means
of identity exploration and help-seeking, and noting con-
tradiction between perceived and actual benefits of these
online activities.
Limitations andFuture Directions
This study had a number of strengths, including the use
of a large clinical sample and focus on a burgeoning topic
in adolescent mental health with limited research. Future
research should build on limitations of this work. First, anal-
yses were based solely on self-report measures and were
cross-sectional in nature. Use of real-time monitoring and
ecological momentary assessment methods may provide
more detailed and objective measures of online self-injury
activities, and feelings/functions perceived in the moment
while engaging in these behaviors. Relatedly, this study
lacked continuous assessment of frequency, severity, and
recency of SITBs, which are likely related to class member-
ship and engagement in specific online activities, and thus
will be important to measure in future research. Such con-
tinuous assessment could also discern nuances in the rela-
tionships between these constructs, such as examining asso-
ciations between frequency, recency, and amount of online
self-injury activities with SITBs and functions. In addition,
although the use of latent profile analysis has limitations,
these results add important initial insight into adolescents’
motivations for engaging in online self-injury activities and
associated risk for SITBs. Future research should explore
these possibilities using in-depth, longitudinal investiga-
tions of individual profiles of online self-injury activities.
Longitudinal studies are needed to inform directional and bi-
directional relationships between online self-injury activities
and SITBs. In addition, future studies should examine the
degree to which the functions and perceived consequences
of online self-injury activities differ from those of actual
engagement in SITBs.
Given efforts to balance detailed data collection with fea-
sibility and ease of administration within the context of usual
care in an inpatient psychiatric setting, the created measure
used may not have assessed the full spectrum of motivations
for, or consequences of, engagement in online self-injury
activities. In addition, each of the online self-injury func-
tions was assessed using only one or two items. Thus, future
efforts should seek to empirically develop and validate a
comprehensive measure of online self-injury activities and
functions, providing measures of concurrent and discrimi-
nant validity and reliability. In addition, such a measure
should include multiple items per function and perceived
consequence to allow for factor analyses. Furthermore, given
the small sizes of subgroups within the groups of youth iden-
tifying as SGM, these subgroups were collapsed for analy-
ses. Future work should examine, for example, potential dif-
ferences between youth identifying as bisexual versus those
identifying as gay or lesbian in regard to online activities and
SITB risk (see Thoma etal.2019). Finally, given that only
adolescent report was collected, it is unknown how peers or
larger social networks respond to online self-injury activities
(e.g., via reinforcement, ignoring/isolation, or connecting
to resources and/or services). Thus, future research should
incorporate a multi-method, multi-informant approach to
capture the broad impact online self-injury activities may

Research on Child and Adolescent Psychopathology
1 3
have on relationships, psychosocial functioning, and psy-
chopathology from various perspectives.
Clinical Implications
Findings suggest that thorough assessment of online activi-
ties, particularly among psychiatrically impaired youth, may
be an important component of clinical practice in the digital
age. Such information may identify those at high risk for
SITBs based on their online behavior, and those who there-
fore may require more intensive monitoring and interven-
tion. Psychoeducation and treatments could be individually
tailored based on identified functions of adolescents’ online
self-injury activities. For instance, adolescents who com-
municate online about their self-injury for affect regula-
tion could learn alternative emotion regulation strategies,
whereas adolescents who do so to reduce social isolation
could learn online and offline social skills to improve the
quality of social interactions and friendships. In addition,
caregiver monitoring of online activity for self-injury related
content may be particularly important for this high-risk
group. Finally, given that rates of online self-injury activi-
ties were greatest among SGM youth, self-injury related
online activities may serve as a potential mechanism under-
lying increased risk for SITBs among SGM youth, which
could be examined and targeted in future intervention work.
Findings highlight the array of online self-injury activities
and potential functions of these behaviors in psychiatrically
impaired youth, associations with risk for SITBs, and the
need to further research into these online behaviors among
vulnerable adolescents.
Funding Dr. Nesi was supported in part by the American Foundation
for Suicide Prevention (PDF-010517) and National Institute of Mental
Health (K23MH122669). Dr. Burke was supported by NIMH T32 grant
MH019927. Dr. Lawrence was supported by the American Founda-
tion for Suicide Prevention (PDF-009519).Any opinions, findings, and
conclusions or recommendations expressed in this material are solely
the responsibility of the authors and do not necessarily represent the
views of AFSP or NIMH.
Compliance with Ethical Standards
Conflicts of Interest The authors have no financial relationships or
conflicts of interest related to this study to disclose.
Ethical Approval The Lifespan Hospitals Institutional Review Board
(IRB) approved this research.
Availability of Data and Material The measures administered for this
study were part of a standardized admission process for all adolescents
admitted to a psychiatric inpatient unit. This study consists of a retro-
spective chart review, and thus the IRB approved a waiver of informed
consent. Given the sensitive nature of such data, the authors will handle
requests for data access on a case-by-case basis.
References
Anderson, M., & Jiang, J. (2018). Teens, social media & technology.
Pew Research Center
Arendt, F., Scherr, S., Romer, D., etal. (2019). Effects of exposure
to self-harm on social media: Evidence from a two-wave panel
study among young adults. New Media and Society, 21(11–12),
2422–2442. https ://doi.org/10.1177/14614 44819 85010 6
Asparouhov, T., & Muthén, B.O. (2014). Auxiliary variables in mixture
modeling: Using the BCH method in Mplus to estimate a distal
outcome model and an arbitrary second model. Mplus Web Notes.
Bandura, A. (2001). Social Cognitive Theory of Mass Communication.
In Media Psychology (Vol. 3, Issue 3, pp. 265–299). Routledge.
https ://doi.org/10.1207/S1532 785XM EP030 3_03
Bell, J., Mok, K., Gardiner, E., Pirkis, J., etal. (2018). Suicide-Related
Internet Use Among Suicidal Young People in the UK: Charac-
teristics of Users, Effects of Use, and Barriers to Offline Help-
Seeking. Archives of Suicide Research, 22(2), 263–277. https ://
doi.org/10.1080/13811 118.2017.13346 09
Branley, D.B., & Covey, J. (2017). Is exposure to online content
depicting risky behavior related to viewers’ own risky behavior
offline? Computers in Human Behavior, 75, 283–287. https ://doi.
org/10.1016/j.chb.2017.05.023
Brown, R.C., Fischer, T., Goldwich, A.D., Keller, F., Young, R., Ple-
ner, P.L., etal. (2018). #cutting: Non-suicidal self-injury (NSSI)
on Instagram. Psychological Medicine. https ://doi.org/10.1017/
S0033 29171 70017 51
Carlyle, K.E., Guidry, J.P.D., Williams, K., Tabaac, A., Perrin, P.B.,
etal. (2018). Suicide conversations on InstagramTM: contagion
or caring? Journal of Communication in Healthcare. https ://doi.
org/10.1080/17538 068.2018.14365 00
Christie, D., & Viner, R. (2005). Adolescent development. BMJ,
330(7486), 301. https ://doi.org/10.1136/bmj.330.7486.301
De Choudhury, M., & Kiciman, E. (2017). The language of social
support in social media and its effect on suicidal ideation risk.
Proceedings of the 11th International Conference on Web and
Social Media, ICWSM 2017.
Dyson, M.P., Hartling, L., Shulhan, J., Chisholm, A., Milne, A.,
Sundar, P., etal. (2016). A systematic review of social media use
to discuss and view deliberate self-harm acts. PLoS ONE. https ://
doi.org/10.1371/journ al.pone.01558 13
Erikson, E.H. (1968). Identity: Youth and Crisis. W W Norton & Co.
Erikson, E.H. (1980). Identity and the life cycle. W W Norton & Co.
Foulkes, L., & Blakemore, S.J. (2016). Is there heightened sensitivity
to social reward in adolescence? In Current Opinion in Neurobiol-
ogy (Vol. 40, pp. 81–85). Elsevier Ltd. https ://doi.org/10.1016/j.
conb.2016.06.016
Frost, M., & Casey, L. (2016). Who Seeks Help Online for Self-
Injury? Archives of Suicide Research, 20(1), 69–79. https ://doi.
org/10.1080/13811 118.2015.10044 70
Gould, M., Jamieson, P., Romer, D., etal. (2003). Media contagion and
suicide among the young. American Behavioral Scientist. https ://
doi.org/10.1177/00027 64202 25067 0
Hawton, K., Hill, N.T.M., Gould, M., John, A., Lascelles, K.,Robinson,
J., etal. (2020). Clustering of suicides in children and adolescents. In
The Lancet Child and Adolescent Health (Vol. 4, Issue 1, pp. 58–67).
Elsevier B.V. https ://doi.org/10.1016/S2352 -4642(19)30335 -9
Heilbron, N., & Prinstein, M.J. (2008). Peer influence and adolescent
nonsuicidal self-injury: A theoretical review of mechanisms and
moderators. Applied and Preventive Psychology. https ://doi.
org/10.1016/j.appsy .2008.05.004
Insel, B.J., & Gould, M.S. (2008). Impact of Modeling on Adolescent
Suicidal Behavior. Psychiatric Clinics of North America. https ://
doi.org/10.1016/j.psc.2008.01.007

Research on Child and Adolescent Psychopathology
1 3
John, A., Glendenning, A.C., Marchant, A., Montgomery, P., Stewart,
A., Wood, S., etal. (2018). Self-harm, suicidal behaviours, and
cyberbullying in children and young people: Systematic review.
Journal of Medical Internet Research. https ://doi.org/10.2196/
jmir.9044
Johns, M.M., Lowry, R., Andrzejewski, J., Barrios, L.C., Demissie,
Z., McManus, T., etal. (2019). Transgender identity and experi-
ences of violence victimization, substance use, suicide risk, and
sexual risk behaviors among high school students. Morbidity and
Mortality Weekly Report, 68(3), 67.
Joiner, T.E. (2005). Why people die by suicide. Cambridge, MA: Har-
vard University Press.
Kann, L., McManus, T., Harris, W.A., Shanklin, S.L., Flint, K.H.,
Queen, B., Lowry, R., Chyen, D., Whittle, L., Thornton, J., Lim,
C., Bradford, D., Yamakawa, Y., Leon, M., Brener, N.,Ethier,
K.A., etal. (2018). Youth risk behavior surveillance - United
States, 2017. MMWR Surveillance Summaries. https ://doi.
org/10.15585 /mmwr.ss670
King, C.A., & Merchant, C.R. (2008). Social and interpersonal factors
relating to adolescent suicidality: A review of the literature. Archives
of Suicide Research. https ://doi.org/10.1080/13811 11080 21012 03
Lanza, S.T., Tan, X., Bray, B.C., etal. (2013). Latent Class Analy-
sis With Distal Outcomes: A Flexible Model-Based Approach.
Structural Equation Modeling. https ://doi.org/10.1080/10705 511.
2013.74237 7
Lewis, S.P., & Seko, Y. (2016). A Double-Edged Sword: A Review
of Benefits and Risks of Online Nonsuicidal Self-Injury Activi-
ties. Journal of Clinical Psychology. https ://doi.org/10.1002/
jclp.22242
Mok, K., Jorm, A.F., Pirkis, J., etal. (2016). The perceived impact of
suicide-related internet use: A survey of young Australians who
have gone online for suicide-related reasons. DIGITAL HEALTH.
https ://doi.org/10.1177/20552 07616 62986 2
Moreno, M., & Kota, R. (2013). Social media.Children, adolescents,
and the media, 435–456.
Muehlenkamp, J.J., Claes, L., Havertape, L., Plener, P.L., etal. (2012).
International prevalence of adolescent non-suicidal self-injury and
deliberate self-harm. Child and Adolescent Psychiatry and Mental
Health. https ://doi.org/10.1186/1753-2000-6-10
Muthén, L., & Muthén, B. BO 1998–2018.Mplus User’s Guide.
Nesi, J., Choukas-Bradley, S., Prinstein, M.J., etal. (2018). Transfor-
mation of Adolescent Peer Relations in the Social Media Con-
text: Part 1—A Theoretical Framework and Application to Dyadic
Peer Relationships. Clinical Child and Family Psychology Review.
https ://doi.org/10.1007/s1056 7-018-0261-x
Nesi, J., & Prinstein, M.J. (2015). Using Social Media for Social Com-
parison and Feedback-Seeking: Gender and Popularity Moderate
Associations with Depressive Symptoms. Journal of Abnormal
Child Psychology. https ://doi.org/10.1007/s1080 2-015-0020-0
Niederkrotenthaler, T., Haider, A., Till, B., Mok, K., Pirkis, J., etal.
(2017). Comparison of suicidal people who use the internet for
suicide- related reasons and those who do not: Survey study in
Austria. Crisis, 38(2), 131–135. https ://doi.org/10.1027/0227-
5910/a0004 32
Niederkrotenthaler, T., Reidenberg, D.J., Till, B., Gould, M.S., etal.
(2014). Increasing help-seeking and referrals for individu-
als at risk for suicide by decreasing stigma: The role of mass
media. American Journal of Preventive Medicine. https ://doi.
org/10.1016/j.amepr e.2014.06.010
Niederkrotenthaler, T., & Stack, S. (Eds.). (2017).Media and Suicide:
International Perspectives on Research, Theory, and Policy.
Routledge.
Nock, M.K., Holmberg, E.B., Photos, V.I., Michel, B.D., etal. (2007).
Self-Injurious Thoughts and Behaviors Interview: Development,
Reliability, and Validity in an Adolescent Sample. Psychological
Assessment. https ://doi.org/10.1037/1040-3590.19.3.309
Nylund, K.L., Asparouhov, T., Muthén, B.O., etal. (2007). Deciding on
the number of classes in latent class analysis and growth mixture
modeling: A Monte Carlo simulation study. Structural Equation
Modeling. https ://doi.org/10.1080/10705 51070 15753 96
O’Brien, K.H.M.M., Putney, J.M., Hebert, N.W., Falk, A.M., Agui-
naldo, L.D., etal. (2016). Sexual and Gender Minority Youth
Suicide: Understanding Subgroup Differences to Inform Inter-
ventions. LGBT Health. https ://doi.org/10.1089/lgbt.2016.0031
Peters, J.R., Mereish, E.H., Krek, M.A., Chuong, A., Ranney, M.L.,
Solomon, J., etal. (2020). Sexual orientation differences in non-
suicidal self-injury, suicidality, and psychosocial factors among an
inpatient psychiatric sample of adolescents. Psychiatry Research,
284, 112664. https ://doi.org/10.1016/j.psych res.2019.11266 4
Prinstein, M.J., Heilbron, N., Guerry, J.D., Franklin, J.C., Rancourt, D.,
Simon, V., etal. (2010). Peer influence and nonsuicidal self injury:
Longitudinal results in community and clinically-referred adoles-
cent samples. Journal of Abnormal Child Psychology. https ://doi.
org/10.1007/s1080 2-010-9423-0
Reynolds, W.M., & Mazza, J.J. (1999). Assessment of suicidal idea-
tion in inner-city children and young adolescents: Reliability and
validity of the suicidal ideation questionnaire-JR. School Psychol-
ogy Review.
Rideout, V., & Robb, M.B. (2018). Social life media: 2018 teens
reveal their experiences. Common Sense Media. https ://doi.
org/10.1016/j.ijhea tmass trans fer.2016.02.015
Selfhout, M.H.W., Branje, S.J.T., Delsing, M., ter Bogt, T.F.M., Meeus,
W.H.J., etal. (2009). Different types of Internet use, depression,
and social anxiety: The role of perceived friendship quality. Jour-
nal of Adolescence, 32(4), 819–833. https ://doi.org/10.1016/j.
adole scenc e.2008.10.011
Smith, P.N., Cukrowicz, K.C., Poindexter, E.K., Hobson, V., Cohen,
L.M., etal. (2010). The acquired capability for suicide: A com-
parison of suicide attempters, suicide ideators, and non-suicidal
controls. Depression and Anxiety, 27(9), 871–877. https ://doi.
org/10.1002/da.20701
Thoma, B.C., Salk, R.H., Choukas-Bradley, S., Goldstein, T.R., Levine,
M.D., Marshal, M.P., etal. (2019). Suicidality disparities between
transgender and cisgender adolescents. Pediatrics, 144(5), 1–9.
https ://doi.org/10.1542/peds.2019-1183
Twenge, J.M., Joiner, T.E., Rogers, M.L., Martin, G.N., etal. (2018).
Increases in Depressive Symptoms, Suicide-Related Outcomes,
and Suicide Rates Among U.S. Adolescents After 2010 and Links
to Increased New Media Screen Time. Clinical Psychological Sci-
ence.https ://doi.org/10.1177/21677 02617 72337 6
Valkenburg, P.M., & Peter, J. (2013). The differential susceptibility to
media effects model. Journal of Communication, 63(2), 221–243.
https ://doi.org/10.1111/jcom.12024
Van Orden, K.A., Witte, T.K., Cukrowicz, K.C., Braithwaite, S.R.,
Selby, E.A., Joiner, T.E., etal. (2010). The Interpersonal The-
ory of Suicide. Psychological Review. https ://doi.org/10.1037/
a0018 697
Ybarra, M.L., Mitchell, K.J., Palmer, N.A., Reisner, S.L., etal. (2015).
Online social support as a buffer against online and offline peer
and sexual victimization among U.S. LGBT and non-LGBT youth.
Child Abuse and Neglect, 39, 123–136. https ://doi.org/10.1016/j.
chiab u.2014.08.006
King, C.A., Hovey, J.D., Brand, E., & Ghaziuddin, N. (1997). Predic-
tion of positive outcomes for adolescent psychiatric inpatients.
Journal of the American Academy of Child & Adolescent Psy-
chiatry, 36(10), 1434-1442.
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.