1985Archives of Sexual Behavior (2019) 48:1983–1992
1 3
“primary” mental health diagnosis, such as autism spectrum
disorder or borderline personality disorder, or as a result of
a severe trauma (e.g., sexual abuse). Another explanation is
that gender dysphoria is inherently distressing, i.e., the marked
incongruence between one’s felt gender and somatic sex—even
within psychosocial milieus that are largely “affirming/support-
ive”—which leads to clinically significant symptoms such as
anxiety or depression. A more common explanation (and one
that is often favored by “gender-affirming” clinicians and theo-
rists) is that the co-occurring mental health issues are simply
secondary to factors such as family rejection or social ostracism
within the peer group vis-à-vis the gender dysphoria (see, e.g.,
Grossman, Park, & Russell, 2016; Janssen & Leibowitz, 2018;
McDermott, Hughes, & Rawlings, 2017).
In this broader context of co-occurring mental health issues,
concern about suicide risk has become a topic of intense focus
in recent years (see, e.g., Tanis, 2016). On the Internet, for
example, one might come across the comment made by some
parents “I would rather have a trans kid than a dead kid” (see,
e.g., Biggs, 2018; Digitale, 2017; “I’d Rather Have a Living
Son Than a Dead Daughter,” 2016) and instances of completed
suicide receive intense media scrutiny (e.g., Bever, 2016; Savva
& Small, 2019). Indeed, Karasic and Ehrensaft (2015) asserted
that completed suicides are “alarmingly high”—a statement
which, in my view, has no formal and systematic empirical
basis. In fact, I would argue that the statement itself is alarming.
So, what do we know about suicidality among adolescents
with gender dysphoria? In addition to the case report litera-
ture (e.g., Acosta, Qayyum, Turban, & van Schalkwyk, 2019;
Alastanos & Mullen, 2017; Rice etal., 2016), a number of
studies from specialty gender identity clinics have reported on
the percentage of adolescents with gender dysphoria with a his-
tory of suicidal ideation and/or self-harm and suicide attempts.
These studies have relied on clinical chart information, parent’s
report or self-report. For example, in one clinical chart study
(N = 69), Di Ceglie, Freedman, McPherson, and Richardson
(2002) found a history of self-harm and self-injurious behavior
in 23% and 22% of the adolescents, respectively. Subsequent
studies have also provided descriptive data on the percentage
of patients where suicidality (thoughts and behaviors) has been
endorsed, with sample sizes ranging from 34 to 203 (Becker,
Gjergji-Lama, Romer, & Möller, 2014; Holt, Skagerberg, &
Dunsford, 2016; Kaltiala-Heino, Sumia, Työläjärvi, & Lind-
berg, 2015; Khatchadourian, Amed, & Metzger, 2014; Olson,
Schrager, Belzer, Simons, & Clark, 2015; Peterson, Matthews,
Copps-Smith, & Conrad, 2017; Skagerberg, Parkinson, & Car-
michael, 2013; see also Mann, Taylor, Wren, & de Graaf, 2019).
Not surprisingly, it is almost always the case that there were
higher rates for suicidal ideation than for self-harm and/or sui-
cide attempts. There is also a literature on suicidality among
non-clinic-based samples of adolescents with gender dysphoria
or who self-identify as transgender (e.g., Butler etal.,
2019;
Johns etal., 2019; Katz-Wise, Ehrensaft, Vetters, Forcier, &
Austin,
2018; Kidd, Gaetz, & O’Grady, 2017; Perez-Brumer,
Day, Russell, & Hatzenbuehler, 2017; Toomey, Syvertsen, &
Shramko, 2018; Veale, Watson, Peter, & Saewyc, 2017). These
studies also report what would appear to be high rates of sui-
cidal ideation and of self-harm or suicide attempts.
Apart from measurement issues (most of these studies used
fairly crude metrics of suicidality), there are at least two other
methodological issues that deserve some reflection. First, in
general, the clinic-based samples did not employ any type of
comparison group, such as a group of adolescents referred
for any other type of mental health concern or even a non-
referred comparison group. Second, in the non-clinic-based
samples, when a comparison group was used, it was limited
to “cisgender” adolescents, but without taking into account
the mental health status of these youth. For example, Perez-
Brumer etal. (2017) reported that the past 12-month rate of
self-reported suicidal ideation among transgender adolescents
was 33.73% (N = 280) compared to 18.85% of non-transgender
adolescents (N = 25,213) (see Perez-Brumer etal.’s Table1 for
their weighted subsample data). Toomey etal. (2018) reported
a very high rate of self-reported lifetime suicide attempts
among transgender birth-assigned females (50.8%) compared,
for example, to a 17.6% rate among cisgender birth-assigned
females. If one wanted to make the argument that at least some
of the transgender students would meet the criteria for a mental
health diagnosis of gender dysphoria, then one would want
to make a comparison with the cisgender or non-transgender
students who also had one mental health diagnosis (say, for
example, anxiety or depression). This would allow for a more
nuanced comparative analysis to see whether or not suicidal-
ity is higher, similar, or lower among adolescents with gender
dysphoria when compared to some type of clinical comparison
group.
de Graaf etal. (2019) measured suicidal ideation and self-
harm/suicide attempts using two items from the Child Behavior
Checklist (CBCL) or the Youth Self-Report Form (YSR) (Item
91: “Talks about killing self”; Item 18: “Deliberately harms
self or attempts suicide”). Both items were rated on a 0–2-point
scale (“Not true,” “Somewhat or sometimes true,” “Very true”),
with the time frame “now or within the past 6months.” The
sample consisted of adolescents referred for gender dysphoria
from three clinics: Toronto, Amsterdam, and London (total
N = 2065). In addition to between-clinic comparisons, the per-
centage of adolescents in which these two items were endorsed
was compared with the CBCL/YSR referred and non-referred
U.S. standardization samples (Achenbach & Rescorla, 2001).
Although there was, at times, significant between-clinic
variation in the percentage of adolescents for whom these two
items were rated as either a 1 or a 2, the key point that I wish to
make here is that the rate of suicidality was, in general, much
more similar to that of the referred adolescents than to the non-
referred adolescents from the CBCL/YSR standardization sam-
ples. For example, on CBCL Item 91 for birth-assigned females