Olson, K. R., Durwood, L., Horton, R., Gallagher, N. M., & Devor, A. (2022). Gender Identity 5 Years After Social Transition. Pediatrics Les artikkelen (åpen tilgang)

Analyse av artikkelen

SEGM Summary

A recent study published in Pediatrics examined the 5-year gender identity development trajectory of transgender-identified children who underwent early social gender transition (SGT). The children were, on average, 6-7 years old at the time of SGT. Five years later, at the average age of 11-12, almost all—97.5%—continued to identify as transgender, including a small subset (3.5%) developing a non-binary identification. Only 2.5% of the children desisted from transgender identification by the end of the study period, and re-identified with their sex.

The authors concluded that detransition among previously socially gender transitioned youth is rare. A significant proportion of the youth in the study had already initiated interventions with puberty blockers (29%) and cross-sex hormones (31%) by the end of the study, and the authors opined that the remainder would likely initiate medical interventions in the future.This finding is in sharp contrast to earlier research demonstrating that most cases of childhood-onset gender incongruence tend to resolve sometime during adolescence and before reaching mature adulthood. However, the children in the prior research were not socially transitioned, and early social transition had been discouraged by prior protocols.

Some of the recent news coverage of this study incorrectly stated that the study confirmed that children who claim a transgender identity rarely change their minds. This statement is only partially accurate. A more accurate statement is that the study suggests that children who claim a transgender identity and undergo early social transition rarely change their minds, at least into their early teen years. This is because the Trans Youth Project, the source of the study's data, is specifically focused on evaluating the effects of early social gender transition in gender-diverse youth, and social gender transition was a prerequisite for participating in the study.

Below is SEGM’s critical analysis of the study: its strengths, weaknesses, and limitations.

What the Study Got Right

1. The authors are correct that little is known about the trajectories of children who undergo early social gender transition.

There is virtually no prior research into the developmental trajectories of socially-transitioned children, because the practice of pre-pubertal social transition was discouraged by the authors of the Dutch protocol. The Dutch clinicians discouraged it because most gender dysphoric children reidentified with their sex during puberty, making “watchful waiting” – rather than early transitioning – common sense. 

The following quotes from the Dutch clinicians encapsulated the Dutch concern with early SGT:

“As mentioned earlier, symptoms of GID [Gender Identity Disorder] at prepubertal ages decrease or even disappear in a considerable percentage of children (estimates range from 80–95%) [11,13]. Therefore, any intervention in childhood would seem premature and inappropriate." (Cohen-Kettenis et al., 2008, p. 1895)

"Because most gender dysphoric children will not remain gender dysphoric through adolescence (Wallien & Cohen-Kettenis, 2008), we recommend that young children not yet make a complete social transition (different clothing, a different given name, referring to a boy as “her” instead of “him”) before the very early stages of puberty." (de Vries and Cohen-Kettenis, 2012, p. 308)

"In making this recommendation, we aim to prevent youths with nonpersisting gender dysphoria from having to make a complex change back to the role of their natal gender (Steensma & Cohen-Kettenis, 2011). In a qualitative follow-up study, several youths indicated how difficult it was for them to realize that they no longer wanted to live in the role of the other gender and to make this clear to the people around them (Steensma, Biemond, et al., 2011). These children never even officially transitioned but just were considered by everyone around them as belonging to the other (non-natal) gender. One may wonder how difficult it would be for children living already for years in an environment where no one (except for the family) is aware of the child’s natal sex to make a change back." (de Vries and Cohen-Kettenis, 2012, p. 308)

2. The authors are correct in their observation that the children in their study who underwent early social transition appear to be persisting in their transgender identity at very high rates.

Historically, 61%-98% of gender incongruent children desisted from their trans identification before reaching adulthood (a finding confirmed by 11 of the 11 studies that studied this phenomenon). However, the current study suggests that 98% of early-socially-transitioned children persist in their wish to undergo gender transition.

Of note, the 61%-98% (or the 85%) historic desistance statistic has been critiqued as inflated, by those who noted that some of the children diagnosed with a “gender identity disorder” were merely extremely gender-nonconforming.  A reanalysis of these data focused on the subset of the children who met the full diagnostic criteria, and found that 67% of them desisted, with the majority growing up to become gay adults.

The high rate of desistance among gender variant children has been recognized multiple times by the pioneering researchers in the field of pediatric gender medicine, and also by the Endocrine Society’s treatment guidelines:

"With current knowledge, we cannot predict the psychosexual outcome for any specific child. Prospective follow-up studies show that childhood GD/gender incongruence does not invariably persist into adolescence and adulthood (so-called “desisters”). Combining all outcome studies to date, the GD/gender incongruence of a minority of prepubertal children appears to persist in adolescence (20, 40). In adolescence, a significant number of these desisters identify as homosexual or bisexual.” (Hembree et al., 2017, p. 3876)

3. The authors are also correct that most of the early-socially transitioned children in the study proceeded to medical transition, and the rest will likely do so as well.

According to the study, by the end of the 5-year period, at least 60% of the youth had already started to take puberty blockers and/or cross sex hormones (the rate may be even higher since 8% of the participants did not return the questionnaires in the final 2 years of the study—if some of those children started on hormonal interventions, the researchers would not know).  Given the focus that “gender-affirming” care places on the provision of medical interventions, it is probable that many of the remaining 40% of youth in the study will proceed to hormones and surgeries.

The Study's Limitations

1. Lack of equipoise, as evidenced by no discussion of risks

In medical research, “equipoise” requires researchers to approach their research question with genuine uncertainty about the effect of an intervention and is the ethical basis for medical research. The UK's independent review of gender dysphoria healthcare, the Cass Review, recently issued an interim report in which it said:

“Social transition – this may not be thought of as an intervention or treatment, because it is not something that happens within health services. However, it is important to view it as an active intervention because it may have significant effects on the child or young person in terms of their psychological functioning.64,65 There are different views on the benefits versus the harms of early social transition. Whatever position one takes, it is important to acknowledge that it is not a neutral act, and better information is needed about outcomes.” (Cass, 2022, p. 62)

Yet the investigators of this study presume that early social gender transition is largely beneficial, and that the only potential downside is the risk of detransition, which their findings suggest is low. They do not engage with evidence that social gender transition may not confer the claimed benefits (including research that disputes their own earlier findings of benefits). Nor do they consider that persistence increases the risk of undergoing invasive medical and surgical interventions with a lifelong burden of risk and aftercare.

With over 60% of study participants having already commenced hormonal interventions, the researchers should have included a discussion of these risks, including the potential harm to bone health, brain development, impaired sexuality, cardiovascular health, as well as the risks of infertility and sterility, whenever puberty blockers are administered at Tanner stage 2 and are followed by cross-sex hormones.

2. Non-representative sample

The data for this study came from the Trans Youth Project, a convenience sample of parents who opted into family meetings with the researchers face-to-face once every 1-3 years until the child turned 12 and, after that, periodic questionnaires. The research team did not provide diagnosis or treatment. The only benefit to the families was a small financial reward for participation, and the ability to contribute to the knowledge base.

Although it is not certain how the study participants were recruited, the resultant sample had a highly unusual composition. A full 17% of the research participants reported a “multiracial” identity, compared to just 2.8% of the US population. The study participants were also significantly more affluent than the average US family: 35% reported incomes over $125,000, compared to 24% in the US population.

It is unclear how applicable the findings from this demographically skewed sample are to the rest of the US population.

3. Poor applicability to youth diagnosed with gender dysphoria

The researchers observed that most of the families in the study did not believe that the DSM diagnosis of gender dysphoria was either “ethical or useful.” Consistent with this finding, the researchers noted that in many cases the “distress” criterion, necessary for the DSM-5 diagnosis, was not met. This allows for the inclusion of children with a certain parent profile: one who is heavily invested in the idea that their child is transgender even if they don’t meet diagnostic criteria for gender dysphoria. Currently, the diagnosis of gender dysphoria is a necessary condition for determining the medical necessity of interventions. Since this study did not require the diagnosis of gender dysphoria, it is unclear whether its findings are applicable to the population of gender-dysphoric youth. The lack of diagnosis also makes it unclear how this study could be reproduced. 

4. Unknown applicability to youth with adolescent-onset gender dysphoria in particular

It is notable that the ratio of natal boys to girls in this study is approximately 2:1. This predominately male presentation, combined with the early age of social transition (average 6.5 years) suggest that most of the sample came from the population with early-childhood emergence of transgender identity. Prior research using the same sample reported that all youth in the Trans Youth Project had excellent mental health function, with “no elevations in depression and slightly elevated anxiety relative to population averages. They did not differ from the control groups on depression symptoms and had only marginally higher anxiety symptoms."

In contrast, the population of youth desiring hormones currently is dominated by young people whose trans identity emerges for the first time around or shortly after puberty and who suffer from high rates of mental illness or neurocognitive comorbidities. In addition, the majority are natal females. It is not yet known whether the findings of persistence of trans identity among early socially-transitioned youth with early-childhood onset of gender dysphoria and good baseline mental health will apply to adolescents with a post-puberty onset of transgender identity which is further complicated by mental health issues.

5. Non-participation bias

Two sources of non-participation bias in the study are apparent. First, the children in the study had been socially transitioned for 1.5 years prior to enrolling in the study. Thus, it is likely that the study under-represents families where the child had experienced a briefer period of social transition and who then detransitioned. Second, a significant number of families (8%) failed to contact the researchers in the past two years, and an unknown number failed to stay in touch in the final year of the study (2020). These non-respondents may be disproportionately parents of detransitioners.

This bias can negatively impact the reliability of the study in several ways, including underestimating the true rate of detransition, and under-representing the experiences of the children who were socially transitioned for a period and later detransitioned.

6. Problems in reporting

There are several inaccuracies in reporting that impede the interpretation of the study results:

• The study refers to a 5-year follow-up, but subjects were only followed for 3.8 years.  The 5-year estimate comes from the fact that the subjects had already socially transitioned for 1.5 years prior to enrolling. While the authors are correct in stating that the identity outcome occurred 5 years after social transition, they should have clearly stated that they could only validate the progression for 3.8 out of the 5 years.

• Although the researchers disclosed that 8% of the study sample failed to return questionnaires in the 2-year time period 2019-2020, they did not make it clear how many had been expected to respond in 2020 specifically but failed to do so. Children’s identity development is often unpredictable and can change quickly. If non-participation in 2020 is not significantly different from 2019, it is a moot point. However, if it is significantly higher, then combining the two years may mask a much higher rate of non-response in the final year of the study. Non-responses may represent “silent” desistance, whereby parents do not officially withdraw from the study, but simply do not consider the study relevant since their children no longer consider themselves transgender. The title of the project (“Trans Youth Project") and its stated goal to study “transgender children” contribute to the possibility that parents of desisters may not wish to stay engaged in this research project.

• The participant demographics presented in Table 1 contain only three variables: race, annual household income, and geographic location. Other important demographics are omitted, including parents’ educational attainment, marital status, and home ownership. Neither the age distribution of the study participants nor the age at which their trans identity emerged are reported. The children and adolescents’ pattern of sexual attraction/ orientation is not noted. This is an important omission, as gender incongruence in childhood is strongly associated with future homosexuality. If a significant proportion of the youth in the study are attracted to the individuals of their natal sex, it would suggest that early social transition poses risks of iatrogenic harm to LGB youth by exposing them to highly invasive and unnecessary medical interventions.

7. Commingling of interventions and lack of control group

The majority of participants started puberty blockers and/or cross-sex hormones during the study, so it is hard to separate the effects of social transition on persistence of transgender identity from the effects of these medications.  For example, four studies confirm that over 95% of children who start puberty blockers, persist in their trans identification and continue to cross-sex hormones. Thus, taking puberty blockers may be in part responsible for the persistence found by the study. More generally, the lack of a control group makes it hard to interpret which of the interventions are associated with the persistence of trans identity, and the study design precludes the determination of causation or its direction.

8. Lack of long-term follow-up

The study noted high rates of trans identity persistence at the age of 11-12. However, little is known how these adolescents will identify as they move through the later stages of adolescence and reach mature adulthood. All aspects of identity continue to significantly change in adolescence and young adulthood, and gender identity may be subject to similar changes. Moreover, based upon the initial ages of the study participants, many did not reach the age by which they would be likely to understand themselves as lesbian, gay, or bisexual (LGB). LGB individuals frequently go through a period of gender dysphoria in childhood. Many detransitioners also have come to understand themselves as LGB adults following a period of temporary transgender identification in their childhood or adolescence. 

9. Limitation in hypothesis

The authors assert that the main concern with early social gender transition is that the process of re-identifying with their natal sex following a period of social transition may be distressing to a child. Thus, their research question only deals with the rates of persistence and desistence. The authors are correct that researchers ­– including the authors of the original Dutch protocol ­– have observed that some children may find it excessively stressful to detransition, particularly when many adults around them are not even aware of the child's natal sex.

However, this is not the only concern about early social transition. Another key concern, also voiced by the Dutch researchers, is that children who are socially transitioned at an early age, and who end up persisting with their trans identity, lose touch with biological reality, and as a result, may have unrealistic expectations of what “gender-affirming” hormones and surgeries can realistically deliver. This may result in disappointing post-surgical outcomes or inability to participate in the lifelong medical maintenance required to preserve the desired appearance:

"Another reason we recommend against early transitions is that some children who have done so (sometimes as preschoolers) barely realize that they are of the other natal sex. They develop a sense of reality so different from their physical reality that acceptance of the multiple and protracted treatments they will later need is made unnecessarily difficult. Parents, too, who go along with this, often do not realize that they contribute to their child’s lack of awareness of these consequences." (de Vries and Cohen-Kettenis, 2012, p. 308)

The Director of the UK's Gender Identity Development Service (GIDS) has highlighted a closely-related complication of early childhood social transition, that it makes it difficult to obtain consent for later medical transition procedures:

“In the UK, we’re seeing much younger people socially transitioning. But sometimes it then becomes almost impossible for them to think about the reality of their physical body. They are living totally the gender they feel they are, but of course their body doesn’t match that, and it becomes something that can’t be talked about or thought about. Clearly, it then becomes quite difficult in terms of keeping their options open and ensuring fully informed consent for any appropriate physical interventions.” (The Times, 29 August 2015)

Another issue raised by researchers is the possibility that early social gender transition may steer a child toward persistence of a transgender identity, which would otherwise have naturally reverted to be congruent with one’s sex:

“With the emergence in the last 10–15 years of a pre-pubertal gender social transition as a type of psychosocial treatment – initiated by parents on their own (without formal clinical consultation) or with the support/advice of professional input [ref. omitted] – it is not clear if the desistance rates reported in the four core studies will be “replicated” in contemporary samples. Indeed, the data for birth-assigned males in Steensma et al. (2013a) already suggest this: of the 23 birth-assigned males classified as persisters, 10 (43%) had made a partial or complete social transition prior to puberty compared to only 2 (3.6%) of the 56 birth-assigned males classified as desisters. Thus, I would hypothesize that when more follow-up data of children who socially transition prior to puberty become available, the persistence rate will be extremely high. This is not a value judgment – it is simply an empirical prediction." (Zucker, 2018, p. 7)

Similarly, Dutch researcher Dr. Thomas Steensma suggested that:

"A childhood transition has an effect by itself and influences the cognitive gender identity representation of the child and/or their future development" and that this "link between social transitioning and the cognitive representation of the self [would] influence the future rates of persistence." (Steensma et al., 2013, p. 150).

The risks of early social transition are also acknowledged by the Endocrine Society’s treatment guidelines:

“However, the large majority (about 85%) of prepubertal children with a childhood diagnosis did not remain GD/ gender incongruent in adolescence (20). If children have completely socially transitioned, they may have great difficulty in returning to the original gender role upon entering puberty (40). Social transition is associated with the persistence of GD/gender incongruence as a child progresses into adolescence. It may be that the presence of GD/gender incongruence in prepubertal children is the earliest sign that a child is destined to be transgender as an adolescent/adult (20). However, social transition (in addition to GD/gender incongruence) has been found to contribute to the likelihood of persistence.” (Hembree et al., 2017, p. 3879)

10. Unnecessary and confusing change in terminology

Undergoing “gender-affirmation” is frequently described as “transitioning” or “gender transition.” When individuals stop the process of transitioning, it is commonly referred to as “detransition.” Detransition, which appears to be a growing phenomenon, is a well-established term used by several recently-published studies.

The authors’ attempt to change established terminology from the well-established “detransition” to the novel “retransition” is at best unnecessary, and at worst confusing. If “detransition” – ceasing to pursue gender transition—is recast as “retransition,” then what shall we call individuals who actually change their minds once again, and do choose to continue to pursue gender transition after all? Several children in the study did just that. Another study has reported on the experiences of transgender-identified individuals temporarily detransitioned.

While it’s admirable when medical terminology evolves to elucidate a concept, we feel that in this case, the change serves to obscure rather than enhance understanding of the topic.

Concluding Thoughts

This study finding of high rates of persistence of transgender identity in children following early social gender transition is consistent with two possible explanations. One is that the study was comprised of the parents who were exceptionally good at predicting their child's future transgender identification. This would imply that although most (61%-98%) of transgender-identified children naturally desist during early puberty, the parents in the study who had made the prediction that their children would persist turned out to be right nearly 100% of the time. While plausible, the probability of this is low. 

The other possible explanation is that early-childhood social gender transition may consolidate an otherwise transient childhood transgender identity. We believe the latter explanation is more likely. The hypothesis that early social gender transition is not neutral but may be a form of psychosocial intervention that predisposes an otherwise transient childhood transgender identity to persist has been voiced before. The study results lend support to this hypothesis. Parents considering undertaking a social gender transition of their gender-variant children need to be advised of this possibility. Notably, the boys in the study (“trans girls”) were transitioned on average one year earlier than girls (“trans boys”): at age 6 rather than 7 years old (Table 2). This may reflect societal discomfort with feminine gender nonconforming boys, which may lead some parents to socially transition gender non-conforming boys at an earlier age.

As the practice of early social gender transition becomes more common, it is reasonable to expect that many more gender-variant youth will persist in their trans identity. This in turn will likely significantly increase the number of young people seeking hormonal and surgical transition, which is of concern because of the poor state of medical knowledge: the longest available set of outcomes of individuals who medically transition in adolescence and young adulthood tracks patients only to an average of age 21, and the best evidence is rated as “low” or “very low” quality.

Currently, active debates are ongoing over the age at which children are old enough to provide meaningful consent (or assent) to undergoing gender transition, due to its inherent risks and uncertainties, as well as the near-certainty of infertility and even sterility, which occurs when puberty blockers at Tanner stage II are followed by cross-sex hormones. Debates whether 16-year olds or 12-year olds can consent to medical interventions with such profound life-long consequences are currently playing out all over Europe, and most recently, in several US states. Since almost all early-socially-transitioned children in the Olson et al study continued with the transgender identity into puberty and over 60% are already undergoing medical transition, the study suggests that many of these life-changing decisions are occurring not at 12, 14, or 16, but effectively at the much younger age of 6 or 7.