© Getty Images
In September 2019, JAMA Psychiatry, a peer-reviewed journal of the American Medical Association, published an article entitled 'Association Between Recalled Exposure to Gender Identity Conversion Efforts and Psychological Distress and Suicide Attempts Among Transgender Adults' by psychiatrists Jack Turban and Noor Beckwith, and epidemiologist Sari Reisner. Turban, the article's corresponding author, is a well-known advocate for trans rights and medical treatment of transgender-identifying individuals. This publication concluded that therapy causes harm, and has been used to promote bans on psychotherapy for gender dysphoria worldwide.
As the title of the published study indicates, Turban and his co-authors (whom we will describe simply as "the authors") set out to examine the effects of what they describe as "gender identity conversion" on the mental health of transgender-identifying individuals. In examining data from the 2015 US Transgender Survey (USTS), they found that survey participants who responded affirmatively to a specific yes/no question — "Did any professional (such as a psychologist, counselor, religious advisor) try to make you identify only with your sex assigned at birth (in other words, try to stop you being trans)?" — reported poorer mental health than those who responded negatively. From this, the authors conclude that what they call "gender identity conversion efforts" ("GICE") are detrimental to mental health, and so should be avoided in children, adolescents, and adults. These conclusions were widely publicized by mass-media outlets, including as a means to advocate for legislative bans on "GICE," a policy step that the study authors have explicitly endorsed.
We agree with Turban's position that therapies using coercive tactics to force a change in gender identity have no place in healthcare. However, we take issue with the authors' problematic analysis and flawed conclusions, which they use to justify the misguided notion that any practice that deviates from reflexively "affirmative" psychotherapy for gender dysphoria (GD) is harmful and should be banned. Their analysis is compromised by serious methodological flaws, including the use of a biased data sample, reliance on poorly constructed survey questions, and the omission of any control variable that tracks subjects' baseline mental-health status.
Further, their conclusions are not supported by their own analysis, even if one puts aside the above-described methodological flaws. While the authors claim to have found evidence that "GICE" is associated with psychological distress, what they actually found was that survey respondents recalling "GICE" were more likely to self-report serious mental illness, a fact that lends itself to a variety of different interpretations. The authors' decision to interpret this association as evidence of "GICE" harm is unwarranted, because neither the existence of causation, nor the direction of any causation, can be discerned from the study due to its cross-sectional design. In fact, an alternative explanation for the found association — that individuals with poor underlying mental health were less likely to be affirmed by their therapists as transgender — is just as likely, based on the data presented. And yet, in the "Conclusions and Relevance" section of their article, the authors offer the baseless claim that "these results support policy statements from several professional organizations that have discouraged this [GICE] practice." It's exactly the sort of easily digestible statement that journalists are happy to report uncritically, especially if it comes with the imprimatur of JAMA editors. But presenting a highly confounded association as causation is simply a serious error, and one that already has served to misinform and mislead clinicians, policymakers, and the public-at-large about this important issue.
What is arguably even more problematic than the authors' flawed analysis is the simplistic idea that one's mode of therapy may be reduced to "affirmation" versus "conversion." The notion that all therapy interventions for GD can be categorically classified into this simplistic binary serves as the implicit foundation for the authors' analysis and conclusions. Yet it betrays a misunderstanding of the complexity of psychotherapy. At best, it overlooks a wide range of ethical and essential forms and ideologically neutral forms of psychotherapy that don't fit into any such binary. At worst, it mis-categorizes ethical psychotherapies that don't fit the "affirmation" descriptor, stigmatizing them as "conversion therapies."
We originally raised our concerns about the quality of the authors' study and the validity of their conclusions in a letter to the editor of JAMA Psychiatry. Our letter was rejected, reportedly due to space limitations. In the months that followed, as we observed the authors' unsupported claims being repeated globally, we felt compelled to write a more detailed critique of the study, which we published in a letter to the Archives of Sexual Behavior on October 21st under the headline One Size Does Not Fit All: In Support of Psychotherapy for Gender Dysphoria. We have adapted that letter here for a general audience. Our aim is to identify the more problematic areas of the authors' analysis and illustrate how heeding their recommendations will serve to limit access to ethical psychotherapy for individuals suffering from GD, further disadvantaging this already highly vulnerable population.
Biased sampling
As noted above, Turban and his co-authors used data from the 2015 USTS survey of transgender-identifying individuals. This survey used what is known as convenience sampling, a methodology that tends to generate low-quality data. Specifically, survey participants were recruited through transgender advocacy organizations, and subjects were asked to "pledge" to promote the survey amongst friends and family. This recruiting method yielded a large but predictably skewed sample. While the authors acknowledge that the USTS may not be representative of the US transgender population as a whole, they proceed to treat it as a valid source of data for major policy recommendations.
In a data table contained in our Archives of Sexual Behavior article, we compared the demographic characteristics of the USTS participants to those of transgender participants from a high-quality probability sample collected by the Centers for Disease Control Behavioral Risk Factors Surveillance System (BRFSS). Even after applying weighting to correct for known survey biases, the USTS participants were far more likely to be young (42 percent were 18-24 years old, versus just 22 percent in the BRFSS data) and educated (47 percent report having completed post-secondary education, versus 14 percent in BRFSS). They are far less likely to own a home (16 percent vs. 55 percent) or to be married or coupled (18 percent vs. 46 percent). They are also much more likely to self-identity with a non-binary gender identity (38 percent vs. 22 percent). In regard to self-reported sexual orientation, only 15 percent of the USTS participants report a heterosexual orientation, compared to 69 percent of the BRFSS participants (it is not clear if sexuality in either case was reported relative to one's sex or gender identity).
A number of additional data irregularities in the USTS raise further questions about the quality of the data. Many of the survey participants (nearly 40 percent) had not transitioned either medically or socially at the time of the survey, and a significant number reported no intention to transition in the future. The information about treatments received does not appear to be accurate, as a number of respondents reported the initiation of puberty blockers after the age of 18 years, which, for obvious reasons, is highly improbable. Furthermore, the survey had to apply a special system of data weighting due to the unexpectedly high proportion of respondents who reported that they were exactly 18 years old. These irregularities raise serious questions about the reliability of the USTS data.
By targeting transgender advocacy groups, the survey underrepresented the experiences of transgender individuals who are not politically engaged. The emphasis on the survey's explicit goal to highlight the injustices suffered by transgender people during the recruitment stage, and in the introduction of the survey instrument itself, made it vulnerable to overreporting of adverse experiences due to "demand bias" (also known as the "good subject effect"). This form of bias presents when researchers reveal their hypothesis and aims, thereby encouraging participants to support the investigator's aims with their answers.
Finally, the experiences of detransitioners and desisters — people who once self-identified as trans, but then reverted to an identity that aligns with their biology — were not included in the survey, as they were disqualified from completing it. Failure to include such individuals in research regarding psychological interventions for GD is a serious oversight. These individuals, whose transgender identification proved to be transient, may well have been hurt by therapies that "affirmed" them as transgender, and may have benefitted from therapies that helped them successfully ameliorate their GD. It isn't surprising that an activist group would see fit to deliberately exclude them from its survey. But it's surprising that academic authors and the editors of a prestigious journal would take the same view.
Invalid indicators
Turban and his co-authors' conclusions rest on the assumption that they had a valid way of determining whether or not a respondent was exposed to unethical forms of conversion therapy, as they define that term. Yet, the USTS question they rely on for this (Question 13.2 in the survey, quoted above) is too non-specific to serve as a valid measure of gender-conversion therapy:
"Did any professional (such as a psychologist, counselor, religious advisor) try to make you identify only with your sex assigned at birth (in other words, try to stop you being trans)?"The question conflates mental-health encounters with a range of completely unrelated professionals. There is no information about whether the recalled encounter was self-initiated or coerced; whether it involved diagnostic evaluation or a specific therapeutic intervention; whether the focus of the encounter was gender dysphoria or another condition; or whether shaming, threats, or other unethical tactics were experienced. This lack of context and detail renders the question useless as a means of differentiating between ethical non-affirmative (neutral) encounters and unethical conversion therapy.
Consider a common situation in which a patient is seeking approval for medical treatment for GD, wherein the role of the therapist is to assess the individual's mental health to ensure that the GD is not secondary to another condition. It has long been known that such encounters can be experienced by patients as a bad-faith attempt to withhold sought after treatment. Further, patients with psychiatric diagnoses, highly prevalent in transgender-identifying populations, can potentially experience or misinterpret neutral interpersonal interactions as invalidating or rejecting. Not only does the survey question provide no detail that would help discriminate between legitimate therapy and unethical conversion therapy, but it arguably biases the recall of neutral encounters toward recall of conversion by using emotionally charged language (e.g., "stop you being trans") and by conflating recall of religiously motivated encounters with clinical ones.
The authors ignored these issues in their article, and instead created a veneer of certainty by treating a positive response to USTS question 13.2 as proof of "GICE," or "gender identity conversion efforts." GICE itself is a novel term that, conveniently, the authors define in no identifiable way except by circular reference to USTS question 13.2 itself. (And oddly, Turban himself referred to the same USTS question by yet another apparently self-invented term, "PACGI" or "Psychological Attempts to Change Gender Identity," in a publication just weeks earlier.) Turban and his co-authors seem to implicitly define "GICE" as any professional encounter that the subject recalls as non-affirmative of their transgender identity. But non-affirmative care is clearly not the same as "conversion," as the latter implies an unethical therapist with a biased agenda and seeking a fixed outcome. This basic logical error undermines all of the arguments that follow.
Data misinterpretation
One finding that emerges from the authors' analysis is that USTS participants who recalled exposure to "GICE" were more likely to report severe psychological distress, as evidenced by their likelihood to score 13 or higher on the widely‐used Kessler K6 non‐specific distress scale. From this, the authors conclude that "GICE" has adverse effects on mental health.
Before addressing the unsupported claim of causation, we'd like to further explore the use of the K6 scale, which itself raises red flags that stand apart from the above-described problems with the underlying survey data. The K6 scale was specifically developed as a tool to discriminate between cases of non-specific psychological distress and cases of serious mental illness (SMI). Scoring at least 13 on this test is predictive of having a Diagnostic and Statistical Manual of Mental Disorders (DSM) diagnosis of schizophrenia, bipolar disorder, and a range of other major mental-health conditions that cause serious functional impairment. Thus, the authors' claimed association between a survey respondent's recall of "GICE" and that respondent scoring 13 or higher on the K6 scale would, if meaningful, suggest that USTS participants recalling "GICE" are more likely to have a severe mental illnesses diagnosis than those not recalling "GICE." Further, if this association could be shown to be causative, which Turban and his co-authors are eager to suggest, this would imply that exposure to "GICE" causes serious mental illness in previously mentally-well populations. This is a highly speculative hypothesis, and, as discussed below, also highly implausible.
Firstly, the authors' analysis contains no control for survey respondents' pre-"GICE"-exposure mental health status. Not only does this critical omission confound the association between exposure to "GICE" and present mental health, it may actually mask reverse causation — i.e., It was the individual's underlying poor mental health that led to their experience of "GICE" in the first place. (We are not herein arguing that such causation exists — because, unlike Turban and his co-authors, we recognize the limitations of the USTS data. We are merely showing readers that the data on which the authors rely lends itself to conclusions very much at odds with their own preferred view.)
Let us revisit the example of a common clinical encounter in which a person with GD and one or more comorbid psychiatric conditions presents for assessment with the goal of obtaining approval for cross-sex hormones. An assessment of such a complex presentation generally requires multiple sessions, and involves ascertaining whether the GD is secondary to another condition. It is also likely that the clinician might focus on treating the comorbid condition(s) first, before pursuing "gender-affirming" interventions. While such a contact would be recalled by the respondent as non-affirmative, and thus likely classified as "GICE" according to Turban's typology, it would be, in this scenario, the patient's poor mental health status that led to the "non-affirming" nature of the encounter, rather than vice versa. If such an individual had attempted suicide in the past, or continued to struggle with mental illness more recently, the analysis used by Turban and his co-authors would serve to attribute these results to "GICE," when in fact, no such attribution is warranted.
In fact, the failure to control for subjects' baseline mental health makes it impossible to determine whether the mental health or suicidality of the survey respondents worsened, stayed the same, or potentially even improved after non-affirming encounters. Given the high rate of co-occurring mental illness in transgender-identifying patients, failure to control for prior mental health status is one of the more serious methodological flaws that undermine the authors' work.
Conclusion
Coercive techniques aimed at forcing unwanted changes in individuals are unethical, and have no place is modern psychotherapy, whether in the form of "conversion therapy" or otherwise. This principle is self-evident and needs no additional justification. However, as we have demonstrated, Jack Turban and his co-authors have failed to prove that "GICE," as they apparently define it, is coterminous with conversion therapy, or with any other unethical therapeutic practice; let alone that it causes poor mental health or suicide attempts in study subjects. Their data consists of answers to a badly convoluted question, conceived by an ideologically motivated group, and administered to an explicitly biased survey sample. By means of unwarranted logical leaps and linkages, and without any of the control data one would normally expect in a study of this kind, they then proceed to extrapolate dubious associations that, through torqued language, are presented as being what a journalist, politician, or ordinary layperson would misinterpret as causative. Rather than appropriately acknowledging the significant study limitations they faced, and calling for more research, the authors instead leveraged their flawed findings to engage in a media campaign aimed at promoting legislative bans of "GICE." Turban himself co-wrote an op-ed — citing himself, naturally — concluding that "It's time for conversion efforts to be illegal in every state, before more people die." The word "conversion" appears 10 times in the article, and appears meant to denote any practice other than unfettered affirmation.
It is hard to know how JAMA permitted the paper to be published in the first place. But having published it, one might easily understand why its editors would be unwilling to publish critiques: The errors are so glaring that they serve to impugn not only the authors but the editorial procedure that led to its publication. To the best of our knowledge, all of the letters written to the editor of JAMA Psychiatry, many by respected academics and clinicians who outlined the serious problems we have described, have been rejected. (Some of them were later submitted as non-indexed comments in the online publication). These omissions serve to stifle scientific debate, and perpetuate the politicization of transgender healthcare, whereby treatment decisions are increasingly legislated by politicians.
It also hurts some of the most vulnerable people in our society. Consider a female victim of sexual assault, who subsequently develops an intense discomfort with her female anatomy and expresses a desire to undergo biomedical interventions to change her body. As psychotherapists, we are personally familiar with this kind of case. We also know that it would be unethical for a clinician to overlook the contribution of sexual victimization to such a case of nascent GD. A therapist enthusiastically supporting this patient's new male identity — the outcome Jack Turban would advocate as the only alternative to his catch-all category of "GICE" — would be failing to provide appropriate treatment for what amounts to a post-traumatic condition, instead providing an inappropriate treatment with the potential to harm. Similarly, a boy who has been traumatized by relentless bullying due to his gender "non-conformity" (e.g., interest in classical music or fashion and avoidance of sports), may conclude that if he were a girl then he would "fit in" and the humiliation would stop. In this case, too, gender-affirming interventions miss the mark when what this traumatized young person requires is psychotherapy.
Another obvious difficulty arises when same-sex attracted adolescents report cross-sex identifications. Research shows that a high number of homosexual adults have experienced periods of "cross-sex" behaviors and cross-gender identification in childhood and adolescence, often to a degree that is severe enough to warrant the diagnosis of GD, or Gender Identity Disorder, as it was previously known. When a dysphoric same-sex attracted young person in the midst of this developmental process presents for mental health care, a clinician overtly affirming the patient's cross-sex gender identity would be failing this patient by not addressing the patient's struggle with same-sex attraction and/or internalized homophobia. In fact, some homophobic families — and, indeed, whole societies — have embraced "affirmative" biomedical trans pathways, as they view a nominally heterosexual trans-identified child as less threatening to their traditional values than a gay boy or lesbian. In cases where the gay child is encouraged to transition, the "affirmative" care is itself a form of gay conversion therapy.
And while one wouldn't know it from reading the results of the USTS, or from papers such as 'Association Between Recalled Exposure to Gender Identity Conversion Efforts and Psychological Distress and Suicide Attempts Among Transgender Adults', GD can present as a transient symptom that resolves either spontaneously or in the context of developmentally informed psychotherapeutic treatment. Some common examples of transient gender-dysphoric states include teen girls, often on the autism spectrum, experiencing distress around the physical and social changes of puberty; or gender-non-conforming young women struggling with shame about being seen as "butch." These individuals, searching for ways to understand and remedy their distress, can incorrectly attribute their discomfort to being transgender. Several case reports indicate that the distress of young people with GD can lessen or resolve with appropriate psychotherapeutic interventions that address the central issues.
If anything other than "affirmation" is viewed as "GICE," it follows that the provision of psychotherapy in these clinical scenarios would be seen as harmful conversion efforts. Yet these therapeutic interventions do not aim to convert or consolidate an identity, but instead aim to help individuals gain a deeper understanding of their discomfort, the factors that have contributed to their distress, and their motivations for seeking transition. These exploratory questions are consistent with the principle of therapeutic neutrality — a cornerstone of ethical psychotherapy. In fact, both (actual) "conversion" and "affirmation" therapy efforts are alike in that both carry the risk of undue influence, potentially compromising patient autonomy. By contrast, the provision of a neutral, unbiased psychotherapeutic process that allows patients to clarify their feelings and assess their various treatment options, which can range from non-invasive to highly invasive, irreversible procedures, is arguably the only way that meaningful informed consent for the latter can be obtained.
The unproven claim by Turban that non-affirming therapies are dangerous — indeed, fatally so — stands in contrast to the documented risks and uncertainties associated with hormonal and surgical interventions that are a core part of the "affirmation" treatment path. Until recently, puberty blockers were considered safe and fully reversible. But there is now emerging evidence of their adverse effects on bone and brain health. Additionally, since almost all of the children treated with puberty blockers proceed to cross-sex hormones, concerns have been raised that puberty blockers may in fact consolidate gender dysphoria in young people, putting them on a life-long path of biomedical interventions.
Cross-sex hormones are associated with cardiovascular complications, including a four-fold increased risk of heart attacks in biological females, and a three-fold increase in the incidence of venous thromboembolism in biological males. "Gender-affirming" surgeries, as they're euphemistically described, can cause urethral stricture, neo-vaginal stenosis and prolapse, and long-term post-mastectomy pain. The effects of "gender-affirmative" care on fertility have not been adequately studied, but infertility is a likely outcome, depending on the specific treatments pursued. (It remains unclear to what extent fertility concerns will be important to this group of patients as they mature. But increasingly, gender treatment centers are recommending fertility-preservation procedures prior to undergoing hormonal interventions.)
Given the absence of robust long-term evidence that the benefits of biomedical interventions outweigh the potential for harm, especially among young people, it is evident that the least invasive treatment options should be pursued before progressing to more risky and irreversible interventions. To the extent that psychological treatments can help an individual obtain relief from GD without undergoing body-altering interventions, ensuring access to these interventions is not only ethical and prudent but also essential.
The rate of regret, detransition, and desistance from transgender identification is largely unknown, notwithstanding the anecdotally observed large increase in detransitioning patients speaking out in social media forums following transitions they have come to regret. What we do know is that the majority of patients with classical, childhood-onset gender dysphoria desist from transgender identification at some time before young adulthood. The minority who persist with their transgender identification into adulthood and undergo "gender-affirmative" surgeries have been reported to have low rates of regret and detransition. Despite the shortcomings of such studies, these results support the need to make transition services available to those who truly do have real and persistent forms of dysphoria. But such cases should not be generalized in a way that erases the high prevalence of desistance among trans-identified youth.
The novel cohort of young GD patients increasingly presenting for help is poorly understood. It is over-represented by adolescent females with recent-onset GD and with comorbid mental health and neurocognitive issues. The trajectory of GD among these young patients, including the rates of desistance and detransition, remains unknown. However we know that in our own clinical practice, we are seeing increasing numbers of detransitioners with adolescent-onset GD who regret not having received exploratory psychotherapy to help them understand their distress and the desire to transition before they underwent irreversible medical and surgical treatments. The same is true of many of our colleagues, some of whom have not yet spoken out, which is difficult in the current environment.
We are especially concerned by the number of youth who report that when doubts about their own transgender status arose, their therapists continued to affirm them as transgender, attributing any doubts to internalized transphobia, and encouraging them to continue medical interventions, which, in turn, unnecessarily exacerbated the psychological and physical harms. Such misguided and harmful forms of clinical practice would only be encouraged by any policy that serves to stigmatize neutral psychotherapeutic services as "conversion therapy."
Advocates of "affirmative care" naturally downplay the risks arising from inappropriate transitions. (Turban himself has described them as largely "cosmetic.") Yet we are seeing increasing numbers of patients who feel deeply traumatized by inappropriate transitions. They suffer from irreversible physical changes, including alterations to their genitals and sexual function, sterility, painful vaginal atrophy, chest/breast alteration and scarring, deepening of the voice, unwanted permanent changes to facial hair growth, male-pattern baldness, urinary incontinence, and other lasting effects. Apart from the distress that these changes cause directly, they also negatively impact many areas of patients' lives, including their ability to form a stable identity (many feel trapped in a "gender no-man's land"), to find romantic partners and supportive social networks, to bear children, or to secure employment. The process of coming to terms with these consequences of their transition is psychologically difficult and can be profoundly painful.
Given the risky and irreversible nature of "gender-affirming" treatments, it is concerning that for many years now, there has been a lack of systematic research into the role that developmentally-informed psychotherapy can play in the amelioration of GD, especially among young people. The need for the continued development and evaluation of non-invasive psychological treatment alternatives for GD has never been more urgent, given the fact that over three percent of young people now report transgender identification or ideation. In the face of this phenomenon, it is time to raise the bar on science — or at least enforce the bar that already exists — and to heed the first and most fundamental tenet of medicine: "First, do no harm."
The authors include Dr. Roberto D'Angelo, a psychiatrist who works with trans-identifying teens and adults, and training and supervising psychonalyst at the Institute of Contemporary Psychoanalysis, Los Angeles; Ema Syrulnik is a healthcare data analytics expert; Sasha Ayad, a licensed professional counselor who works with adolescent-onset gender dysphoric teens and young people; Lisa Marchiano, a psychotherapist who works with detransitioners; Dianna Theadora Kenny, formerly professor of psychology at the University of Sydney, is currently a practicing psychologist and psychotherapist working with trans-identified adolescents and their families, and author of a book on gender dysphoria; and Patrick Clarke is a psychiatrist in Australia. All authors are affiliated with the Society for Evidence-Based Gender Medicine (SEGM).
Comment: