Editors’ note: The link to the original book review, which is now a retraction notice, can be found here. Dr. Hall’s book review has been republished by Michael Shermer on Skeptic here. On July 13, Dr. Hall republished a revised version of her review on her own blog. While the revised version is clearer that the original study by Lisa Littman was not good science, it still contains most of the problems and errors we found in the original version.
Introduction: Why Dr. Harriet Hall’s review of Abigail Shrier’s Irreversible Damage was retracted
Science-Based Medicine as a philosophy, blog, and approach to medical practice is primarily about optimizing and affirming the relationship between science and critical thinking with healthcare. One of our core principles has always been that the best evidence available must always inform the standard of care, and that therefore an objective and clear-eyed view of that evidence is critical.
Naturally, we tend to focus the most attention on issues that are controversial or where there is a motivated campaign of anti-science, pseudoscience, or science denial, because that is where the principles of SBM are most needed. Examples include cancer quackery, antivaccine pseudoscience, and, for the last 16 months, misinformation and disinformation about the COVID-19 pandemic and the public health measures instituted to slow its spread. SBM is not apolitical. It can’t be, because the promotion of effective science-based laws and regulations in medicine is an inherently political activity. However, we do strive to be non-partisan. It is not our job to promote a specific political ideology, but we do strive to do our best to make sure that political, social, and personal decisions are informed by the best science possible.
For further background, the SBM blog articles published here are not considered peer-reviewed, but we do have a strict editorial policy. Outside submissions undergo review by our full editorial board, and most are rejected or require revisions prior to publishing. Our own editors, however, have earned the privilege of publishing articles without prior review, since they have a proven track record. They post, and often neither Dr. Gorski nor I see the articles until they are live. That does not mean, however, that we do not have mechanisms of quality control.
If any concerns about accuracy, fairness, or completeness come to our attention, we deal with them in a number of ways. Often, clarification in the comments is sufficient. Sometimes we make corrections to the original text of the article. When we receive outside complaints, one or more of the non-author editors (usually David Gorski or myself) will re-read the original article to determine if, in our view, it was fair and appropriate, which is almost always the case. That doesn’t mean we have to agree with everything in the article, as long as the article adheres to defensible science-based standards.
It further needs to be noted that because of the nature of SBM we often tackle topics from two perspectives simultaneously. Our articles are primarily critical analyses of various topics through the lens of SBM principles. Alternatively, they consist of a discussion of those principles themselves. Often pseudoscience is fairly blatant, so that even general medical knowledge, combined with a thorough understanding of SBM and review of the relevant literature, is sufficient to grasp the relevant science. However, some topics are scientifically more complex, requiring reliance on topic experts if not within our areas of expertise. We are also aware that we have “holes” in our scientific expertise, given the professions and areas of expertise of our editors and contributors.
That is why, when discussing topics outside of our professional areas of expertise, we are acting more like science communicators than scientific experts. Our relevant expertise there is SBM itself, but otherwise we defer to the consensus of expert opinion with regard to technical analysis of the literature and relevant basic science. This is standard science journalist practice. However, this can be a delicate balance.
Two weeks ago, one of our editors published a book review that raised concerns with Dr. Gorski and me, as well as at least one other editor, soon after it published. Reading it, we both feared that this book review had probably strayed beyond evidence or expert opinion and thus required a robust response. This was a review of a book by Abigail Shrier titled Irreversible Damage: The Transgender Craze Seducing Our Daughters. This particular book discussed a complex area of medical practice that also happens to be one embroiled in heated political debate. Because of the context of this topic, we believed it especially critical that SBM be perceived as a politically neutral and reliable source of information about the relevant science. Unfortunately, Dr. Hall’s fellow editors were concerned that the review in question did not achieve this goal.
Our first step was to carefully review the article and then discuss our concerns directly with Dr. Hall to hopefully find a solution. The challenge here was that, while we had enough background knowledge to immediately see there were serious problems with the review, none of us are topic experts. Reviews outside SBM by those with expertise in this area seemed to be making valid scientific criticisms of the opinions and claims in Ms. Shrier’s book, which the review took at face value.
Clearly what we needed was time to do a deeper dive on this complex controversy, to wrap our heads around the published evidence, and to vet the claims and arguments on both sides. This is something we would have preferred to do prior to publication, but we no longer had that luxury. Giving an immediate half-baked analysis would not do SBM readers justice. Ultimately, we decided to hit the “pause” button, to withdraw the review for a time while we consulted outside experts and did our own internal review. Since Dr. Hall indicated she would publish her article on an alternate site (and immediately did), we saw no pressing need to leave the article on SBM while this review was underway.
Having completed what we consider to be a minimally sufficient review, David Gorski and I now present a summary of our reading of the primary literature addressing the core claims of Ms. Shrier’s book, in addition to articles and interviews by Ms. Shrier and other reviews and commentary on the topic. Because Dr. Hall has already republished her review on Michael Shermer’s website, we link to it here and include a link to where her article was originally published on SBM, which now contains our first retraction statement. We also have solicited outside expert reviews which we will publish if after editorial review we deem them to meet the publication standards of SBM.
To be absolutely clear, we encourage full and open scientific debate on this and any topic relevant to SBM. All the SBM editors are passionate about our project because we care deeply about every patient receiving the best medical care possible. This requires not only that medicine be science-based, but that we demand open and transparent discussion of medical science and practice without political or ideological tampering. We regret that many people outside SBM have simply assumed (without evidence or just cause) that our actions were based on a desire to censor open discussion. Nothing could be further from the truth, and our multi-year history of tackling controversial topics and taking many unpopular positions is evidence of that.
Further, we wish to make it clear that Dr. Hall still remains an editor of SBM in good standing. She has worked tirelessly to promote SBM and its principles, contributing over 700 articles to SBM since 2008, all without any compensation or possibility of reward beyond public service. However, at SBM quality matters first, and so we have to remain open to correction when necessary.
Abigail Shrier’s claims and transgender care
In her book, Ms. Shrier creates a very specific narrative. In the last two decades there has been a dramatic increase in the number of youth openly identifying as transgender and seeking gender-affirming interventions. This observation is not controversial. What is controversial is Shrier’s claim that this increase in the number of adolescents “coming out” as trans is likely due to a social contagion, in which social media influencers convince young people they are trans when they are not, which leads them to be herded into irreversible transitional medical procedures they will later regret, all aided and abetted by doctors whose treatment is based more on ideology than science. This narrative, however, is not supported by any evidence and is cobbled together with a gross misreading of the scientific evidence. Unfortunately, Dr. Hall did not discuss the numerous errors of science in Ms. Shrier’s book and, even though she did recognize that much of the evidence was anecdotal, still more or less accepted Ms. Shrier’s narrative at face value.
Let’s review what research there is on transgender care, starting with what is the standard of care for trans youth, since that seems to be the focus of much of the political attention. In her book, as quoted in Dr. Hall’s review, Ms. Shrier likens the current state of transgender care as if one were dealing with anorexia as, “If you think you are fat, then you are. Let’s talk about liposuction and weight-loss programs”. For emphasis Dr. Hall writes: “We are asked to disregard DNA and accept the ineffable feelings of an eight-year-old”.
The analogy here is not apt. Eating disorders are clearly disorders, with now well-established diagnostic criteria and medical risks. The assumption behind the analogy is that feeling as if you are a different gender than the one assigned at birth is also a harmful disorder that should be “cured”. This assumption is not valid and is itself likely harmful. In fact Ms. Shrier relies on an outdated definition from the DSM-IV to make this case. The DSM-V now recognizes that having a gender identity that differs from the gender assigned at birth is not a disorder. Having dysphoria resulting from that fact combined with social factors is.
Further, DNA is demonstrably not determinative when it comes to sexual orientation and gender identity. There are developmental factors, for example. Elsewhere Dr. Hall characterizes the notion that a person can have a female brain in a male body as “biologically nonsensical”, but this is also demonstrably not true. It absolutely is possible to have the secondary sexual characteristics of one gender with the gender identity of another because of other genetic, hormonal, and developmental factors.
Further, looking at published standards and countless interviews with practitioners, the notion that those involved in gender care are caving to political correctness rather than best practice is an unfair caricature that is clearly motivated more by ideology than science and medicine. For example the World Professional Association for Transgender Health publishes treatment guidelines that are widely accepted. First, however, we have to distinguish between transgender children (pre-puberty) and adolescents (something which Ms. Shrier often conflates, causing confusion). The standard of care for children is that they get no medical intervention at all – no hormones, no surgery. Their treatment is entirely confined to psychological assessment and psychosocial interventions. Medical treatment options begin only after children enter puberty (the exact stage varies from country to country). At that point, there are standards that apply.
In order for adolescents to receive puberty-suppressing hormones, the following minimum criteria must be met:
- The adolescent has demonstrated a long-lasting and intense pattern of gender nonconformity or gender dysphoria (whether suppressed or expressed);
- Gender dysphoria emerged or worsened with the onset of puberty;
- Any coexisting psychological, medical, or social problems that could interfere with treatment (e.g., that may compromise treatment adherence) have been addressed, such that the adolescent’s situation and functioning are stable enough to start treatment;
- The adolescent has given informed consent and, particularly when the adolescent has not reached the age of medical consent, the parents or other caretakers or guardians have consented to the treatment and are involved in supporting the adolescent throughout the treatment process.
For those who are interested, the entire document is worth reading. It gives a fairly concise review of the literature.
These standards are for fully reversible interventions. Partially reversible interventions, including cross hormone therapy, have stricter criteria for assessment and informed consent. For irreversible surgical interventions, the following must be met:
Genital surgery should not be carried out until (i) patients reach the legal age of majority to give consent for medical procedures in a given country, and (ii) patients have lived continuously for at least 12 months in the gender role that is congruent with their gender identity. The age threshold should be seen as a minimum criterion and not an indication in and of itself for active intervention.
In other words, surgical interventions are reserved for legal adults who demonstrate sufficient maturity and at least 12 months living in their congruent gender identity.
Other published standards are similarly rigorous. The Endocrine Society clinical practice guidelines recommend hormone therapy for gender dysphoria only for those who, “have undergone psychiatric assessment, and have maintained a persistent transgender identity”.
Of course, these are standards, and not every practitioner adheres perfectly to the standard of care in any aspect of medicine. But we don’t take outliers and use that to criticize the standard or pretend it is typical or common. Interviews with those involved in transgender care indicate that adherence to rigorous standards as outlined above are the norm.
What about the claim that those who identify as trans are going through a phase not likely to persist into adulthood? This, too, is false or misleading. Dr. Hall wrote:
Historically, the conviction that one’s gender doesn’t match one’s anatomical sex typically began around age 2–4. It affected only .01 percent of children, almost exclusively boys. And in 70% of cases, they eventually outgrew it. Prior to 2012, there was no scientific literature on girls age 11–21 ever having developed gender dysphoria at all.
The term “children” above is vague and also critical. Nowhere does Dr. Hall explicitly explain that the literature on the trans phenomenon specifically divides children into prepubescent and adolescents. This is critical, because some of the statistics she is citing above apply only to prepubescent children – that they outgrow their trans identity and that they are “almost exclusively” boys. That latter point is an exaggeration. Data shows that the ratio of AMAB (assigned male at birth) to AFAB (assigned female at birth) prior to puberty identifying as trans is between 6:1 and 3:1 – more boys than girls, but not “almost exclusively”.
Furthermore, the fact that young children may alter their identity as they mature is based on highly flawed research, studies with methods so fatally flawed that the results cannot be trusted, let alone cited as facts. But even if this statistic were reliable, it is less relevant to the discussion. As stated above, there are no medical interventions for this group of children. Medical gender-affirming interventions are reserved for adolescents and older. The data on adolescents is very different. While more data is certainly desired, the statistics we have indicate that nearly all adolescents who identify as trans maintain that identity into adulthood.
Therefore, what Ms. Shrier, and by extension Dr. Hall, are doing is misleadingly using (flawed) statistics for young children to criticize interventions reserved for adolescents and adults who do not share those statistics. To be fair, Ms. Shrier also uses those statistic to argue that young children should not be receiving even social interventions affirming their gender identity, Dr. Hall, however, conflates these age groups. No medical interventions will happen until trans youth are at an age when their trans identity is very likely to be permanent, which is further supported by extensive evaluation according to published standards of care.
What about the alleged spike in people identifying as trans and the shift from predominantly AMAB to AFAB? This is misleading also. Part of the change in male to female ratio is just shifting statistics to different age groups. While young children historically have been 6:1 to 3:1 AMAB, adolescents and adults have always been from 2:1 to 1:1 AMAB to AFAB. The ratio has been shifting over the last two decades, but bringing the ratio closer to parity of 1:1.
Let’s also delve further into the prevalence statistics.
Again, there is no question that the apparent prevalence of trans individuals has increased in the last two decades. The immediate question should be, however, is this a real increase in the underlying phenomenon or simply an increase in the counted number? We see this all the time in medicine. Most analogous to this situation is perhaps the dramatic increase in the number of autism diagnoses from the early 1990s to the 2010s. We have discussed this at length on SBM. It is clear from the data that most of this increase is due to an increase in awareness, availability of services, shifting diagnostic criteria, and diagnostic substitution. Screening can also play a role. As Dr. Gorski likes to say, if you look for something more intensively, you will find more of it. Even with all these factors, it is difficult to rule out a small real increase, but that is certainly not the core of the phenomenon.
There are even examples that are not psychological or neurodevelopmental disorders, like autism, for which the diagnostic criteria can be more subjective than they are for “hard” diagnoses like cancer. In fact, let’s talk about cancer. There is a phenomenon known as ductal carcinoma in situ, which is a premalignant lesion that frequently progresses to breast cancer. Before 1970, it was a very uncommon, even rare, diagnosis. However, since the 1970s, the incidence of DCIS has increased by 16-fold. How could this be? Remember, this is a “hard” pathological diagnosis. The diagnostic criteria for DCIS have not changed; the diagnosis still requires a biopsy showing the characteristic lesions growing within the milk ducts. So what has changed? As discussed here many of times, it’s screening. The increase in incidence of DCIS corresponds very closely to the introduction of mammographic screening programs in the late 1970s. Similar increases have been observed for prostate cancer after the introduction of PSA screening.
Our point here is not to revisit the controversies over screening for cancer and the potential for overdiagnosis engendered by overly intensive screening. These have been discussed on SBM many times before, and not just by Dr. Gorski. Dr. Hall has also written about overdiagnosis. Rather, by bringing up the examples of autism and DCIS, we simply wish to emphasize that it is not at all unusual for the incidence of a medical condition or disease to increase, sometimes dramatically, based on factors that have nothing to do with an underlying change in the prevalence of the condition, factors such as screening, acceptance, and changes in diagnostic criteria. Moreover, this sort of phenomenon has been observed even for diagnoses that are considered “hard”, not just “softer” diagnoses like autism or other neurodevelopmental or psychological conditions. When such an increase in a condition is observed, it is not surprising that scientists (and the public) look for causes. Unfortunately, some will be tempted to latch onto causes that most jibe with their preexisting beliefs. It happened for autism, which was blamed on vaccines. It is happening for trans youth, which is being blamed on a number of causes not well supported by science.
Basically, what appears to be behind the increase in young people identifying as trans is similar to what is behind the increase in autism prevalence. Over the last two decades, the culture has shifted, with greater acceptance of trans individuals and openness to a non-binary approach to gender. There also has been an increase in the availability and visibility of support services for trans individuals and of gender-affirming interventions. Further, the historic numbers prior to the cultural shift were extremely low, around 0.1% in the US, which almost certainly was a gross undercounting. So of course the numbers have significantly increased – we would predict nothing else. Recent surveys put the figure at about 0.4%, a four-fold increase.
The real question is – are these known factors, increased acceptance and availability of resources, enough to explain the increase in people identifying as trans, or do we need to invoke a new phenomenon, such as Ms. Shrier does in supporting the notion that a social contagion is convincing children they are trans when they are not? We clearly need more and better evidence to address this question, but the evidence we currently have strongly favors the former, and there is no evidence to specifically support Ms. Shrier’s view.
For example, a 2019 review of the demographics found:
Flores and colleagues note that their 2016 estimate of the percentage of transgender-identifying adults in the U.S. is double their 2011 estimate, which they attribute to improvement in survey methods (11). Arcelus et al. also describe a general increase in TGNB [transgender non-binary] individuals in Europe within their study period. Greater visibility and acceptance of TGNB individuals, and as a result, greater willingness to self-identify as TGNB and seek transition, may contribute to these trends.
But also, according to USTS results by age 20, 94% of transgender respondents began to feel that their gender was different from the sex assigned at birth, 73% of respondents began to think they were transgender, and 52% began to tell others that they were transgender. These findings were consistent across respondent age groups, which may suggest that reports of growing TGNB prevalence could be attributed to increased awareness, acceptance and self-reporting of TGNB identities among younger generations than an actual increase in prevalence.
It should also be noted that prevalence data is almost all based on those presenting to clinics. This is a self-selective population, and is certainly a subset of the overall trans population. What we are seeing is an increase in those presenting to clinics, with no reason to think there is an increase in the total population.
Ms. Shrier’s narrative is that mostly young girls are being convinced that they are trans, and she offers as evidence the fact that the ratios of AMAB to AFAB trans individuals are shifting. But as we discussed above, we are not observing a flood of AFAB exceeding AMAB, but rather a coming to 1:1 parity, or close to it. The most parsimonious explanation here is that AFAB were previously underreported, and in this case the most parsimonious explanation is likely to be the most accurate. More AFAB are presenting to clinics requesting gender-affirming interventions. And this has two obvious and likely causes. One is the availability of top surgery for AFAB individuals, which is more available and less costly than genital surgery. The other is that the shifting of the gender ratio could mean that there was likely a backlog of AFAB demand for services.
One way to differentiate between these dueling narratives about the causes of the increase in people identifying as trans and requesting gender-affirming interventions is to look at the rate at which trans individuals who did undergo a medical intervention later express regret or detransition. If Ms. Shrier’s fears are valid then we should already be seeing an increase in the incidence of regret and detransition.
About this Dr. Hall writes:
We are starting to see desisters (those who stop identifying as transgender) and detransitioners (those who had undergone medical procedures, regretted it, and tried to reverse course). No statistics are available on how often this happens.
This is misleading. First, it is not true that we have no statistics. But further, if it were true, then how would we know that the incidence is increasing? Her two claims, which appear to come from Ms. Shrier’s book, contradict each other. It turns out, both are wrong.
A 2021 meta-analysis of regret following gender affirming surgery, combining 27 studies and 7,928 transgender patients, found that the pooled prevalence was 1%.
A 2018 survey of surgeons about their own patient statistics found that out of 22,725 patients who underwent GAS only 62 later expressed regret, and out of these only 22 said it was because of a change in their gender identity. The rest cite reasons such as conflict with family or dissatisfaction with the surgical outcome.
Other reviews also find extremely low rates of regret, ranging from 0.3% to 3.8%. Further, if anything the rate of regret is decreasing over time as social support and surgical procedures improve.
The bottom line is that regret is rare and decreasing, which convincingly falsifies Ms. Shrier’s narrative that it is significantly increasing (based entirely on anecdotal evidence). In fact, this is powerful evidence against Ms. Shrier’s entire social contagion narrative.
To bolster the “social contagion” hypothesis, Dr. Hall cites the controversy over “Rapid Onset Gender Dysphoria (ROGD)”. This idea was proposed in 2016, based on a single study that even Dr. Hall now acknowledges to have been bad science. The journal later published a “correction” that mainly added a proper discussion of the preliminary nature of the study. At this point is it not valid to cite ROGD to support the social contagion hypothesis, nor is it valid to suggest that there is now a burden of proof to rule it out, or that it should in any way inform medical practice. This is a thin hypothesis based on shoddy science; it should no more affect medical practice than similar-quality studies should impact the vaccination program.
Further, Dr. Hall’s review creates the impression that gender-affirming interventions are risky or harmful, but to support that narrative she distorts the evidence. She states:
Suicide is common, but there is evidence that factors other than gender dysphoria may be causing the suicidal ideation, and there is evidence that affirmation does not ameliorate mental health problems. In one study of adult transsexuals, there was a rise in suicidality after sex reassignment surgery.
We do not believe that this is a fair assessment of the evidence. A 2016 systematic review of psychological well-being following gender-affirming interventions found:
Two studies showed a significant improvement in psychological functioning at 3–6 months and 12 months compared with baseline after initiating hormone therapy. The third study showed improvements in quality of life outcomes 12 months after initiating hormone therapy for FTM and MTF participants; however, only MTF participants showed a statistically significant increase in general quality of life after initiating hormone therapy.
A 2021 systematic review of gender affirming surgery specifically found:
Findings from this review indicate that GAS [gender-affirming surgery] can lead to multiple, significant improvements in psychological functioning.
This is the first study in which associations between access to pubertal suppression and suicidality are examined. There is a significant inverse association between treatment with pubertal suppression during adolescence and lifetime suicidal ideation among transgender adults who ever wanted this treatment. These results align with past literature, suggesting that pubertal suppression for transgender adolescents who want this treatment is associated with favorable mental health outcomes.
There are more studies, but the summary is that the research over years has consistently shown that gender affirming interventions improve mental health and reduce risk of suicide in trans patients.
Where we agree with Dr. Hall is that the current state of this evidence is far from ideal. Mainly for practical reasons, most of this research is not blinded or controlled. To put this into context, however, most surgical interventions are not studied in blinded trials, and sham surgical interventions are rare. You cannot blind a trans individual to whether or not they received a gender affirming intervention.
But we do agree that given this reality, we need to continue to study and monitor such interventions for both medical and psychological outcomes. This is where an informed medical and ethical discussion should take place, balancing the risks and benefits of interventions given the limitations of the research. There is also a meaningful ethical conversation to be had about the proper age of consent and balancing that with risks vs. benefits of gender-affirming interventions.
Conclusion: There is science supporting transgender care
Abigail Shrier’s narrative and, unfortunately, Dr. Hall’s review grossly misrepresent the science and the standard of care, muddying the waters for any meaningful discussion of a science-based approach to transgender care. They mainly rely on anecdotes, outliers, political discussions, and cherry-picked science to make their case, but that case is not valid.
Most significantly, they warn about medical interventions for children, citing mainly the notion that children are not able to make such choices at such a young age and will likely change their minds, regretting their decision because their gender identity is still developing. However, the age group for which they cite (fatally flawed) statistics do not receive medical interventions, and the age group that is eligible are not likely to change their gender identity. This is a statistical bait-and-switch.
The standard of care waits until children are at an age where their gender identity is generally fixed, and then phases in interventions from most reversible to least, combined with robust psychological assessments. Further, regretting these interventions remains extremely rare, and does not support the social contagion hypothesis.
At this point there is copious evidence supporting the conclusion that the benefits of gender affirming interventions outweigh the risks; more extensive, high-quality research admittedly is needed. For now, a risk-benefit analysis should be done on an individual basis, as there are many factors to consider. There is enough evidence currently to make a reasonable assessment, and the evidence is also clear that denying gender-affirming care is likely the riskiest option.
Steven Novella, MD, Founding Editor, SBM
David Gorski, MD, PhD, Managing Editor, SBM
Coming up – Part II of this series will include a far more detailed discussion of the key claims in Abigail Shrier’s book and where she goes wrong by an expert in the care of trans children and adolescents.