NRPLUS MEMBER ARTICLE {C} oncerns about the efficacy of so-called gender-affirming care stem from the universally poor quality of studies that inform the topic. These worries could be addressed through a well-executed randomized control trial, conventionally the gold standard of medical research. A new study released this week is the first ever randomized control trial on the issue, but it was so poorly executed that it only obfuscates and misleads.
The study — “Early Access to Testosterone Therapy in Transgender and Gender-Diverse Adults Seeking Masculinization” — was published in the Journal of the American Medical Association (JAMA), a once-esteemed journal that has become so deeply politicized that it’s now a preeminent source of medical misinformation.
In the study, Australian researchers randomly divided 64 adult women with gender dysphoria into two groups. The 32 women in the first group received immediate treatment with testosterone; this was the treatment group. The women in the second group adhered to the standard-of-care three-month waiting list for testosterone treatment; this was the control group, and during this three-month experiment, these women received no intervention at all, not even a placebo. The researchers administered surveys to capture baseline measures of gender dysphoria (i.e., distress), depression, and suicidality in both groups. They then administered the same surveys three months later to compare changes in the treatment and control groups.
Randomized control trials are generally seen as empirically rigorous because the only difference between the treatment and control group should be their assignment to that group. Consequently, differences in outcomes can be attributed to the intervention itself. However, the new study was not a “blinded” trial — one in which subjects are unaware of whether they are receiving the treatment. It is therefore impossible to distinguish how much of any change in outcome is caused by the treatment itself or by a placebo effect that could be produced by something harmless that the subjects believed to be the treatment.
Other studies have demonstrated placebo effects for reducing depression that dwarf the benefit claimed for testosterone in the JAMA study. For example, in studies where subjects begin with high levels of depression, a placebo can lower reported depression by more than 9 points on the PHQ-9 scale (the same scaled used in the JAMA study) three months later, while in studies where subjects have milder initial depression, a placebo can lower depression scores on that same scale by more than 4 points after three months.
In the JAMA study on testosterone, subjects who received testosterone reported a reduction of 5.6 points on the PHQ-9 measure of depression after three months relative to subjects who knew that they had received nothing. The benefits of placebos in reducing depression suggest that researchers could have produced that size benefit by injecting subjects with a harmless placebo instead of testosterone.
Similarly, other research indicates that subjects receiving placebos can also experience reductions in suicidal ideation over time, as measured on the SIDAS scale, which the JAMA study also uses. For example, in one study of an intervention to reduce suicidal ideation, subjects who received the placebo intervention experienced a reduction on the SIDAS scale of 4.42 points after 4.5 months, not dramatically different from the claimed benefit of 6.5 points on the same 50-point scale in the JAMA study.
The JAMA study does not demonstrate the benefits of testosterone for reducing self-reported depression or suicidality. It probably demonstrate the benefits of subjects believing they are receiving treatment rather than knowing they are receiving nothing. The women in the control group may also have been negatively biased in their outcomes by waiting for a treatment that they believed would help them but were not yet receiving.
That the researchers observed improvements in feelings of dysphoria in the treatment group offers perhaps the most damning evidence that they were in fact observing a placebo effect. Testosterone treatment takes many months to produce changes in physical appearance, the mechanism by which gender dysphoria is claimed to be treated. In fact, no visible physiological changes occur in the first three months of testosterone treatment other than skin oiliness and acne and cessation of menstruation. It takes several more months for testosterone to cause facial-hair growth and a change in muscle mass to make women look more like men, which is the proposed cure for dysphoria. In the JAMA study, the change in the dysphoria measure had to be caused by the anticipated effects of receiving a treatment, which is a placebo effect, rather than by the testosterone itself.
The usefulness of the study (or lack thereof) is also limited by who was featured in it. The sample included 64 individuals ages 20 or older. Ethical and legal debates about gender medicine focus on whether it is sensible to allow children to undergo irreversible medical treatments. If recent history is any guide, activists will nonetheless champion these JAMA findings to lobby in favor of pediatric sex-trait modification, a practice that the rest of the world is rapidly abandoning.
As the first randomized control trial on gender medicine, the study will feature prominently in trans advocacy. Unfortunately, it only underscores that such advocacy is detached from scientific reasoning and not animated by it.
Jay P. Greene is a research fellow at Do No Harm. Ian Kingsbury is the director of research at Do No Harm.