Hormone Therapy & Suicidality: What My 2025 Follow-Up Study Shows.

In 2019, coming out of my dissertation (Allen et al., 2019), my co-authors and I published a small study (N = 47) examining changes in well-being and suicidality among transgender adolescents following hormone therapy.

This year, we conducted a much larger follow-up with more than 400 youth to see whether those early findings would hold up when we replicated the study with larger cohort and over even a longer observation window. As we know, there is a replication crisis in the sciences (Ioannidis, 2005; Open Science Collaboration, 2015), and there is the decline effect (i.e., as replication increases, initially large effects tend to shrink or vanish; see also my Blog on Purple Hat Therapies. So, to have increasing confidence in outcomes, we need larger, better designed studies. At the same times, even studies with “low quality evidence” but consistent results, or “low quality evidence” in which intervention show large effect sizes, can increase our confidence in the intervention. This is in fact, how most medical science works.

In the 2025 study, Changes in Suicidality Among Transgender Adolescents Following Hormone Therapy: An Extended Study, the results were consistent with the first 2019 study. This time, we observed a medium effect size (rather than large), which is what you might expect with a larger sample size. When that same pattern appears in larger samples, it suggests the improvement is genuine, and less likely to be due to random chance or idiosyncrasies of the first study.

This follow-up study matters because evidence in medicine accumulates over time. It would be rare that one study settles any question (though, I am in favor of devoting resources to larger, better studies rather than multiple small studies). In any case, evaluating evidence requires asking the limitations and quality of individual studies, as well as replication, convergence, and consistency across methods and populations.

In the area of gender care for transgender and non-binary youth, the findings have been remarkably consistent, and they tend to point in the same direction. Across dozens of studies, transgender youth who begin hormone therapy typically experience measurable improvements in a wide range of outcomes (Turban, 2025; Budge et al., 2025).

That does not mean the evidence is perfect. If one were to take the most cautious view, it’d be safe to say that it appears there is no evidence of harm for most patients. And the harm that does occur, appears less frequent than harms we accept in other areas of medicine. Given recent socio-political events (bans on access/providing hormone therapy [HT] for minors), one would think we have very clear evidence that almost everyone is harmed or regrets HT during adolescence. That is not the case. In fact, we have dozens of studies, one after another, all showing the same general trend of benefit.

But evidence-based medicine (EMB) has never just been about the data. The evidence alone cannot tell us what to do. That depends on the patients and stakeholders’ values and preferences. Some people place a high value on current distress and potential benefits of an intervention. Others may place a low value on current distress, and high value on potential harms of a given intervention. More evidence will solve those underlying value-laden questions. This brings us to a related question: what can clinicians responsibly claim about the scope and limits of therapeutic interventions?

As a mental health practitioner, it is also important to be clear about what therapy can and cannot do. There is no credible evidence that psychotherapy can alter a person’s gender identity or reliably resolve gender dysphoria by changing it. People come to therapy to understand themselves, to explore, and sometimes their sense of identity does evolve and change—but that is discovery and exploration, not conversion. Therapy’s role is to help people clarify what is true for them and to support informed, values-aligned decisions about their lives.

If we are practicing good gender care, we are never telling any patient “You are transgender and you need X intervention.” It is about supporting the patient as they figure that out themselves (if they don’t already know) and providing accurate information about treatment options, which may include non-treatment.

When we step back and take a broader view – across medical research, psychological outcomes, and lived experience – the evidence of benefit is consistent. It may not meet the highest standard of certainty, but the signal is strong and stable over time. That is what good science looks like in the complexity of real life.