Costa et al, 2015

Psychological Support, Puberty Suppression, and Psychosocial Functioning in Adolescents with Gender Dysphoria


Clinicians followed two groups of gender dysphoric adolescents from an initial pool of 201 at the Gender Identity Development Service (GIDS) in London for 18 months. Clinicians measured participants’ psychosocial functioning (Children’s Global Assessment Scale (CGAS)) at four six-month intervals (baseline, six, twelve, and eighteen months).

All participants received psychological support only for the first 6 months. The clinicians them divided participants into two groups. The first group were “immediately eligible” for puberty blockers and received psychological support and puberty blockers for the final 12 months. The second group were “delayed eligible” and received psychological support only for the same period.

For reasons that are unexplained only 71 of the original 201 completed the full study period.


Study Design

One hundred participants, of the 201 who reached the six-month interval, were not immediately eligible for blockers due to ”possible comorbid psychiatric problems and/or psychological difficulties.“. On this basis, clinicians allocated them to the ”delayed eligible” group. Clinicians assessed 101 as ”immediately eligible”. Consequently, the two groups were not homogenous, and we cannot draw firm conclusions regarding the impact of blockers. A lack of improvement in the delayed eligible group could be a consequence of their more severe problems (Biggs, 2020).

It is also unclear what happened to the 80 participants who did not reach the twelve-month interval or the 50 who withdrew between the twelve and eighteen-month marks. It remains possible that participants withdrew due to adverse effects of treatment or they no longer wished to pursue further intervention. The two groups did have a similar withdrawal rate during the final interval. However, if the reasons for withdrawal differed between the two study groups, then the remaining 35 immediately eligible and 36 delayed eligible would represent biased subsamples.


At the six month interval, after receiving psychological support only, the 201 participants improved their CGAS score from 58 to 61 (discussed below).

The immediately eligible began puberty suppression and both groups continued to receive psychosocial support. During this interval, 41 immediately eligible, and 39 delayed eligible subsequently withdraw, for unknown reasons. 61 delayed eligible, and 60 immediately eligible participants remained at the twelve-month interval after six months of differing treatment.

Over this period, the delayed eligible further improved their score to 63 and the immediately eligible improved to 64.

During the final six month period, 25 in each group subsequently withdraw for unknown reasons, leaving 36 delayed eligible, and 35 immediately eligible at the eighteen-month interval. At this point, the delayed eligible showed no further improvement, and the immediately eligible continued to improve their CGAS score to 67.


The authors note that the improvement over time in each group was significant for both groups, but as Biggs, (2020) points out it is the differences between the groups that matter:

…this difference is not statistically significant: a 2-tailed t-test for the difference between group means yields P = .14, far beyond the conventional .05 threshold. In sum, the samples were so small that we can draw no conclusions. Therefore, the article provides no evidence that GnRHa improves psychosocial functioning.

The study has further ethical and research quality issues in its discussion of the initial improvement in the six months when all participants received psychosocial support only.

The authors attribute this to a possible ”timely addressing of psychosocial problems”, or to the knowledge that ”puberty suppression will be performed within a reasonable time and refer a distress reduction because of their accepted and understood requirements”. In the case of the latter, the implication is that a dysphoric child becomes happier when clinicians agree to provide the treatment they desire, not as a result of the treatment itself. The same placebo effect could be present when the adolescent begins treatment, and would bias results in the treatment’s favour, An especially important issue for studies that use subjective assessments as the primary outcome measure. The more emphasis that the peers and role models place on the supposed necessity of the treatment, the greater the placebo effect will likely be.


The study has multiple flaws and finds no significant differences between the group treated with puberty blockers and the group who received psychosocial support only. When referencing this study, medical professionals must inform parents that their child’s mental health may improve with psychosocial support only.