Thread by @SEGMtweets

Gender medicine does not have a monopoly on bad science. But if poor research were an Olympic event, it would arguably be a favorite to win the gold. SEGM researchers analyze a recent systematic review of puberty blockers, highlighting key problems./1
segm.org/flawed_systema…

Since individual studies can be unreliable, clinicians prefer to rely on systematic reviews (SR) of evidence. SRs scrutinize all the studies about a topic using rigorous and reproducible methods, separating the “signal” from the “noise.” However, a misleading SR can cause harm./2

A SR by Rew et al. of puberty blockers concluded, "the evidence to date supports the finding of few serious adverse outcomes and several potential positive outcomes." This contradicts the UK NICE SR, which found "little change with GnRH analogues from baseline to followup." /3

How could two systematic reviews, conducted during the same period, and tackling the same topic, have come to such different conclusions? The answer lies in the rigor (or the lack thereof) of the review process itself, which is sorely lacking in the Rew et al. SR. /4

The systematic review by Rew et al. identified only 151 of the 525 potentially eligible studies found by NICE. Notably, Rew et al. failed to include several studies that highlight problems with puberty blockers. /5
#MedTwitter #psychtwitter

They also failed to accurately assess the quality of the studies, as exemplified by their poor analysis of Turban et al.'s puberty blockers study. The problems with that study (selection bias, confounding, inadequate exposure measure, etc) put it at a critical risk of bias. /6

The review by Rew et al. inappropriately wades into the territory of making treatment recommendations (which is the prerogative of guideline developers) while skipping the required steps. They inappropriately suggest to professionals that puberty blockers are "life-saving."/7

The Commentary uses the example of the Turban et al. 2020 study on puberty blockers & suicidality to show how weak studies are "strengthened" by subsequent publications. This inadvertent game of telephone, at times bordering on evidence-laundering, seems endemic in the field. /8

Given the problems in the studies uncovered by reviews done by public health authorities (US, UK, Sweden, Finland), consumers of SRs (including clinicians, patients, policy-makers) should be wary of claims that suggest certainty, even if they appear in peer-reviewed journals. /9

The Commentary aptly asks, when the evidence used to recommend treatment comes from such poor quality, contradictory data, can patients really be considered to be giving informed consent?
We encourage you to read the full commentary at the link below. /10 acamh.onlinelibrary.wiley.com/doi/pdf/10.111…