Thread by @JLCederblom

AusPATH likes to put out statements that look good on the surface but quickly deflate when you dig into them. Recently they put out this: auspath.org/gender-affirmi…. Let's take a look at a specific part of it which I am very familiar with.

The authors make two claims: that adhering to carefully developed guidelines is associated with little regret and "de-transition", for which they offer three references; and that "de-transition" is often a consequence of external pressures, for which they offer one reference.

I will leave aside the fact that they use loaded language and try to sneak in "carefully developed" when that isn't the case, and just consider the extent to which their references support the claims they make.

The first thing to ask is "did these studies follow these guidelines the authors speak of?", to which the answer is no. All of them partially or fully predate those guidelines, and to various degrees deviate from those later guidelines.

Already, the sentence has collapsed. "As a consequence of adherence to these carefully developed guidelines" is simply unsupported. It instead becomes a claim backed only by the authority of the authors, people who have already attempted to mislead the reader.

However, we can continue dissecting the sentence. The claim that remains is that "gender affirming healthcare is associated with very low rates of regret", which the references could support. So let's look at the three references individually and see what they say.

de Vries et al., 2014, is required reading for anyone getting involved in this field. And it does explicitly say "all young adults in this study were generally satisfied with their physical appearance and none regretted treatment", which seems to support the claim.

But wait, what's this, "1 transfemale died after her vaginoplasty owing to a postsurgical necrotizing fasciitis"? And "refusal (n = 2), failure to return questionnaires (n = 2)"? And "being medically not eligible (eg, uncontrolled diabetes, morbid obesity) for surgery (n = 3)"?

What de Vries et al., 2014, does is simply exclude anyone who had a bad outcome. You don't get to say "because those who had good outcomes did not regret it in the short term then no one had regrets". That's patently false.

This is further complicated by the original sampling method, which would exclude early dropouts, potentially further distorting their findings. So the first reference on its surface supports the claim, but actually does not when you take a serious look.

The second reference is Wiepjes et al., 2018, where the authors looked at a large number of patient records and found very few regretful cases - only 14 among two and a half thousand patients. Again, it looks like it supports the claim. Let's dig deeper.

When we use the word "regret" we usually think of it as meaning the person "wishes it had not happened" or "would not do it again". That's not what took place in this paper, which defined regret as "going back on natal sex hormones and expressing regret in patient records".

This combination of medical detransition - while staying at the same clinic - and expressing regret in patient records, does not give us anything like what we would normally consider the word regret to mean. The fact that they note a 36% loss to follow-up should clue AusPATH in.

It also misses someone feeling they're "too far gone to go back" - a real possibility with public examples - or people who originally detransitioned for non-regret reasons. If you can think of a scenario other than the a straight forward instant clarity regret, they'd be missed.

These are all very obvious as soon as you consider the limitations of the study, but that is not in the interest of the authors, because they are not approaching this from a scientific angle, they are simply propagandizing, or even proselytizing.

They see a low number and they want it to be true, so they will not question it further. They are firmly in the clutches of confirmation bias and they get upset when others don't join them.

The final reference is Brik et al., 2020, which does not deal with regret, but is presumably included because it reports on how many children continued onto cross-sex hormones after puberty blockers. This is an odd inclusion, because calling that detransition would...

...imply that puberty blockers are a form of transition, which is practically true, but usually a big no-no for them to say, they must describe it as a "pause button" providing "time to figure out your gender identity".

I don't think the authors understood this implication.

So none of their references really support their regret statement past surface level, and in some ways they contradict or undermine it. Which means that the statement rests only on the authority of a group of people who have been shown to not understand what they're citing.

Not understand, or likely not care to understand. I suspect they're more than capable of intellectually working it out, but they are emotionally bound not to.

As for their final statement, "where de-transition occurs", this is Turban doing his USTS15 data dredging thing again. A survey where current transgender identity was required, recruited via advocacy groups, all the usual issues.

Very obviously, if you ask people who currently identify as trans why they detransitioned, you will only get one type of answer. It's not serious work to exclude people who used to identify as trans but no longer do from a survey about detransition.

Nor can you make any general statements about detransition based on such poor data. Again, this is something these people will know for every area other than trans - until they disagree with it.

This is something we saw with tandfonline.com/doi/full/10.10… - suddenly trans activist researchers are well aware of sampling bias, applicability, causality and general limitations. It becomes very transparent all of a sudden.

When dealing with the kind of statements they write, it is very similar to dealing with Gish gallops (or 'the asymmetry of bullshit'), so I decided to focus very narrowly on these two sentences and show how they are broken not just in small ways, but fundamentally so.

Every single paragraph can be drilled into to this depth, it is simply a matter of already being deeply familiar with the literature or investing a lot more time in debunking it than they did in writing it. I am generally in favor of meeting bad ideas with good ideas, however...

...with these organizations, and AusPATH in particular, it definitely becomes a question about how many times you engage with bad faith actors. And even when you dismiss them, it is important to be careful and thorough, and continue to give them the chance to improve.