Tavistock puberty blocker study published

OldCrone

In the end section of the paper they said they lost some patients’ data. Could that be the reason ?

Thanks for that link NotBadConsidering. I knew I'd seen something about this on here before. This is another paper about this case - this one is open access.

jme.bmj.com/content/medethics/46/11/743.full.pdf

pediatrics.aappublications.org/content/145/2/e20191606

I don’t for one minute believe this is theoretical, I think they’re prospectively protecting themselves for a case they have who is probably 15-16 and doesn’t want CSH.

It’s grotesque.

I'm thinking now that this is because at the time of this report in 2015, it's possible that the other 14 hadn't yet had a full year on puberty blockers, so their results wouldn't have appeared here, neither for the SRS test nor for the 'psychological functioning' test. I don't think it's because some of them have 'aged out' of the study before they had had a year on puberty blockers, because if that were the case, their T0 values would still appear.

I think that the comment from my earlier post is still true, that of the children whose SRS scores were given for both T0 (start of trial) and T1 (after a year on puberty blockers), the number (as well as the percentage) of children who showed autistic traits has risen:

Normal:
T0: 53.3% (representing 16 children out of 30)
T1: 41.4% (12 out of 29)

Mild to moderate:
T0: 30.0% (9 out of 30)
T1: 37.9% (11 out of 29)

Severe:
T0: 16.7% (5 out of 30)
T1: 20.7% (6 out of 29)

This must be the same cohort of children, since it would be nonsensical to state the percentage of one group of children with autistic traits at the start of the trial, and the percentage of a different group of children with autistic traits after a year on puberty blockers.

That’s a good Maya thread.

From a v good thread by MF.

twitter.com/mforstater/status/1338096822947024899?s=21

MissLucy

I think they thought of themselves as brave medical pioneers and expected to get a pat on the back for their ground-breaking work. They probably weren't too bothered about statistical integrity.

This is exactly why we have a regulatory system.

It may not be the clinicians’ (main) day job to design clinical trials and write scientific papers but it is what the regulatory system and those who work or volunteer within it specialise in.

OldCrone

I was wondering exactly how many children moved into the 'mild to moderate' or 'severe' categories after a year on puberty blockers, since these numbers were higher than at the start of the trial.

I don't think that children moved into the more severe category. What's happened is that children age out of the study at different rates, because they are not followed up beyond age 16.

All the participants are scored in after 12 months, but only those who are still under 16 are re-scored at 24 and 36 months.

What this suggests is that children who started PBs at a younger age (14 year-olds and younger) scored higher on the SRS than older adolescents (age 15). Which doesn't strike me as particularly encouraging.

If they realised it was rubbish, I doubt Russell Viner would have put his name to it. I think they thought of themselves as brave medical pioneers and expected to get a pat on the back for their ground-breaking work. They probably weren't too bothered about statistical integrity.

They should never have approved it and I think there is a legal case against the HRA.

Isn’t it the research team and sponsor’s job to make the research fair, capable of finding out about its proposed subject etc? I thought the people and organisations listed on the published paper are responsible. I notice this paper hasn’t been peer reviewed so hopefully those ‘peers’ will highlight all the research design or analysis problems when that is done?

I've just been looking again at the table about autistic traits in the trial participants from the board meeting in 2015 (this isn't covered in the newly-published paper). I was wondering exactly how many children moved into the 'mild to moderate' or 'severe' categories after a year on puberty blockers, since these numbers were higher than at the start of the trial.

16.7% scored as 'severe' on the SRS test at the start of the trial. 16.7% of 44 is 7.3. I would expect this to be a whole number representing the number of children in that category.

After a bit of number crunching, it looks as though only 30 of the children took this test at T0 (the start of the trial) and 29 at T1 (after a year on puberty blockers), since this is the only way to get whole numbers of children for all the percentages they state.

They clearly state on that page that the number of children is 44, so I assumed all of them would have been tested. Why didn't they test the other 14 (15 after 1 year)? Is it because they already had an autism diagnosis, so they felt it was unnecessary? What else could be the reason? They don't even mention that not all the children were tested (or not all their scores were recorded).

It's on page 50 of the board meeting document (Appendix 7) if anyone else wants to have a look. It's possible that I've misinterpreted something.

tavistockandportman.nhs.uk/documents/142/board-papers-2015-06.pdf

Not sure what you’re referring to gardenbird but apparently this was an unfunded study.

I think some very hardworking clinicians were pressured into doing a clinical trial they didn’t want to run. It was unfunded, underpowered and uncontrolled.

I hope I am wrong in thinking that this was the trial that a Gender clinic was paid £1.3m to run?

Yes, (re hardworking clinicians being pushed into things that they didn't want to do) and I imagine that the huge staff turnover will have had an effect as well. And the ongoing pressure from external agencies to find certain outcomes maybe? Or at least a reluctance to publish something that would bring more pressure?

They mention the statistical plan but have not included it.

This is really not helping with the transparency of such a high profile piece of research which has already come in for so much criticism and controversy. And is so delayed in its reporting.

I think some very hardworking clinicians were pressured into doing a clinical trial they didn’t want to run. It was unfunded, underpowered and uncontrolled.

Ultimately, due to the poor quality of data, it is unusable.

I agree that it reads like a cry for help at times.

But, someone went to the trouble of getting it through an alternative ethics committee, when the first one rejected it.

And that ethics committee approved it.

They should never have approved it and I think there is a legal case against the HRA.

Years ago, back in the mid-1990s when I was a teenager, I remember reading an article about a child who had been born female but who had severe gender dysphoria and wanted to be a boy called Fredd. The article would have probably been in either Sugar or J17 (neither of which exist anymore, afaik). I remember reading the article at the time and I'm sure the child interviewed mentioned the prospect of going on puberty blockers. I remember Fredd saying something like "I want to go on them because the last thing I want is for my body to start developing into that of a woman." It was the first time I had ever heard about puberty blockers.

I've since tried to trace the article online but have been unable to. However, I did find a brief reference to a TV documentary featuring a transgender child called Fredd that was aired around 1995 or 1996. I suspect the article I read followed the TV show. I haven't been able to find anything else about Fredd since then.

Anyway, my point is that clearly PBs have been used for cases of gender dysphoria for some time (unless my memory of the article is playing massive tricks on me). If kids like Fredd were put on these blockers back in the mid-90s, surely we must have some long term data by now?

Of course, since I can't find any follow-up information on Fredd, it is possible that Fredd desisted and never medically transitioned at all.

It gets worse. From strengths and limitations section of the paper:

“ Scoring of psychological questionnaire data was rechecked at the completion of the study however this was not possible in very small numbers of participants in whom only scale scores rather than individual item data were preserved during data migration in hospital clinical information systems.”

sultana fair point on criteria for SAE.

What about it being reported as an adverse event?

From discussion section of the paper:

“Psychological distress and self-harm are known to increase across early adolescence. Normative data show rising YSR total problems scores with age from age 11 to 16 years in non-clinical samples from a range of countries.[29] Self-harm rates in the general population in the UK and elsewhere increase markedly with age from early to mid-adolescence, being very low in 10 year olds and peaking around age 16-17 years.[53-56] Our finding that psychological function and self-harm did not change significantly during the study is consistent with two main alternative explanations. The first is that there was no change, and that GnRHa treatment brought no measurable benefit nor harm to psychological function in these young people with GD. This is consonant with the action of GnRHa, which only stops further pubertal development and does not change the body to be more congruent with a young person’s gender identity. The second possibility is that the lack of change in an outcome that normally worsens in early adolescence may reflect a beneficial change in trajectory for that outcome, i.e. that GnRHa treatment reduced this normative worsening of problems. In the absence of a control group, we cannot distinguish between these possibilities. We aimed to use normative reference data to examine this issue. However age- and gender-standardised t-scores for ASEBA and other outcomes cannot answer this question as they cover a very broad age range (e.g. 12-18 years). We had anticipated that z-scores on the YSR available by calendar year for two comparable countries (Netherlands; Australia) might be informative however confidence intervals were too wide to draw reliable inferences.“

My translation for the lay person ( I welcome any corrections or comment):

They acknowledge that because they didn’t run a controlled study, they don’t know whether the drug had any impact at all on psychological distress and self harm.

They planned to compare the data to known data for the general population however they were unable to.

I would like to suggest that their chosen method of measuring this data (using median scores and combining two separate questionnaire answers) was always going to lead to an issue comparing to the published data for the general population.

WHY did they test an experimental treatment on young children, if not to reduce those children’s’ psychological distress??? They failed.

The one thing in clinical trials which is worse than a bad outcome, is an outcome which you can’t interpret or explain.

They failed.

everythingthelighttouches

1) this is a clinical trial. You are supposed to plan your outcome measures and how you will calculate them in advance in the statistical plan for the trial documents. if you change the plan part way through a trial you have to issue an amendment and rationale. All very normal... the purpose is To avoid data manipulation or questions about data manipulation. I wonder whether we can get these docs?
I agree. It appears it was never a particularly rigorous trial from the outset. I suspect the initial protocol didn't go into these details.

2. Do we have the raw data? Can we see if the mean scores for these two suicide questions were statistically significantly different at the later time points??
Everything I've seen I've posted in this thread. There is a lot more about the interim data than the final data. You can probably work out the raw data from looking at the table.

3. Why THE HELL didn’t they report an increase in their own suicide measures to the regulator as an Serious adverse event (SAE) at 12 months??
Ultimately it's the decision of the investigator whether a safety signal constitutes an serious adverse event or not. I think you could make an argument that it doesn't meet the criteria below.

Serious Adverse Event (SAE)
Any untoward medical occurrence that at any dose:

• results in death,
• is life-threatening,
• requires inpatient hospitalization or prolongation of existing hospitalization,
• results in persistent or significant disability/incapacity,

• is a congenital anomaly/birth defect

Thank you so much everything and sultana that’s very kind of you to give links.

ChattyLion

You could have a browse around the ICH Guidelines. These are internationally agreed and then the corresponding signed- up regulatory agencies (US, EU, Japan, others...) incorporate them into their own guidance
www.ich.org/page/efficacy-guidelines

E6 Good Clinical Practice (GCP) is an important one

E9 Statistical Principles for Clinical Trials also addresses many of the issues we've been discussing on this thread

This guidance (as well as that of the region and countries) must be followed in order to get new drugs, and new indications, approved.

In this case no new drug licence is being sought. The drugs are being used off-label for an unlicensed indication. This in itself is not a problem, off-label prescribing is fairly common especially in children.

However had anyone sought to make treatment of gender dysphoria a licensed indication for PBs, they would have needed to generate a substantial evidence package demonstrating that the drug worked, that it had an acceptable safety profile, a favourable benefit-risk assessment and much else.

On bone mineral density. My reading of the methods is that they did take account of age and sex adjusted scores:

“BMD z-scores for age and birth-registered sex appropriate to this machine were calculated.[26]”

I think what would’ve happens/is happening is that GIDS screen for ASD traits, it the score is high then they refer back to the local ASD assessment services for a recommendation for a full assessment. The local services will then decide whether to proceed with the assessment. It won’t be linked to the GIDS treatment. If the patient, or their parents don’t want an assessment then they won’t get one. Local services are stretched, waiting lists often over a year or longer.

You’ve then got the added complication of being in two systems for two different things.

With the push by vociferous voices for quicker intervention, they’re not going to wait for 1 year plus for an ASD diagnosis.

The gold standard should be that GIDS is linked to an expedited ASD assessment service. Though that will never be commissioned due to gender dysphoria now being classed as a stand alone condition even though it can be linked to some mental health conditions and ASD ridged thinking.