Cass Report

[Emily's Cat]

Out of curiosity, has anyone here ever been in an RCT as a subject?

I have never had a condition that was bad enough for me to risk participation.

 

[theprestige]

Is it just me, or is the lack of "gold standard" research yet another damming nail in the coffin of trans affirming science?

 

[Emily's Cat]

Precisely this. The act of randomly assigning patients with a specific condition into groups which are not self-selected is part of what makes RCTs much more powerful than other forms of studies.

Randomization serves the purpose of being able to measure the with and without cohorts and calculate the differential between them. Blinding specifically controls for placebo effects.

There are other ways to do testing and trials, but the error bounds on results are necessarily larger and any conclusions will have lingering caveats.

As a side note - randomization only works if your testing population is large enough to ensure that any correlated factors are mitigated. For example, if your population is 10 people and you randomly assign 5 to control and 5 to test... the likelihood of there being a statistically significant difference in something like age or sex or geographic location or income is very high - which means that those factors can't be ruled out as contributors to the measured outcome. The sample populations have to be large enough that the sample statistics of non-tested measures can be shown to be significantly the same. Without that, you end up doing semi-random testing, with defined cohorts on each potentially correlated measure and then you do random assignment from within each granular cohort. You might end up with a quota for each 5-year age bracket, sex, and MSA for example.

 

[Ziggurat]

Is it just me, or is the lack of "gold standard" research yet another damming nail in the coffin of trans affirming science?

The lack of high quality research is certainly a big problem for advocates of medical transition. I'm not sure about "gold standard", though. That's often taken to mean double blinded randomized placebo trials. But you can't effectively do that with trans treatment. I mean, there's obviously no possible placebo for cosmetic surgery. But even with hormone treatments, if you're getting the actual hormones, you're going to know pretty quickly. If you're just getting the placebo, you're going to figure it out shortly as well. So there's really no possibility of blinding, and no possibility of placebo. Even if you just wanted to do a randomized trial (no placebo, no blinding), I don't know how you would actually get people to sign up. Nobody is going to want randomized treatment, they're going to want whatever treatment they're already convinced is the best. So that really hamstrings what's even possible to do with a study.

But even within that framework, there are better studies and worse studies. If your timeline is really short, your data is worthless. If your sample size is really small, your data is worthless. If a lot of your participants drop out and you don't follow up with them, your data is worthless. So my above statements shouldn't be taken as letting researchers off the hook for bad study design. But in fairness, we can't really do the kinds of studies that would produce the most robust possible data.

 

[Emily's Cat]

Since the salient impact of trans medicine is supposed to be mental health related, there are two types of studies that would be reasonable and appropriate - and would likely still be able to qualify as high quality.

The first would be a longitudinal study. You'd need a sufficient number of starting participants, which would probably require several years of initial collection. Each participant's mental state would need to be documented with some standardized measure across associated variables prior to the beginning of any treatment, and they'd also need to collect any comorbidities or other potential causal factors for those mental health issues. For example, if a participant was autistic, they'd need to capture that as a potentially correlated or conflating factor. Similarly a history of childhood abuse or sexual assault could be a potential cause of their mental distress. Any treatment modalities would be documented throughout the process. Then they'd need to do repeated follow-up assessments of mental state over a long period of time.

One of the weaknesses of many studies done so far is that they have a very short follow-up measure. They tend to report great improvements in mental state... but that's immediately following the treatment. It's not at all surprising that when a distraught kid says "I'm super sad because I want to be a male, but I'm a female", then you give them testosterone, they're going to have an observed improvement in mood because their belief is being validated and they're experiencing affirmation of their belief. But give it a few years, and anecdotally the initial mental health distress returns - because it wasn't actually caused by being transgender, rather their gender identity incongruence was caused by some other underlying and still-untreated factor.

The second would be a matched cohort study (I don't know if this would be the right term for this context). In that situation, you wouldn't necessarily be declining treatment for individuals. Rather, you'd set up side-by-side assessment and monitoring for two different but similar populations. In this case, you'd want a group of patients with mental distress seeking transgender interventions, and you'd want another group of patients with similar mental distress patterns - similar rates of childhood abuse, sexual assault, age, sex, geography, etc. - who are NOT seeking transgender interventions. Then you'd compare starting mental state and changes in state over time. This would likely be more challenging to arrange - you'd essentially want to do repeated assessments for pretty much anyone presenting with a cluster of correlative conditions across the board. Then you mine that data after the fact and create matched cohort pairings to determine whether the treatment was more or less effective than other treatments like talk therapy or antianxiolytics or whatever else.

I've done the second type of study, albeit not in a clinical study. I've reviewed longitudinal before/after studies in the context of disease management outcomes and cost effectiveness, but I've never performed the studies myself.

 

[angrysoba]

Since the salient impact of trans medicine is supposed to be mental health related, there are two types of studies that would be reasonable and appropriate - and would likely still be able to qualify as high quality.

The first would be a longitudinal study. You'd need a sufficient number of starting participants, which would probably require several years of initial collection. Each participant's mental state would need to be documented with some standardized measure across associated variables prior to the beginning of any treatment, and they'd also need to collect any comorbidities or other potential causal factors for those mental health issues. For example, if a participant was autistic, they'd need to capture that as a potentially correlated or conflating factor. Similarly a history of childhood abuse or sexual assault could be a potential cause of their mental distress. Any treatment modalities would be documented throughout the process. Then they'd need to do repeated follow-up assessments of mental state over a long period of time.

One of the weaknesses of many studies done so far is that they have a very short follow-up measure. They tend to report great improvements in mental state... but that's immediately following the treatment. It's not at all surprising that when a distraught kid says "I'm super sad because I want to be a male, but I'm a female", then you give them testosterone, they're going to have an observed improvement in mood because their belief is being validated and they're experiencing affirmation of their belief. But give it a few years, and anecdotally the initial mental health distress returns - because it wasn't actually caused by being transgender, rather their gender identity incongruence was caused by some other underlying and still-untreated factor.

The second would be a matched cohort study (I don't know if this would be the right term for this context). In that situation, you wouldn't necessarily be declining treatment for individuals. Rather, you'd set up side-by-side assessment and monitoring for two different but similar populations. In this case, you'd want a group of patients with mental distress seeking transgender interventions, and you'd want another group of patients with similar mental distress patterns - similar rates of childhood abuse, sexual assault, age, sex, geography, etc. - who are NOT seeking transgender interventions. Then you'd compare starting mental state and changes in state over time. This would likely be more challenging to arrange - you'd essentially want to do repeated assessments for pretty much anyone presenting with a cluster of correlative conditions across the board. Then you mine that data after the fact and create matched cohort pairings to determine whether the treatment was more or less effective than other treatments like talk therapy or antianxiolytics or whatever else.

I've done the second type of study, albeit not in a clinical study. I've reviewed longitudinal before/after studies in the context of disease management outcomes and cost effectiveness, but I've never performed the studies myself.

I don’t know how you could ethically do that study since the control group would require treatments that their doctors see fit according to their need and not for the purpose of the clinical trial. The control group is therefore unlikely to be a have all of the other factors controlled for. For example, say that there is a form of therapy that happens to be very useful such as CBT or more simply an intervention which is better for that patient such as removal from their environment, then while that treatment is good for them it does not mean it would be as effective at treating gender dysphoria. It might be specific to the issues that individuals in the control group have.

 

[theprestige]

I don’t know how you could ethically do that study since the control group would require treatments that their doctors see fit according to their need and not for the purpose of the clinical trial. The control group is therefore unlikely to be a have all of the other factors controlled for. For example, say that there is a form of therapy that happens to be very useful such as CBT or more simply an intervention which is better for that patient such as removal from their environment, then while that treatment is good for them it does not mean it would be as effective at treating gender dysphoria. It might be specific to the issues that individuals in the control group have.

Which raises many questions about the purported science that recommends social transition and gender affirming care for minors.

If the necessary science can't be done ethically, then what are we supposed to advocate in good conscience? Preferred pronouns? Not supported by ethical science. Transcending sex segregation with fiat self-ID? Not supported by ethical science.

The most notable thing about the Cass Report is that there is no countervailing report that indicates social or medical transition as a scientifically ethical approach.

 

[angrysoba]

Which raises many questions about the purported science that recommends social transition and gender affirming care for minors.

If the necessary science can't be done ethically, then what are we supposed to advocate in good conscience? Preferred pronouns? Not supported by ethical science. Transcending sex segregation with fiat self-ID? Not supported by ethical science.

The most notable thing about the Cass Report is that there is no countervailing report that indicates social or medical transition as a scientifically ethical approach.

Not disagreeing with that.

 

[Emily's Cat]

I don’t know how you could ethically do that study since the control group would require treatments that their doctors see fit according to their need and not for the purpose of the clinical trial. The control group is therefore unlikely to be a have all of the other factors controlled for. For example, say that there is a form of therapy that happens to be very useful such as CBT or more simply an intervention which is better for that patient such as removal from their environment, then while that treatment is good for them it does not mean it would be as effective at treating gender dysphoria. It might be specific to the issues that individuals in the control group have.

I assume you're talking about a matched cohort study?

The entire point of matched cohort is that it doesn't have a control group. It doesn't compare treated to non-treated. It compares efficacy of a particular type of treatment against all other treatments. Having a non-treated control group allows for evaluation of placebo effect in addition to efficacy. Matched cohort doesn't even try to evaluate placebo.

Let me try to give you an example. The study I did was looking at medical claims cost for employer groups who had a particular benefit integrated in their coverage, versus those that had the benefit provided by a different insurer and not integrated. I couldn't very well tell some groups that they weren't allowed to have the benefit, nor could I blind it in any fashion. What I could do, however, was to take all of the historical claims data for both types of groups over a long period of time. Then I could classify each individual in both groups on the types of characteristics that are known to affect claims - age, sex, diagnosis groups, location, etc. Then I ran tens of thousands of random samples of matched pairs from both groups, and compared the resultant average claims costs that resulted from the with and without groups. Basically, the simulation would randomly select an individual from the with group, then would go find all of the individuals in the without group that had the same age, sex, diagnosis set, and lived in the same location and randomly pick one of those.

 

[jt512]

The lack of high quality research is certainly a big problem for advocates of medical transition. I'm not sure about "gold standard", though. That's often taken to mean double blinded randomized placebo trials. But you can't effectively do that with trans treatment. I mean, there's obviously no possible placebo for cosmetic surgery. But even with hormone treatments, if you're getting the actual hormones, you're going to know pretty quickly. If you're just getting the placebo, you're going to figure it out shortly as well. So there's really no possibility of blinding, and no possibility of placebo. Even if you just wanted to do a randomized trial (no placebo, no blinding), I don't know how you would actually get people to sign up. Nobody is going to want randomized treatment, they're going to want whatever treatment they're already convinced is the best. So that really hamstrings what's even possible to do with a study.

There would be no need for a placebo arm because you would be comparing two treatments: medical vs. non-medical (eg, counseling). Also, given that the option of counseling exists, a placebo control would be unethical.

I agree that you couldn't ethically or practically randomize people who aren't seeking medical treatment to the medical arm of the trial, but you could ethically and practically randomize people seeking medical treatment to the other arm. In fact, this is a routinely practice in clinical trials. This design is especially appropriate in jurisdictions (eg, the U.K, as I understand it) where medical transitioning for minors is only permitted for participants in clinical trials. For a patient who wants the medical treatment, enrolling in the trial and having a 50% chance of getting it is better than not enrolling in the trial and having a 0% chance of getting it.

The second [observational study design] would be a matched cohort study (I don't know if this would be the right term for this context). In that situation, you wouldn't necessarily be declining treatment for individuals. Rather, you'd set up side-by-side assessment and monitoring for two different but similar populations. In this case, you'd want a group of patients with mental distress seeking transgender interventions, and you'd want another group of patients with similar mental distress patterns - similar rates of childhood abuse, sexual assault, age, sex, geography, etc. - who are NOT seeking transgender interventions. Then you'd compare starting mental state and changes in state over time. This would likely be more challenging to arrange - you'd essentially want to do repeated assessments for pretty much anyone presenting with a cluster of correlative conditions across the board. Then you mine that data after the fact and create matched cohort pairings to determine whether the treatment was more or less effective than other treatments like talk therapy or antianxiolytics or whatever else.

I assume that by "mental distress," you mean gender dysphoria, since the scientific question that the study would aim to answer would be, what is the better treatment for gender dysphoria, not just any mental distress.

 

[Elaedith]

Unfortunately, it isn't clear that they are going to use randomisation in the proposed trial of puberty blockers announced by the NHS. Although I don't think the protocol is released yet, all the descriptions I have seen suggest they intend to follow a group that self-selects to take puberty blockers and a group that does not, without random assignment. This approach won't adequately address some key questions, including how often gender dysphoria spontaneously resolves after going through puberty without transitioning but with support (ie 'watchful waiting') and what factors predict this.

 

[Emily's Cat]

I assume that by "mental distress," you mean gender dysphoria, since the scientific question that the study would aim to answer would be, what is the better treatment for gender dysphoria, not just any mental distress.

No, not gender dysphoria. Anxiety, depression, and suicidality.

I get where you're going - the issue is that there seems to be a misattribution of cause in a whole lot of cases, especially cases where the individual is a young female or is autistic. What seems to happen is that a young person shows up with a complaint of being depressed, anxious, suicidal, or otherwise emotionally distressed. That young person then tells the counselor that their emotional distress is because they're trans.

The core question to me is whether or not trans is the actual causal factor, or whether trans is a coping mechanism.

A large portion of the people who have detransitioned have reported that they were convinced that the source of their distress was that they must be trans... but transitioning didn't actually fix anything. After having transitioned, they had a relatively short period of relief, but the distress returned. And after having matured some and recognizing that they weren't any better off, they realized that they were never actually trans in the first place. For them, trans was a way to cope with the source of their distress, it was something that was easy to point to and say "oh that must be why" and avoid actually dealing with their issues.

 

[theprestige]

I spent most of my life thinking I was suffering from clinical depression.

Now I'm taking anti-anxiety meds, I'm pretty sure depression was never the problem.

 

[jt512]

Unfortunately, it isn't clear that they are going to use randomisation in the proposed trial of puberty blockers announced by the NHS. Although I don't think the protocol is released yet, all the descriptions I have seen suggest they intend to follow a group that self-selects to take puberty blockers and a group that does not, without random assignment. This approach won't adequately address some key questions, including how often gender dysphoria spontaneously resolves after going through puberty without transitioning but with support (ie 'watchful waiting') and what factors predict this.

It seems to me that following patients who elected not to take puberty blockers would provide pretty good answers to those questions. The question it won't answer with much confidence is the counterfactual: How would the group that took puberty blockers have faired had they not done so?

Perhaps the NHS's thinking is that there is too much evidence of harm of puberty blockers to conduct an RCT at this tine. Perhaps if the observational study shows otherwise, the NHS follow up with an RCT.

 

[Emily's Cat]

I spent most of my life thinking I was suffering from clinical depression.

Now I'm taking anti-anxiety meds, I'm pretty sure depression was never the problem.

My sister was on antidepressants for several years before they finally figured out they were bipolar. I'm glad they've got you sorted

 

[novaphile]

I don’t know how you could ethically do that study since the control group would require treatments that their doctors see fit according to their need and not for the purpose of the clinical trial. The control group is therefore unlikely to be a have all of the other factors controlled for. For example, say that there is a form of therapy that happens to be very useful such as CBT or more simply an intervention which is better for that patient such as removal from their environment, then while that treatment is good for them it does not mean it would be as effective at treating gender dysphoria. It might be specific to the issues that individuals in the control group have.

I was part of a RCT for an additional asthma treatment.

(I have no idea what the additional treatment was, but I was given pills to take twice per day in addition to the existing treatment regime I was on.)

The study lasted for about eight weeks, and I was called (weekly IIRC) to complete a telephone survey about symptoms.

Since then, I've been approached about six times, over about 15 years, to see if I was willing to participate in other studies.

So far, I have always failed the entry criteria, because my symptoms are well managed by the existing treatment.

RCTs in that format, appear to deal with the ethical dilemma of withholding treatment.

I have a suspicion that withholding treatment is never an option in a drug trial.
(Unlike TV shows where the volunteers die because life-saving magic is denied to them.)

 

[Elaedith]

It seems to me that following patients who elected not to take puberty blockers would provide pretty good answers to those questions. The question it won't answer with much confidence is the counterfactual: How would the group that took puberty blockers have faired had they not done so?

If patients choose whether or not to take puberty blockers, and amongst those who don't a majority desist after puberty (as shown in every study ever conducted prior to the use of the affirmative-approach), while almost all who take them progress to cross-sex hormone treatments (as shown in numerous studies), activist clinicians will say that this proves the patients knew whether or not they were 'really trans' and this was reflected in their decision on whether or not to take them. If the group not taking blockers includes some who wanted them but were advised against it and some of these desist, this will be taken to show that clinicians and/or parents can predict persistence.

Perhaps the NHS's thinking is that there is too much evidence of harm of puberty blockers to conduct an RCT at this tine. Perhaps if the observational study shows otherwise, the NHS follow up with an RCT.

I'm not clear how allocating some patients who want blockers to an alternative treatment would be less acceptable if puberty blockers are thought to be harmful? However, I suspect an RCT would be difficult to conduct in practice because there has been so much propaganda pushing medical transition as essential, safe and effective, that patients not assigned to get them might abandon the study.

 

[Emily's Cat]

I have a suspicion that withholding treatment is never an option in a drug trial.
(Unlike TV shows where the volunteers die because life-saving magic is denied to them.)

Most clinical trials are either testing a new treatment for a condition that hasn't had a treatment available, or they're testing an alternative/additive treatment.

A lot of cancer trials, for example, are testing new treatments. Their participants are often terminal patients. In that case, some get the experimental treatment while others get a placebo. When it's experimental, that's kind of the whole point of the trial - we don't know if it's going to work at all. Sometimes the trials fail and the treatment is deemed ineffective. Sometimes the side effects are worse than no treatment. Sometimes you get boner pills.

 

[jt512]

If patients choose whether or not to take puberty blockers, and amongst those who don't a majority desist after puberty (as shown in every study ever conducted prior to the use of the affirmative-approach), while almost all who take them progress to cross-sex hormone treatments (as shown in numerous studies), activist clinicians will say that this proves the patients knew whether or not they were 'really trans' and this was reflected in their decision on whether or not to take them. If the group not taking blockers includes some who wanted them but were advised against it and some of these desist, this will be taken to show that clinicians and/or parents can predict persistence.

So we shouldn't make medical policy based on science, not ideology.

I'm not clear how allocating some patients who want blockers to an alternative treatment would be less acceptable if puberty blockers are thought to be harmful?

I agree. If they design the study the was I suggested, namely, just among those seeking medical treatment.

However, I suspect an RCT would be difficult to conduct in practice because there has been so much propaganda pushing medical transition as essential, safe and effective, that patients not assigned to get them might abandon the study.

Could be. That's why I suggested that the study be conducted in a country where medical treatment was otherwise unavailable, which I thought, actually, was the case in the U.K.

 

[Elaedith]

Could be. That's why I suggested that the study be conducted in a country where medical treatment was otherwise unavailable, which I thought, actually, was the case in the U.K.

Puberty blockers are currently unavailable outside clinical trials in the UK either on the National Health Service or privately, but that is a recent change following the Cass report. Prior to that they were prescribed through the 'Gender Identity Development Service' (GIDS), who didn't keep proper records or do follow up. We had the same propaganda about them being safe and essential to prevent suicide, and watchful waiting is conversion therapy, as is pushed in the US. Obviously lots of people are unhappy about the ban.
I don't think the fact that there is no other way to obtain puberty blockers would necessarily stop patients randomised to the non-medical arm of an RCT from withdrawing from the study (or being withdrawn by parents) once they realise they are not getting the blockers.