Hacker News Reader

Comment by fshbbdssbbgdd | Hacker News Reader

I’m skeptical that a placebo control for MDMA can be valid. The effects of MDMA are pretty noticeable. The study actually asked participants if they thought they were on MDMA, and most answered correctly.

If the participants can tell, it somewhat defeats the purpose of having a placebo control group.

View on HN

If you start with people who know what MDMA feels like, I suspect you're sunk for the reasons you state.

But if for instance the placebo were niacin... niacin makes most people feel funny. If you don't know what MDMA or a niacin flush feels like, you might just know something is happening but not what.

And then you test people who don't experience niacin flush with or without a placebo to make sure that niacin isn't a therapeutic chemical.

Blinding isn’t just about the patients - It’s about the therapy providers as well. We would like to think they’d provide the same treatment regardless of what the patient received but over-eager researchers tend to get more excited about the active group and provide them more attention and enthusiasm.

Niacin isn’t a great placebo these days because it’s not the 70s any more and people going into these studies will probably Google what to expect from the drug. Also Niacin is short acting and makes people itchy and uncomfortable, so it’s not really compatible with therapy.

Isn’t that why many psychiatric experiments engage in some misdirection?

Talked to a therapist at the MAPS conference who said that the way you know a patient is in the trial is about halfway through their therapy session, they inevitably say, "But enough about me, how are YOU doing??"

That is the most MDMA thing ever.

My understanding is that in trials like this they’ll usually use a benzodiazepine as the control. Not psychedelic, but psychoactive enough that a drug-naive person wouldn’t know for sure what they were on.

This was not an active placebo study. The placebo group did not receive any medication.

Benzodiazepines would not be an appropriate active placebo anyway as they are a sedative, not a stimulant. They would also be expected to interfere with the therapy because they impair memory and alertness, so that’s a no-go.

It's a control. A baseline. That's the whole point. And especially with the therapy trials they're running, its not like it'll hurt people to have been given a placebo (unlike if you were to say, trial an HIV antiviral with a placebo)

Yeah this is going to be hard with all the psychoactive substances, but really we shouldn't discard placebo effect as "not real," especially for disorders that we already know are highly mediated by beliefs (about the self or the world). Reliably and persistently mutating those beliefs toward an actually improved life seems perfectly fine to me.

Right, but if you observe a high enough effect size, it’s in some sense irrelevant whether there was a placebo arm. If somehow there is a reliable d=0.4 placebo effect, then that is just as good as if the drug “really works”. To be clear on magnitudes, there was a something like 70% increase in the treatment effect vs. therapy+placebo.

Indeed what does it even mean for the cure to be “just in your head” or “a placebo” if condition is purely psychological, the trauma is resolved, and there is no remission? It’s a very different game vs. a mechanistic acting drug which either does or does not reduce cholesterol or whatever.

I would be surprised if that wasn't a glaring point for them, but you do control as best as you are able to.

This should be the second of two Phase III studies, which are pretty rigorous. Some drugs are just very hard to make a valid control for.

Some studies also use a drug placebo and a control group with a combination of interventions, or little to none at all, as best as I understand.

One can use statistics afterwards to help bolster claims of efficacy if a sham/placebo group is difficult to source.

This is my best understanding of the matter.

Rick Doblin has some talking points on this that I found fairly compelling... but now I can't find them. Basically that their trial is as good as you can do given that.

Why not use another group taking an amphetamine in addition to one taking an inert placebo?

As effective as that might be in terms of a proper active control group, it does not seem terribly ethical to give amphetamines to patients with PTSD without any medically sound reason to do so.

That and you’d have to correct for undiagnosed ADHD which suddenly improved with a course of amphetamines.

Reply on news.ycombinator.com