MAPS The Blinding Problem

[any major dude]

This has been something of an issue in psychedelic research since it's inception. It's clearly obvious to the subject, and often obvious to the researchers who has gotten placebo and who hasn't.

There have been a number of different methods to attempt to counter this issue, but they are not without their problems as well. Active placebo (a smaller dose of the substance being studied) has become fairly common, but I worry this may offer some therapeutic effect in the "control" group, thus showing lower efficacy in the experimental group. Other substances like methylphenidate & niacin have been used in the past as well, but I'm unsure as to how much less obvious this is than placebo...

I'm at a bit of a loss for other ideas at the moment.,definitely curious as to your thoughts on the matter.

 

[Reverend Random]

I really think there's no obvious way out of this. It seems probable to me that anything you might gave a patient in these trials that makes it harder for both subject and scientist to assess which substance they've been given, might also qualitatively influence the outcome of the control group.

To me it appears that what's happening is that psychedelics are showing the limits of the placebo-controlled model for clinical trials. Which is somehow pretty unfortunate, as I think that psychedelic research has more of a fair chance if it didn't continually challenge the current scientific model at some level, instead of fitting in nicely with the mainstream models.

You have to ask yourself how bad it really is when it becomes pretty obvious who's taken the active drug and who hasn't. But because that's the best we can presently do in terms of showing efficacy of new treatments and medication, that's what we have to work with...

I agree with any major dude that giving an active placebo is also problematic, by the way. Same goes for niacin etc. I remember this great quote from Pahnke's Marsh Chapel Experiment where one subject thought he'd been given psilocybin until he heard his neighbor talk about the blinding light of the candles. :D Let me see if I can find that somewhere.

 

[any major dude]

It is pretty tricky. And yes, RCTs do have their limits, though I think it's necessary for psychedelics to show value via this method to be accepted as medicine.

Another possible avenue would be to have the therapy be the experiment, i.e. use either no therapy or sham therapy in a control group. Granted this has it's limitations as well, but could still provide useful data.

 

[Reverend Random]

It is pretty tricky. And yes, RCTs do have their limits, though I think it's necessary for psychedelics to show value via this method to be accepted as medicine.

Yes, I agree.

Another possible avenue would be to have the therapy be the experiment, i.e. use either no therapy or sham therapy in a control group. Granted this has it's limitations as well, but could still provide useful data.

Well, unfortunately that's also tricky. Most of the clinical studies that are being with done psychedelics are allowed for patients who've shown to be resistant to all other conventional interventions. I think it'd be pretty unethical to give some other patients sham therapy. No therapy at all wouldn't really be a proper control, since psychedelics are typically used as an adjunct to (psycho)therapy, not simply as a medicinal intervention.

I know I'm only pointing out difficulties, but I think it's a pretty difficult problem to solve, because psychedelics and RCTs just don't seem like such a perfect match.

 

[Reverend Random]

This was the quote from Pahnke's dissertation btw.

Nothing much more happened and within another 40 minutes, 45
minutes, everybody was really quiet and sitting there. Y.M. was
sitting there and looking ahead, and all of the sudden T.B. says to
me, "Those lights are unbelievable." And I said, "What lights?" He
says, "Look at the candles." He says, "Can you believe that?" And I
looked at the candles, and I thought, 'They look like candles." He
says, "Can't you see something strange about them?" So I remember
squinting and looking. I couldn't see anything strange. And he says,
"You know it's just spectacular." And I looked at Y.M. and he was
sitting there saying, "Yeah." And I thought, "They got it, I didn't."

(Taken from Doblin 1991)

 

[B9]

You could attempt to interview the subjects by using a machine - if you've taken a decent dose this may prove problematic - but researchers not interacting with the subject at all would cut down the possibility of inadvertently influencing the subject in some way.

If your subject is of reasonably robust mental health there's no particular reason a professional needs to sit & hold hands with them.
If the session was intended for therapy then probably it'd be difficult/unwise to replace the human element, but if experiments are conducted purely for research into the effects of psychedelics then I'd be happy to be alone in a comfortable environment.
Perhaps the subject ought to be given the opportunity to explain in their own words what they feel about the experience - maybe questionnaires can alter peoples ability to correctly convey their experience & instead they attempt to respond to what they believe is being asked of them framing their responses in the type of language used in the questions.

 

[desertshore]

Uh.. hope it's not too much complicated, but you could maybe do something like this:
two experimenters (therapeuts?) are introduced to the experimental subjects and conduct the pre-treatment part of the study together (assessment, clinical interview, whatever). Only one experimenter is present during the pharmacological session (s) with a given patient, only the other experimenter will be present during follow-up sessions and post-treatment testing. Counterbalance experimenters between subjects. Maybe recruit enough subjects so that you have enough statistical power to check the effect of the different experimenters in the phases of the experiment.

Also, there are clinical interviews that can be recorded and then trascripted/rated/validated by other people (who can be completely naive to the experiment).

If this is still not good enough... state in your final reports/paper what you did, try to compute meaningful effect sizes and make an educated guess about how much those could be overestimated because of the flaws in the experimental design. After all, ad impossibilia nemo tenetur

 

[AJsocket]

What about this: Tell the clinical observers that there are two groups, one that is getting the treatment and one that is getting the placebo. However, only actually give the treatment or placebo to 30% of each group. The other 70 percent of each group will be actors who have been trained to act in the same way and also don't know who the actors are and who got the drug. Then you have the clinical observers rate each person individually.