May 25, 2001

12:38 AM

What if the 'placebo effect' is as unreal as a sugar pill?Danish researchers who have looked at 114 clinical trials involving placebos found "little evidence in general that placebos had powerful clinical effects. Although placebos had no significant effects on objective or binary outcomes, they had possible small benefits in studies with continuous subjective outcomes and for the treatment of pain. Outside the setting of clinical trials, there is no justification for the use of placebos." News links here and here.
posted by pracowity (13 comments total)

I don't mean to be flip, but my first response when I read this was, "No kidding! Sugar pills don't fix anything, who woulda thunk it?"

I've heard of placebo effects on minor pain. What other "clinical effects" were they looking for?
posted by RylandDotNet at 1:23 AM on May 25, 2001

The only thing that comes to my mind about this is: "well, duh!"
But I prefer being more vociferous than that.
posted by ZachsMind at 1:31 AM on May 25, 2001

Well I'm surprised. At least here in the UK there's quite a groundswell of opinion in support of "alternative" treatements, and I had presumed that this was because of placebo effects.

The authors claim that the (evidence for) placebo effect has never been rigorously evaluated, which surprised me. That they do find an effect for pain relief could be seen as evidence that they've got things right - it's perhaps the area where you'd expect palcebo to be particularly good.

Makes me wonder about a physio I dealt with who recommended a homeopathic treatment. At the time I thought she might be hoping for a placebo effect (since I'd actually thought that I reckoned it probably wasn't for me ;-) - maybe she just wanted me to be quiet while a rest did its work!
posted by andrew cooke at 2:51 AM on May 25, 2001

I wish they would use the term "Pez" for sugar pill.

By the way, Claritin makes me WAY more tired than a sugar pill. I'm more interested in whether side-effects are WAY underestimated based on what is now false-comparisons to "sugar pill."
posted by ParisParamus at 5:17 AM on May 25, 2001

Common thinking is that whether a placebo works or not is based on the individual and how the placebo is convincingly pitched / given to the patient. So when you show me a study that says placebos have little effect, my first question is, uhhh, how do you know the researchers weren't just ineffective in how they gave the placebo? We're talking about an effect that, to begin with, would be very difficult to test under clnical conditions.
posted by fleener at 5:36 AM on May 25, 2001

> "No kidding!" and "well, duh!"

A number of doctors -- and not necessarily the so-called Alternative Health loons -- think that people who believe they are being given real treatment often actually physically improve even when they are receiving no treatment other than a placebo. That's why it's news.

The Danish researchers say that their study shows little or no such effect, no significant physical change attributable to belief in the treatment. One study does not the truth make (or unmake); others will certainly run independent tests; but if further studies back up this one, I think it will make it harder for people to argue, for just one example, that having "the will to live" keeps people alive longer than they would have otherwise. Such beliefs are quite common.

> how do you know the researchers weren't just
> ineffective in how they gave the placebo

It was a study of more than a hundred different clinical trials that were presumably not chosen for their sloppiness; if the placebos were given improperly, it seems that everyone is doing it improperly.
posted by pracowity at 6:09 AM on May 25, 2001

There is a major difference between placebo treatment in a medical trail and placebo treatment in a clinical setting. Due to informed consent in medical trails, doctors have to inform patients they actually are at risk of receiving a placebo. In a clinical setting, if a doctor (for whatever reason) decides to give a patient a placebo treatment, the patient usually doesn't know it.

Receiving a placebo and knowing it, probably has a different (potential) effect than receiving a placebo and believing you're receiving a medical treatment.

Ignorance is a bliss...
posted by nonharmful at 6:59 AM on May 25, 2001

nonharmful, is there any consensus around the ethical use of placebo treatment in the clinical setting as you describe? Maybe I'm naive, but I thought that physicians pretty much can't be giving out placebos to patients who think they are getting something else.

Which isn't to say that they don't, of course. Is there a physician out there who has firsthand experience with this? As far as the "well, duh" reaction, the placebo effect is, according to my social psychologist partner, widely accepted as quite real, and she was very surprised to hear of evidence against the effectiveness of placebos as given in clinical trials.
posted by BT at 7:11 AM on May 25, 2001

BT, I've never prescribed placebo myself, although one could argue that prescribing fancy sounding salt-solutions for nasal congestion is a form of placebo treatment (or antibiotics for viral infections for that matter). However, some people with a 'vague complaint' don't leave the consulting-room without a prescription. Ethical? I don't think so. Practical? Absolutely.
The most powerful medicine is time, and 'believing that something is done' will speed this up. Not believing that a therapy works, decreases the likelihood of it actually working (example: paper on biofeedback: The mean magnitude of these previously nonspecified or placebo factors is closer to 70% when both the therapist and patient believe in the efficacy of the therapy ).

My point was that in research settings both the therapist and the patient don't believe the placebo works. They believe the tested drug is the effective therapy. Therefore, it's not a 'placebo treatment', but a comparison between a potentially effective drug and a placebo.
posted by nonharmful at 7:47 AM on May 25, 2001

I know a thing or two about clinical trials. Here's a greatly simplified example. Say you have new Drug X and 100 participants, all of whom have some medical condition (like an illness). Now, the idea is that 50 folks (Group A) will get Drug X, and 50 people (Group B) will get a placebo. The group getting the placebo is the Control Group. So if, at the end of the trial, 20 people in Group have a specific change in their illness (like: they have recovered from it) and 5 people in the Control group have the same change, then a clinician can say "Well, 5 people recovered with no treatment and 20 people recoved with treatment, so treatment affects roughly 15 people out of 50."

If the participants don't know which group they are in (i.e. if they are getting Drug X or ma placebo), the trial is saind to be "Blind". If both the participants *and* the clinicians don't know which is which (i.e. the guy giving out the pills dowsn't know if they are real or not) the trial is said to be Double-blind. Most FDA trials are double-blind trials.

Nonharmful, your comment "physicians pretty much can't be giving out placebos to patients who think they are getting something else" is true in clinical settings: they can't say to a participants "This *is* Drug X" unless they know that for a fact (or, at least, they're not s'posed to). But in clinical trials, the Informed Consent tells the participants they they may-or-may-not be actually getting Drug X,, and if the trial is double-blind the clinician doesn't really know anyhow.
posted by Shadowkeeper at 7:50 AM on May 25, 2001

What's astonishing is that they realized that the commonly-held figure of 35 percent was not backed up by rigorous research, and it took them some time to find out where it came from, the single 1950s paper that was the germ of all the later cites. Kudos to them for backing up their suspicions with a long-overdue study. Obviously this means there will have to be more studies done with a specific intent to measure the placebo (non-)effect, for instance, to judge the course-of-illness hypothesis of the authors.
posted by dhartung at 10:49 AM on May 25, 2001