Critical Appraisal

[seren]

Does this belong here? I don't know.

Anyway, I am v. pleased because as part of my work I am going to be trained in critical appraisal as part of a group developing clinical guidelines. They've sent me a sample RCT and cohort study and (one assumes) they are going to teach us how to appraise them. Critically. Yay.

Anyway, I was being good and reading through the meeting papers in advance (rather than frantically thumbing through them at super-speed on the train on the morning of the meeting, through blurry pre-caffeine sleep-head, which I've never, ever done before :angel:). Suffice to say there's a little table outlining our role as guideline developers which says we're supposed to read the literature around various interventions for a set of related disorders, and compare their efficacy in prevention and management against other interventions, placebo, or nothing at all.

Now, I am new to all this (can you tell?) and I think I'm misunderstanding. What is the difference between the placebo and the nothing at all? Is it important to distinguish between people who think they're being treated, and people who just aren't?

[bobdezon]

well a placebo would be a seemingly beneficial result or feeling stemming from a belief in the efficacy of the proceedure/medication, and nothing at all could mean no discernable effect detectable? maybe?

[Mongrel]

This fact sheet about homeopathy (www.senseaboutscience.org.uk/pdf/SenseAboutHomeopathy.pdf) from the Sense about Science guys has a good, basic explanation about the placebo effect

[Fiona]

I think that there is no doubt that there are placebo effects and that it is important to test those separately from "nothing at all", which I take to mean just that: taking no action and giving no advice at all, but letting things take their course.

[Admin]

A placebo is different from the placebo effect.

When a medicine, lets say in pill form, is being trialled it will be given to 50% of the patients with the other half getting the placebo (dummy pill which should be indistinguishable from the real one).

Everyone should be blinded as to who's getting what until after the trial is over and then the two groups can be compared.

If the real treatment works then it will have given better results than the placebo comparison.

All treatments give rise to placebo effects which is why a placebo control is used rather than doing nothing. The placebo control effectively balances out placebo effects.

Now we know why the likes of Chiropractors and Homeopaths claim that randomised placebo-controlled studies are not the best way to evaluate their treatments - because when they are evaluated that way their treatments are no better than placebo: i.e. they don't work.

I have a masterpiece 8) of an article on placebo and placebo effects on my to do list. I just can't be arsed to write it up. ::)

[seren]

So, if I've understood you right, there is no reason to have the third option of "nothing at all"?

What I don't really understand is that we're also talking about advice, not just medicine. Advice about prevention (exercise, posture, diet, that kind of thing) and management. I can't see how it fits with that. What's placebo advice? Especially with exercise, how can you be sure advice will make absolutely no difference? Maybe this is where the "nothing at all" bit comes in, and the placebo bit is just about medicine. That seems to make more sense.

[Fiona]

Well I do still think doing nothing at all is important in medicine. A double blind trial will distinguish between a placebo effect and a real effect, certainly. What it might not do is show what happens if you really don't do anything. Say a medicine is effective but what it does is get you over a cold a week quicker than you otherwise would be. And a placebo gets you over it 2 days quicker (at least in terms of how you feel and function) The double blind would still show the truth of the medicine but it would not have anything to say on cost and rationing. Does that make sense?

[Admin]

Some trials are done by comparing treatment to placebo and they also compare it to a similar group by doing nothing - these people are usually on waiting lists.

It's not always ethical, however, to use a 'no treatment' group when treatment is actually available.

It's not even always ethical to test a new treatment against a placebo! If there's already a proven treatment for a condition the new treatment will be tested against the established one to see if it is better.

[Cuddles]

Well I do still think doing nothing at all is important in medicine. A double blind trial will distinguish between a placebo effect and a real effect, certainly. What it might not do is show what happens if you really don't do anything. Say a medicine is effective but what it does is get you over a cold a week quicker than you otherwise would be. And a placebo gets you over it 2 days quicker (at least in terms of how you feel and function) The double blind would still show the truth of the medicine but it would not have anything to say on cost and rationing. Does that make sense?

The thing is, placebo doesn't have an actual physiological effect. Giving a placebo does not help you get over anything any faster than doing nothing at all. What it does alter is the reporting. For example, if you give two people with a broken arms placebo treatment and nothing at all, the placebo person will likely report it feeling better, even though by any possible objective measure they are exactly the same.

This is why claims that alternative medicine makes people feel better even if they don't actualy dure them are bad. It might be fine for minor, self-limiting illnesses like colds, where it doesn't matter that you are still ill even if you feel better. For more serious problems where your body can't just heal itself, the placebo effect simply hides the problem and can lead to the problem getting much worse. It is a similar to why you shouldn't take painkillers for a sprained ankle. If you hide the pain, you are likely to damage it even more by walking on it.

[Fiona]

I see. Thanks all :)

[Matt]

It's not entirely accurate to say that placebos do not give a physiological effect. There is an endorphin response that has pain killing effects.
http://www.worldwidehealth.com/article.php?id=610