h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

On Placebos:

A placebo is a dummy drug, usually a sugar pill, which is used in a clinical trial. You will often give half the people a active drug and half a dummy drug, and not let on which is which. The difference between the real drug and the placebo is the effect of the drug. The really *weird* thing is that if you gave half the people a sugar pill and tell them that they are being treated, and half the people no treatment at all, the quite often the people who believe that they are being treated will get better faster than those that don't.

This is called the placebo effect. It's commonly believed that it's the mind healing the disease because it thinks that it's been cured.

The placebo effect doesn't work for everything: it doesn't work for broken legs for instance. In fact a lot of what is thought to be the placebo effect can be explained by a number of factors known as 'biases'.

Observer bias - the person assessing the patient thinks that they are better when they are not better, or the patient thinks that they are better. This was what was happening with Ben and the migraines.

Or you can forget how bad things were - ones brain likes to arrange things in a neat order into a story. And if you have a treatment you want things to get better afterwards so you need to This is more effective if you pay for the treatment. You don't want to think of yourself as a fool who spent a lot of money on an ineffective treatment so you convince your self that you are better, so you don't think that you wasted the money. (This is called cognitive dissonence, and if you don't think you do it read 'Mistakes were made, but not by me'.

Reporting bias. This happens a lot, particularly if you have a nice bedside manner, the patient doesn't want to make you feel bad, or be rude so will say that things are better. 'Well he was ever so nice, so I didn't like to tell him that it was rubbish'.

Follow up Bias: If a treatment is ineffective the patient is unlikely to come back for more, so they are less likely to be followed up.

A placebo is a simulated or otherwise medically ineffectual treatment. Examples would include a sugar pill, an injection of saline solution, or cutting someone open and fitting them with a heart pacemaker, but not switching it on.

The entire point of placebos is to act as a control against a drug or procedure under investigation.

For instance, it's perfectly reasonable to ask: does aspirin cure headaches? How would you find out?

Give 100 people aspirin and see if their headaches clear up? Problem there is, unless you're very unlucky, all 100 people are going to report that their headache cleared up sooner or later.

OK... Give 50 people aspirin and 50 people nothing at all. Except now the people getting the pills know they're getting a treatment, and because they *expect* to get better quicker, weirdly, they will. How to prevent that?

OK... Give 50 people aspirins, and 50 people tictacs. Except the people getting the tictacs will know they're not getting aspirin, so although they're getting A treatment, they'll know it's not THE treatment. Not good enough.

OK... Make the test "blind". Give 50 people aspirins, and 50 people a sugar pill that looks and tastes EXACTLY like the aspirin tablet, and don't tell them which is which. Except... the doctors giving the pills know which is which, and will EITHER subconsciouly communicate this to the patients OR, deliberately or otherwise, massage the results because of whether they expect the aspirin to work or not.

OK... DON'T tell the doctors which is which. Make the test "double blind". Just give the pills a code number and don't tell the doc what it means, so when he reports the results he won't know whether the pill he gave out was the real or the pretend medicine. Surely that's all? No. Because the 50 people you gave the aspirin to were all men, and the fifty people who got the placebo were all women.

So... finally... RANDOMISE who gets what, properly. Double blind the test. And have a control group who get a placebo.

Here's the thing: IF your drug works statistically significantly better than placebo - it works! IF your drug doesn't work statistically significantly better than placebo BUT causes no side effects - it's safe! It does nothing, but it's safe.

Of course, it's always possible that it has other effects - sildenafil citrate was being tested for its anti-depressant qualities, and patients on the real stuff didn't report any improvement in their mental state, but wanted to carry on using it after the trial had ended. When they investigated why, its other effects were discovered, and it's now sold as Viagra. Hurrah!

Here's the other thing, though - it's very important that the placebo and the real treatment look, taste and feel identical. Because in your trial you'll find people who get the placebo get better, quicker than if they had nothing. There's an effect going on.

Trials - proper trials - have shown that some placebos work better than others. Red pills WITH NOTHING IN THEM work better than green pills WITH NOTHING IN THEM. Four pills containing NO active ingredients works better than two pills containing NO active ingredients. An injection with no drugs in it works better than a pill with no drugs in it. People have got better from heart problems after being fitted with a pacemaker EVEN IF THE THING ISN'T SWITCHED ON, because they expected to. These are properly mind-boggling and counter-intuitive results.

Simulpost! I'd just like to point out that Z is an actual doctor, while I am a chemical engineer and NOT a doctor. So make of that what you will.

And they actually don't contraindicate each other.

Turmeric:

Alzheimers I don't know about. But isn't that one of the herbs (see above) that has caused controversy because of Big Pharma trying to patent it? Another one is neem. (I use neem toothpaste).

This doesn't mean that the herbs do anything - although their popularity may be anecdotal evidence that they do. It could just mean that the companies think it's cost effective to patent them on the offchance something shows up, and/or that they are known to have all sorts of complex chemicals in them which look promising. If the latter you also get a natural production facility into the bargain - you only have to extract and purify rather than synthesise. Cannabis, for example, has *tons* of complex goodies in it which may or may not be useful. Other than for the obvious porpoise.

>>
Observer bias - the person assessing the patient thinks that they are better when they are not better, or the patient thinks that they are better. This was what was happening with Ben and the migraines.

Husband bias. They might have actually got better.

Anyone interested should subscribe to Mark Crislips Quackcast podcast as well as reading Ben Goldacre's excellent book.

Before we went out she told me her migraines were cured.

Once I started dating her I noticed she was still getting regular headaches and vomiting every few weeks, that was causing me to have to cancel whatever we had planned. I deduced that migraines weren't cured. I have the unused theatre tickets.

>>Of course, it's always possible that it has other effects - sildenafil citrate was being tested for its anti-depressant qualities, and patients on the real stuff didn't report any improvement in their mental state, but wanted to carry on using it after the trial had ended. When they investigated why, its other effects were discovered, and it's now sold as Viagra. Hurrah!


I'm surprised nobody reported an improvement in their mental state. Maybe they just weren't putting the side-effects to good use.

Neem?

Whassa?

I thought that's what Monty Python's knights said...

B

"Once I started dating her I noticed she was still getting regular headaches and vomiting every few weeks"

Hmmm....really? Did these coincide with any medical lectures or cheap Italian restaurants you were promising to take her to?

B

She never went to any lectures with me.

However we did go for curries in Selly Oak..

Ah, of course, you were in Birmingham at the time, weren't you?

In that case, I think there's a very good case to be made for the fact that the intermittent headaches and vomiting were caused by the state of your kitchen.

B

Yes, there was that.

But they also happened in Ben's very nice flat in a very nice area, with a clean kitchen.

I know exactly what placebo is. I want to know the mechanism of how it works.

Simple example. Years ago I had a wart on my leg for years. I asked a very close woman friend to help me make it go away. She did some woo woo as would be called here..within weeks it disappeared.

Mu Beta, I was about to say that!

What happened with the homoeopathy and the migraines was

1) A nighmare 5 day series of migraines
2) A noticable reduction in frequency followed by
3) Reversion to the norm

This took slightly over a year. The homoeopath, who was also one of my dearest friends, explained (1) as the remedy clearing out the system and (2) as the remedy working.

Who knows? My life was pretty unstable at the time, with bouts of intense work overseas followed by three or four months off, interspersed with a private life like a restoration drama.

My best ever me but I'm gullible story is when an ineffective treatment triggered the placebo effect and cured an imaginary illness.

The difference between me and other people is that I **know** the placebo effect works a treat on me. I've seen it happen again and again over the years.

B

Multiple

>> I want to know the mechanism of how it works.

Ah... well ... whoever manages that should be up for a Nobel.

Effers : We don't know exactly how the placebo works, but I described some of the potential mechanisms why it may seem to work in the post above. Specifically regression to the mean, observer bias and reporting bias.

Well that's why I'm saying people shouldn't just bandy the word around as if everyone knows what it is.

People don't understand placebo.

But its so easy to just say 'placebo'.

Dr. Zen just gave me a description of everything I already knew.

We...ll...

There's also mind-over-mind. Like the time that I had pains from my wrist to my shoulders that resisted pain-killers but reduced in 30 minutes with a magnetic wrist-band and vanished completely in a period of a couple fo weeks. Very dramatic and not co-incidence; more like auto-hypnosis.

B