Clinical trials/studies explained
I wrote this article some time ago for another site, and it never got used. I thought it’d be perfect for the supplement sub-forum.
Why bother with trials?
For the most part, people rely on what’s known as anecdotal evidence. An example of this might be: “My friend who had a cold ate a gram of sawdust and recovered within a day”. For a lot of people, this might be reason enough to try eating sawdust. In reality, there are a number of problems:
- Was the sawdust really responsible for his cold being cured? How can we know that he didn’t recover spontaneously on its own, and the sawdust being taken at the same time was just a coincidence?
- If the sawdust did have an effect, is it the case that it would have the same effect on every person who took it, or just select individuals?
- Could the placebo effect be responsible?
- How is it possible to verify the contents of the sawdust ingested?
- How can you be sure that the friend really ingested only one gram of sawdust?
- How can we know if the sawdust caused any later side effects?
- For that matter, how can we even be sure that this friend exists?
Not all trials are created equal
In our sawdust example, we could solve a lot of the problems by taking a group of 20-30 people with a cold and giving half of them a gram of sawdust, and half of them nothing. We’d then examine their recovery rates and see how much faster the sawdust group recovered compared to the group that got nothing (this second group is known as the control group. If the sawdust group recovers significantly faster, we can assume that the sawdust was responsible. Except, this isn’t strictly true.
The Placebo Effect
The Placebo Effect is a commonly misunderstood phenomenon. If I were to give a person with a headache a placebo pill, and their headache were to be cured, one might assume that the person had simply “imagined” that their headache had gone away, or pretended to not feel it anymore. The strange thing is, placebo treatments can actually cure many conditions biologically. Nobody is entirely sure exactly how the placebo effect works, and how it is capable of treating headaches, depression, anxiety, and a whole host of other physical and mental problems. Bottom line is: it does, although usually not nearly as strongly or effective as genuine treatments. If you’re interested in finding out more about the placebo effect, there’ll be another article coming up on that exact subject.
So, how can we separate out a placebo treatment from a genuine treatment? How can we tell that our sawdust is actually doing anything in the body, and that the body is not just taking cues to treat itself? By using a placebo-controlled trial. This time, instead of giving half the group nothing, you give them a powder that looks and tastes like sawdust, but is just plain old sand. At the end of the trial you can compare the improvement in the sawdust group and the placebo group. If the numbers are not significantly difference (i.e cannot be put down to random chance), we can conclude that the sawdust is effective in treating colds. Except, we can’t.
In the example we just gave, you’ll notice I said that the sand was made to taste just like the sawdust. This is so that the study participants do not know who is getting the real medicine, and who is getting the placebo treatment, this is called blinding. Oddly enough, knowing that a treatment is a placebo changes how effectively it works. Other factors that effect the efficiency of a placebo include:
- Pill color Placebo pills with a “hot” color (such as red, orange, yellow) work better as stimulants while placebo pills with a “cool” color (such as blue, green or purple) work better as depressants. This is true in many countries except Italy, apparently.
- Pill size Large pills seem to have a larger placebo effect
- Price Higher priced placebos work better.
- Injections are more effective than pills
In a study by Gracey et al in 1985, a trial was set up to investigate the effect of three substances on pain reduction. The three possible candidates were fentanyl (an anaesthetic), naloxone (a pain-relief blocker, would increase pain), and a placebo. Before the experiment was conducted, the doctors were told that the supply of fentanyl had been held up. They were now aware that the patients they were giving treatment to had no possibility of pain reduction. After half of the patients were treated, the doctors were told that the fentanyl supply had come through, and they now knew that the patients had the possibility of receiving a pain-relieving substance (though still had no idea which patient was getting what, remember that this study is blind). The results were quite staggering:
http://img440.imageshack.us/img440/6851/placebo.png (or see attached)
In the PNF group, the doctors knew the patients had the possibility of getting pain relief. In the PN group, they knew there was none. Credit to mindhacks.com for the scanned image and Daniel Moerman's book “Meaning, medicine and the 'placebo effect'” for the original diagram.
The doctor’s expection heavily influenced the placebo effect, even if the doctor never told the patient they were receiving an inactive treatment. What this demonstrates is the importance of a second layer of blinding.
So, in our sawdust experiment we have to make sure our person handing out the sawdust and sand doesn’t know who is getting which. Our study is now double-blind. If our sawdust group now recover a lot faster than the sand group, we have some convincing evidence that a gram of sawdust is an effective treatment for colds. Sadly, as science is always in the pursuit of higher standards, there’s more to it than that. In the next article I’ll cover triple-blinding, randomized and cross-over trials, and other stringent measures. For now, enjoy the fact that you’re now better informed on science-based medicine!
Start (Initially August 2008, Properly November/December 2008): 6.9x4.9
Currently (September 2010): 8.5x5.75 (6.25 inch head girth)