Sunday, August 16, 2009

Good Research

One problem that alternative medicine has that they say they produce good quality research. They want their pets to be taken seriously. There are some criteria for papers to be taken seriously. One of the best descriptions is in "Science Based Medicine". But I will try to summerize here. First is the mechanism for action plausible? A good example of how something is not plausible is homeopathy. It is based on like cures like and the more dilute it is the better. Take something that causes the symptoms of a disease and then dilute it hundreds or thousands of times and that make the cure. Several problems with plausibility is when one dilutes something to extent they say there is a point where the only thing left is water. The number you learned in high school chemistry class (that is assuming you took high school chemistry) is Avagado's number. It is 10 dilutions x 23 times. That is the number of particles in a mole of stuff or if you dilute something that is when the thing you diluted becomes so dilute that there is only 1 particle, atom or molecule in the entire volume of dilute whatever. Homeopathy prefers to go higher and that means nothing is in the cures. They then invoke magic, the water has memory. How does water have memory? So plausiblity is important. Next is the experiment reproducable in someone elses lab, not just yours? Case in point is the cold fusion thing few years back. Big media blitz. When the experiment was run again countless times in other labs, no results, nothing, noda. Next credibility of the researcher. Does the researcher have a track record in good journals that have good review teams that look for the errors, bad methodology, etc. Does the design of the research published give a clear understanding of potential confounders. Confounders such as placebo effect have an influence on the results. Was the study randomized, controlled, double blind and placebo controlled. Was there enough subjects in each group to minimize potential statistical false positives (usually more than 50). Was the conclusions overstated for the statistical power of the study and the methodology? If all this can be answered in the affirmative then it is a good study. If not, it is suspect. Homeopathy, accupuncture, chiropractic, among others fail on all counts.
Skeptical DoDo

No comments: