Boozers are Choosers of Placebos

October 29, 2012

There has been a lot of talk lately about the power of placebos to produce health benefits, even when the subjects knew they were using a placebo. This makes establishing the positive performance of a drug candidate especially challenging, since a drug will not obtain approval if it does not exceed the performance of the “placebo arm” of a clinical trial by a statistically significant margin.This was noted recently in the trial of the alcohol dependence drug nalmefene.

The drug is touted as the first treatment to counter heavy drinking, and indeed the Phase III trial showed an impressive 66% reduction in total alcohol consumption. But the drug, which presumably eliminates the brain's pleasure response to drinking, doesn't fare so well when you shift from percentages to numerical comparisons with a placebo. In fact, in three different trials, the reductions in alcohol consumption and binge drinking days slid dramatically, but so did the behavior of those on the placebo and the results were statistically indistinguishable.

These observations are in line with a CBS 60 Minutes story (http://www.cbsnews.com/8301-505269_162-57380096/inside-60-minutes-placebo-story/) on placebos versus antidepressants, in which the psychologist behind the study stated that there was virtually no difference in response  among patients in the two groups.

So what can we conclude from this? First of all, our minds are very powerful, perhaps so powerful that they may exert profound influence over the pleasure responses that drive our behavior. Secondly, without rigorous double blind studies that deliver large, repeatable, highly significant differences between the arm of the trial that gets the actual drug and the arm that gets the placebo, there is NO reason to adopt the drug in question. And finally we must accept the fact that there is not a shred of scientific evidence to support the alternative (“herbal”) medicine industry, since rigorous studies have never been done with such remedies.

Comments

Submit a Comment

Please be sure to fill in all information. Comments are moderated. Please no link dropping, domains as names; do not spam and do not advertise.