From a peer w great insight.
I just skimmed the headings, but I can guess where this is going.
People that volunteer for studies are more motivated and have better compliance than the average Joe on the street.
I'll go read it now, but I figured I'd go out on a limb.
Okay, I'm wrong--they look like they're randomized. But totally agree with what I've read so far. Exhausted and will go on my rant about the need for more evidence-based research with objective measureable meaningful outcomes!
Thanks for the article!
There are a lot of pitfalls in interpreting medical research literature. A few--1) many studies are supported by the drug or equipment companies and often the researchers are compensated. In some cases the researchers contracts forbid them from publishing negative results. 2) The use of relative vs absolute improvement. Example--drug x improves the survival from cancer from 3% to 6%--is that a 3% improvement or a 100% improvement. Usually the latter is reported. (And drug x costs $50,000 for a course of treatment and makes you sick as a dog.) 3) Study too small--a famous study showed that the survival rate from breast cancer was better if the breast biopsy was done in the middle of the woman's menstrual cycle. This put everyone in a tizzy until more studies were done, and adding the them all together the timing of the biopsy didn't matte. 4)--this one is the most interesting--publication bias. A treatment is the standard of care. Because everyone knows it's the best treatment no one will publish a study proving that--not interesting. Then someone does a study proving the old treatment is bad, and the article is published. (medical journals like controversy as much as the popular press). Only articles contradicting the old treatment get published. After a while the new wisdom becomes the old wisdom, and now studies showing that the original treatment was best start to get published again, while studies showing it to be bad can no longer get published. So the treatment pendulum swings back and forth (over a period of 5 to 10 years usually) from one extreme to the other.