SoylentNews is people
SoylentNews
from the let-me-sell-you-some-snake-oil dept.
AlterNet reports
[...]the prestigious British Medical Journal has joined [The New Yorker, the US Senate, and the Twittersphere in condemning Mehmet Oz, MD for the pseudoscience on his TV show].
In an article published this week, a group of health experts analyzed a random sampling of episodes of "The Dr. Oz Show" (along with another syndicated show, "The Doctors"). The upshot: the evidence supports less than half of what he says. Which, in practical terms, means you should have reasonable doubt about all of it.
The researchers sat through 40 episodes of the "The Dr. Oz Show"; from those, they identified 479 separate recommendations he or his guests made to his TV audience. After winnowing the selection down to more forceful recommendations, they randomly selected 80 and weighed them against the existing medical literature, evaluating each claim for "consistency and believability."
Only 46 percent of the advice, they found, had evidence supporting it, and just 33 percent of the time were those claims supported by "believable or somewhat believable evidence." For just more than 1 in 3 recommendations, they weren't able to find any supporting information at all (despite, they note, "being quite liberal in the type and amount of evidence we required").
The sad part is how many people get their "information" from television.
(Score: 0) by Anonymous Coward on Tuesday December 23 2014, @06:48PM
To do a full-on double-blind placebo study on a properly sized N from a group of random ethnicity, gender, age is extremely difficult, time and money consuming. And it needs to be repeated several different times. And then, and only then does it have a modicum of likely of being correct. The statistics in this area is abused so much, and there's so much bias as to even make this kind of studying have a very high probability of being "wrong".
So what's left. If you have a supplement, that is natural and not likely to harm you like a real drug, if it doesn't help the patient, the patient has only lost some $. But if it does help, whether due to a placebo effect or not, then you have given the patient (or subject) a chance of gaining a benefit. If you are using potent drugs with potent side-effects a gun-shot approach is dangerous.
This is the basis of the art--not science.
If you are relying on published findings to validate his recommendations, you are relying mostly on things that are false: http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.0020124 [plosmedicine.org]
My soapbox: Most so-called science isn't science. And it's really expensive. And it's an epistemological nightmare.