Только кусочек процитирую.
Let’s look at the following graph of the number of trials completed versus those that were published. In 2008, the company Sanofi completed 92 studies but only a piddly 14 were published. Who gets to decide which gets published and which does not? Right. Sanofi. Which ones do you think will be published? The ones that favor its drugs, or the ones that prove their drugs do not work? Right. Keep in mind that this is the only rational course of action for Sanofi, or any other company to pursue. It’s idiotic to publish data that harms yourself. It’s financial suicide. So this sort of rational behavior will happen now, and it will not stop in the future. But knowing this, why do we still believe the evidence based medicine, when the evidence base is completely biased? An outside observer, only looking at all published data, will conclude that the drugs are far, far more effective than they are in reality. Yet, if you point this out in academic circles, people label you a quack, who does not ‘believe the evidence’.....
.....Imagine you have a coin flipping contest. Suppose a player call ‘Big Pharma” chooses heads, and also pays the coin flipper. Every time the coin flipper pulls up tails, the results don’t count. Every time it comes up heads, it counts. This happens 28% of the time. Now, instead of a 50/50 split of heads and tails, it’s more like a 66/34 split of heads/tails. So the ‘evidence based medicine’ lover claims that heads is far more likely to come up than tails, and castigates people who don’t believe the results as ‘anti-science’.
UPD Русским языком про то же самое по мотивам (ссылка из обзора у dok_zlo)