Quote:
Originally Posted by David Sklansky
They have a long position in OREX. There is always an article defending on thing or another. And this is a ridiculous spin piece, with no discussion of the huge negatives (and why this is likely to be false given prior results).
But let's get away from the BS and consider this: we're talking about a difference of 40 people out of 9000. In a study with a very high dropout rate, where patients were kicked off due to lack of compliance with the diet guidelines. Where information about early results was widely disseminated to a large number of parties, indicating extreme leakiness of information. For a company who is live-or-die based on this study. There are dozens of reasons why this result could be found despite the drug being harmful. A few are:
- Weight loss was greater on the drug than an placebo (this is established as it is an approved weight loss drug like others), and those that were kicked off from the drug were those that didn't lose weight. This effect would be stronger in the drug group as they lose more weight if they follow the guidelines, lowering cardiac risk.
- Fraud. The early release and wide dissemination and patent application makes fraud quite likely. There are billions of dollars in an all-or-nothing scenario riding on this
- Lack of blinding. Non double blinded studies are worthless. Given how leaky this was, we can assume that the study is tainted. Perhaps an early statistical artifact created strong excitement. The fact is that in most studies, both doctors and patients often know who's taking what, despite the claim of double blinding.
- A selection mechanism within the drug (given that it's psychoactive), where the most sick stop taking it and drop out because of how it makes them feel (as one of the drugs increases blood pressure, for example - where the effect is strongest, they have dropped out)
- A placebo response given that the drug has identifiable psychoactive effects
- A reduction is risky behavior given that the drug inhibits pleasure seeking activities (smoking, etc).
This is why you let studies run their course. You need to check if follow up is reliable and if these biases are occurring. These are biases more likely to occur at the start of the study than at the end, and the run of the statistics over the course of trial will weed out a lot of the nonsense that can happen.
This is just dodgy on so many levels and why the FDA is all over this. Don't believe the analysis of shills with a long position to protect.
Last edited by ToothSoother; 03-06-2015 at 05:21 PM.