Tuesday, September 25, 2012

Must read

Today's must-read: The drugs don't work: a modern medical scandal, by Ben Goldacre in The Guardian, about the systematic corruption of science by the pharmaceutical industry. Reading it, I'm left with the unpleasant feeling that our system for approving drugs is utterly dysfunctional and is routinely manipulated by companies to make money. It also seems that Big Pharma is below even social science in the credibility stakes. Yes, we joke that a 95% confidence interval means one social science result in twenty is crap (happened by chance alone) - but at least social scientists regard this as a flaw. Meanwhile, Big Pharma goes actively looking for that one-in-twenty crap correlation, then publishes it while suppressing the other nineteen results, because they can use it to sell something and make some money. In doing so, they're little different from snake oil merchants.

Fixing this is going to mean some radical changes to the pharmaceutical industry. If we want honest results, then we need to take the job of testing drugs away from the industry, and give it to independent regulators. Until we do, we are going to see more and more problems with drugs that don't work, or are actively harmful, because money will continue to corrupt the scientific process.