Bad Science

Submitted by Carlyn Zwarenstein on March 27, 2014 - 06:52

If you are not familiar with Ben Goldacre, you really should do something about that.  A doctor (working now as an epidemiologist) who writes accessibly about the use and misuse of statistics and science (on his website, via Twitter, in a column in the Guardian newspaper, and in the books Bad Science and Bad Pharma), Goldacre has done a lot to help readers overwhelmed by junk statistics and lousy science and medical journalism, quack claims and alarmist Facebook posts.

In his Guardian column a couple of weeks ago, he turned his attention to his own work. Or rather, to lazy media reporting of a paper he co-authored, reporting which suggested that statins have no side effects, a different conclusion from the one drawn in the paper (the researchers found that people report typical statin side effects even when taking placebos). And then, he moves on and actually describes, with clear and detailed explanation, a central flaw in the paper itself (in lack of full source data relating to side effects). It is definitely worth a read, as it gets at a number of his pet issues and some central questions for both academic publishing and popular understanding of medical research.

Also interesting is a question Goldacre raises through this column about the extent to which side effects data is hidden in the translation from unwieldy but complete clinical study reports to brief academic journal publications. 

Commenting on this Blog entry is closed.