Bad Science: Quacks, Hacks, and Big Pharma Flacks

Recently read Bad Science: Quacks, Hacks, and Big Pharma Flacks by Ben Goldacre (Faber & Faber, 2010).  Dr. Goldacre is an epidemiologist in London who has written the "Bad Science" column for The Guardian since 2003 and has a blog of the same name.

It's a humorous and entertaining book, written much in the same style as his blog, and concentrating, naturally enough given his background in medicine, on bad science in medical and health issues.  He discusses issues like "alternative medicine", homeopathy, nutritional pseudoscience, and immunizations (especially the MMR/autism nonsense).  He also has an extensive discussion of the placebo effect (which I found fascinating) and a great discussion of how probabilities are reported in such a way as to exaggerate risk.

Goldacre also skewers the media (deservedly) for its terrible reporting on health issues.  Let me give you a couple of examples from the book.  Suppose you read a health story in the newspaper saying that men in the fifties with high cholesterol are 50% more likely to have a heart attack.  Sounds scary, doesn't it?  This is how the news media typically reports health news.  But it's misleading bullshit!  Why?

The risk of heart attack of men in their fifties with normal cholesterol is 4 out of 100 men.  For men with high cholesterol, it's 6 out of 100 men.  That's a 50% increase.  It's also a 2% extra risk.  Would you see a newspaper story, however, saying that the extra risk of a heart attack for men in their fifties with high cholesterol is 2%?  Of course not, it's not dramatic and scary enough (and most men would say, "Only 2%, I can live with those odds!").

Another example was a news story in Britain about how cocaine use had doubled in schools in some area.  The actual data was a survey that asked school kids if they used cocaine in the past year and the number went from 1% to 2%.  In actual fact, however, the raw data was really 1.4% to 1.9% (0.5% increase) since the numbers were rounded for the report.  Is that increase even of statistical significance?

The media has also aided and abetted the supposed MMR-autism link, giving voice to Hollywood celebrities and their half-assed ideas yet ignoring the scientific data which has conclusively shown that there is absolutely no link between the two.  As an epidemiologist, Goldacre is rightfully concerned about the loss of herd immunity for childhood diseases that have the potential to kill because of unfounded fears over vaccinations.

I have a rant about media and science which I'll be writing soon so I won't dwell much more on Goldacre's points which I mostly agree with.  I'll save another quote from his book for tomorrow's post.

If you're interested in science and health, this book is a good read.  I found it in the local library.