Bad Statistics Lead to False Hope

Newspapers around the world are all agog with the story of a British Man, 25, ‘cured of HIV’; that headline from the normally reserved BBC.  Scot is first in world to beat HIV says, (can you guess?), the Glasgow Sunday Mail.  The more cosmopolitan, but doubly wrong, Medical News Today says, Man is Cured of AIDS.  Other newspapers are reporting that doctors are "stunned," "mystified" and wondering whether this man holds the key to curing AIDS.

The story is pathetically simple once one gets past the headlines.  A man tested positive for HIV, he took a lot of vitamins and just over a year later tested negative (several times).  Now what are you going to believe that he cured himself of HIV or that the first test was wrong?  HIV tests have high accuracy but when millions of people take these tests it’s an easy bet that there will be significant numbers of false positives.

It is even possible that in low-risk populations there will be more incorrect diagnoses than correct ones!  Doctors may be stunned but to a statistician results like this are banal.  Unfortunately, in about a dozen articles that I took a look at, many doctors were quoted (sadly, even the skeptical doctors were skeptical for the wrong reasons – they think the guy must still have HIV!) but not a single statistician.  For the correct statistics see here or my earlier post, Why Most Published Research Findings are False, which analyzes a different application of the same idea.