The Doctor Weighs In

I must confess, I use Google to complete the NY Times Sunday crossword puzzle (please don't tell Will Shortz).   And I regularly use Google instead of the phone book.  I plan my vacations, do research for my blogs, and try to locate old friends using Google search.  In fact, I use Google for just about everything, but I haven't been using Google to diagnose disease.

Turns out, Google is not too bad as a diagnostician.  Some of you may have heard about a doctor who astonished her colleagues by diagnosisng a rare condition (immunodeficiency, polyendocrinopahty, enteropathy, X linked syndrome or IPEX for short) by putting the patient's symptoms and other features into a Google search box. 

More recently, researchers in Australia decided to put Dr. Google to the test, as reported in the November 10, 2006 online issue of British Medical Journal.   Every issue of the New England Journal has a discussion of a particularly difficult case.  It is set up in a way that doctors reading it try to figure out the diagnosis.  A super-specialist in the relevant field discusses the findings of the case and suggests a diagnosis.  At the end of the article, the real diagnosis is revealed.  It is a sort of "stump the expert" type of exercise.

Hangwi Tang and Jennifer Hwee Kwoon Ng, Australian physicians looked at a year's work of these case studies--there were 26 of them -- but they didn't look at the last page where the final diagnosis was listed.  They then entered findings from the case into Google.  For example the entered "cardiac arrest sleep" from a case that turned out to be an unusual type of heart rhythm abnormality (Google got this one). 

Overall, Google came up with the correct diagnosis somewhere on the first three to five pages of the search results in 58% of the cases.  What the authors of this paper didn't tell us was how often the superspecialist who discussed the case came up with the right diagnosis.  I know from reading these cases over the last 20+ years that the specialists don't always get the diagnosis right.  These are really tough cases.  I want to know whether the specialists were as good or better than Dr. Google.

The authors suggest that, because they were doctors, they may have been able to enter terms that non-medical people would not have entered.  Thus, their searches might yield the correct diagnosis more often than the searches of the general public.  I wonder?....it sounds like that is the next study these clever doctors should do.