The BMJ recently published a paper written by a team of Australian doctors testing the use of Google as a diagnostic tool.

The article concluded that “Our study suggests that in difficult diagnostic cases, it is often useful to “google for a diagnosis.”Web based search engines such as Google are becoming the latest tools in clinical medicine, and doctors in training need to become proficient in their use.”

The study found that the correct diagnosis was found in 58% of cases.

At first glance only finding the correct diagnosis on only half of cases seems pretty poor (especially compared to what you would expect from a real life doctor).

However, getting any diagnosis correct is a considerable achievement for a tool not built specifically for this purpose and getting over half right is certainly significant. It may even be useful practically in difficult cases where experienced doctors may be looking for suggestions they hadn’t thought of.

It will be interesting to see how this figure progresses over time. How long will it take (if ever) for a search engine like Google to approach the accuracy rate of healthcare professionals.

Google currently doesn’t have access to the range of databases and research available to other diagnositic engines such as Isabel. Perhaps when more databases and research papers are indexed it’s rate of correct diagnosis will increase.

There has been a wide ranging discussion on this paper on the blog-o-sphere:

The IT for Healthcare Blog highlights the authors concerns about patient’s self-diagnosing with Google and points to this counter argument in Modern Healthcare.

The Clinical Cases weblog wonders whether the “wisdom of the crowd” basis of search engines is good enough to be relied upon for diagnosis making.

Tom Roper and Phil Bradley wonder why more medical librarians aren’t publicly commenting on the issue.

The Krafty Librarian has answered this call pointing out that the 58% success rate is only marginally better than flipping a coin.