Diagnosis by Google

As reported in the press,(CBC; Medical News Today) a recent study published in the British Medical Journal examined the effectiveness and accuracy of Google to diagnose difficult cases. Doctors at a hospital in Brisbane 'Googled' the symptoms of 26 difficult-to-diagnose cases drawn from cases published in the New England Journal of Medicine.

The doctors chose three to five search terms from each of the individual cases without knowing the diagnosis. They then selected the three most prominently displayed search results, and chose the one they felt was the most relevant and compared the diagnosis to those in the journal. The results from Google were accurate in 15 of the 26 cases, or 58% of the time.

The researchers of this study believe that search engines such as Google are useful tools in a world where medical information is expanding so rapidly. Doctors are unable to keep up to date with the changes. As represented in the press, the tone of the research team was one of satisfaction with Google's ability to provide them with correct diagnoses.

Obviously, doctors have the knowledge to properly evaluate the information that they retrieve from the Internet, however a 58% accuracy rate, in my opinion, is not encouraging. Perhaps there is more to this study than has been reported in the press, as there often is, but it seems rather unscientific and reckless to choose the three most prominently displayed results without evaluating the sources from which this information has been obtained. If this is indeed the case, this is the method they followed, hopefully no doctor in the real world would search in such a manner.