Dr. Google, I Presume
Dr. Google, I Presume
Abstract & Commentary
By Allan J. Wilke, MD, Dr. Wilke is Residency Program Director, Associate Professor of Family Medicine; University of Alabama at Birmingham School of Medicine—Huntsville Regional Medical Campus, Huntsville. Dr. Wilke reports no financial relationship to this field of study.
This article originally appeared in the January 29, 2007, issue of Internal Medicine Alert. It was edited by Stephen Brunton, MD, and peer reviewed by Gerald Roberts, MD. Dr. Brunton is a Clinical Professor, University of California, Irvine, and Dr. Roberts is Clinical Professor of Medicine, Albert Einstein College of Medicine. Dr. Brunton is a consultant for Sanofi-Aventis, Ortho-McNeil, McNeil, Abbott, Novo Nordisk, Eli Lilly, Endo, EXACT Sciences, and AstraZeneca, and serves on the speaker's bureau for McNeil, Sanofi-Aventis, and Ortho-McNeil. Dr. Roberts reports no financial relationships relevant to this field of study.
Synopsis: Searching Google yielded a correct diagnosis in greater than half of cases.
Source: Tang H, Ng JH. Googling for a diagnosis—use of Google as a diagnostic aid: internet based study. BMJ. 2006;333:1143-1145.
After the appearance of anecdotal reports of physicians and patients coming up with obscure diagnoses by use of the internet search engine, Google, Tang and colleagues decided to study its accuracy. They took the entire New England Journal of Medicine's (NEJM) case records for 2005 (excluding management cases) and chose 3-5 terms that described the essence of each case. They then searched Google with these descriptors and chose its top 3 diagnoses. Comparing Google's "differential diagnosis" with the diagnosis as presented in NEJM, Google got it right in 15 of 26 cases (58%). As an example, in one case, an 80-year-old man presented with fatigue, unsteady gait, confusion, and insomnia, leading to death. Tang et al searched the terms ataxia, confusion, insomnia, and death. Google suggested spongiform encephalopathy; the diagnosis in NEJM was Creutzfeldt-Jakob disease (CJD).
Commentary
When I first began teaching in the mid-1980s, I played around with an expert system called DXplain. In 1984, when it was developed at the Massachusetts General Hospital Laboratory of Computer Science, it was DOS-based, knew 500 diseases, and was available only over dial-up on AMANET. It was so 20th century!
It eventually migrated to the World Wide Web and joined many other proprietary, web-based, decision support tools (DST). Why are there so many DSTs on the Web, on your desktop computer, in your PDA, or integrated into your electronic health record? Because there are just too many darn facts and connections linking those facts for us to remember! We have become, of necessity, information masters, and the practice of medicine, art plus science, has changed forever.
The promise of DSTs, as yet unfulfilled, is to make the science instantly available, allowing more time for the art. Why has Google overshadowed other DSTs? One factor is it's free, and they are proprietary. The other is its breadth; Tang et al report that it has access to 3,000,000,000 articles! However at this time, Google can't make a diagnosis; it can find articles that include the search terms that you give it. The better the search terms, the better the match. It still requires the skills of a physician to sort though the articles to find those that make the most clinical sense.
It's about to get better (or worse), depending on your stance in the humanity-vs-machine debate. An editorialist discusses the development of the "semantic web," a melding of technologies "which aims to create a universal medium for information exchange by putting documents with computer processable meaning on the worldwide web." The future is so bright, we'll all need sunglasses! When I interviewed residency candidates during this recruitment season, I asked them, "How do you want to be remembered, as a great diagnostician or a great healer?" With Google they could have it all.
Reference
1. Tang H, Ng JH. Googling for a diagnosis—use of Google as a diagnostic aid: internet based study. BMJ. 2006;333:1143-1145.
Searching Google yielded a correct diagnosis in greater than half of cases.Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.