Home / Featured / Watson: Extreme Evidence Based Medicine

Watson: Extreme Evidence Based Medicine

By Lynn Lucas-Fehm, MD, JD

Most of us recall the literary character Dr. Watson who served as the steadfast confidant, supporter, physician and assistant to the brilliant detective Sherlock Holmes.  Now there is a new Watson in our midst, an artificial intelligence computer developed by IBM and named after IBM’s first president Thomas J. Watson.

After handily defeating the formidable human Jeopardy champions, Brad Rutter and Ken Jennings, Watson’s developers have expanded the computer’s medical databases to create what may become the ultimate digital collection of medical information. However, what truly sets Watson apart is that it can analyze facts gathered in natural language and generate differential diagnoses.  It even assigns a ranking to each diagnosis based on its understanding of medical knowledge in textbooks, journals, and case reports.

At a recent demonstration for The Associated Press, Watson was asked to consult on a fictional patient with an ophthalmic disorder. As additional symptoms, medical history and personal information was presented — blurred vision, family history of arthritis, Connecticut residence — Watson’s suggested diagnoses evolved from uveitis to Behcet’s disease to Lyme disease. It gave its final diagnosis a 73 percent confidence rating.  Dr. Herbert Chase, a Columbia University medical school professor and Watson consultant commented that “You do get eye problems in Lyme disease but it’s not common… you can’t fool Watson.”


IBM’s Dan Pelino, general manager for global health care said possible future uses for Watson include allowing a doctor to connect to Watson’s database by speaking into a hand-held device using speech-recognition technology with cloud computing serving as the database for the most advanced research.

There is little doubt that Watson will help physicians with a major problem in modern health care: information overload.  The potential to have immediate access to every resource of evidence based medicine is exciting.  However, Watson developers and consultants will go a step further. Dr. Chase stated that anecdotal information — such as personal blogs from medical websites — may also be included.

Chase remarked, “What people say about their treatment … it’s not to be ignored just because it’s anecdotal. We certainly listen when our patients talk to us, and that’s anecdotal.”

This is where the issue of Watson’s involvement in medical practice becomes a slippery slope.  Since Watson has the ability to interpret natural language will it be able to listen to patient’s complaints, analyze them and give a differential diagnosis?  Watson could become the ultimate physician’s assistant.  In fact, with the increasing reliance on “apps” one might wonder if Watson could truly become “Dr. Watson”.

I have always embraced new technology, reveling in the acquisition of every new smart phone technology, iPAD app and software innovation.  However, reliance on such technology to the exclusion of the human factor in the practice of medicine may lead to increased productivity but suboptimal patient care.  A balance must be achieved in the upcoming decade where technological tools are utilized to provide the physician with the evidence based medicine to assist in accurate diagnosis while allowing the irreplaceable intuition, compassion and humanity only we as physicians can provide.


Lynn Lucas- Fehm, MD, JD, is a radiologist at Abington Memorial Hospital and the 150th President of the Philadelphia County Medical Society.





  1. Martin S. Kohn, MD

    Your concern about the human element is entirely appropriate. That is one of the reasons that we emphasize Watson as a “clinical decision support” tool, not a “clinical decision making” tool. It is also why we specifically avoid the appellation “Dr. Watson.” With it’s ability in NLP processing IBM Watson is able to understand the nature of the decision to be made in a clinical situation. It then reviews large volumes of material, which may include the patient’s EMR if it exists, and returns to the decision maker prioritized information, in the form of suggestions, that the decision maker can use to help make a better decision. Thus, for example, a physician will still consider all the human factors in making a decision about a patient, but will be supported by access to a range of information that the physician would not otherwise have the time to acquire. Watson’s forte in NLP fills in the gaps in human decision behavior, it does not replace the decision maker and does not preclude the human factor.

    You are right in saying IBM Watson will never replace a physician and we are in complete agreement! — Dr. Martin Kohn, Chief Medical Scientist, IBM Research

  2. As a patient with neuroendocrine cancer, one of the “zebra” diseases, I think Watson is a brilliant tool. It took doctors nine years to figure out that I had a tumor on my pancreas, and my husband just learned that he has a prolactinoma that has probably been growing on his pituitary gland for 20 or 30 years. Since doctors tend to look for the common cause and often overlook rare disorders, I love the idea of having a computer at the ready to do some checks and balances. We certainly cannot replace doctors with a computer, but if Watson can help us come to a diagnosis quicker, we could save lives and medical dollars.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.