Talk about TMI. One trillion devices are moving at breakneck speed, generating 2.5 quintillion bytes of data per day. In the world of medicine, information is doubling every five years. In 2010 alone, 700,000 new articles were catalogued by the National Library of Medicine.
That’s the landscape that Martin Kohn, M.D., Chief Medical Scientist, Care Delivery Systems, IBM Research, describes before he explains why IBM believes a supercomputer like Watson can help. How is the typical physician, who spends less than five hours a week reading two or three relevant journals, supposed to keep up or more importantly wade through the meaningless data to find what’s important to patients? Even if the knowledge exists, finding it can be impossible because so much of the information out there is unstructured.
Perhaps we shouldn’t be surprised that 30 percent to 50 percent of health care dollars are spent on things that have no value. It’s simply not humanly possible to keep up.
IBM’s Supercomputing Solution
What IBM has imagined is a tool for physicians that would read thousands of articles and within seconds extract the ideas precisely relevant to the patient decision at hand. “It’s a natural language processing tool that reads and understands English and brings back ideas for you to think about,” says Kohn. Watson wouldn’t just link the physician to the pages or journal articles, Kohn explains. It would say, “I have just read a couple of thousand articles. I found 500 relevant articles. Here are my suggestions that you should consider in making a diagnosis.” Furthermore, it ranks its hypothesis according to how confident it is in its reliability and relevance. And then asks the physician to be more specific.
Watson became famous playing Jeopardy in February 2011. During Jeopardy, Watson read 200 million pages of text in two seconds, and was able to understand the intentionally arcane language and obscure clues of Jeopardy to defeat the two reigning champions.
Using Watson as a Physician-Assistant Tool
Does Watson have the potential to make physicians irrelevant? Absolutely not, says Kohn. “Watson is a tool designed to help physicians make better decisions, not to make the decisions for them.” While no supercomputer, however smart, could take the place of a human hand or mind, harnessing that power to take better care of patients is, at the very least, an intriguing idea.
What do you think of Watson?
"Putting IBM Watson to Work In Healthcare,"
The full presentation Martin Kohn, M.D. recently gave at the University of Utah Health Sciences
Amy Albo is the Publisher for University of Utah Health Sciences.