Days after the IBM computer known as "Watson" trounced two Jeopardy grand champions with impressive ease, the announcement that the technology would be applied to healthcare and could function in the role of a "physician's assistant" sent a minor ripple through the PA profession. Could Watson, which has the ability to process 60 million pages of information per second, replace a clinician?

Absolutely not, according to Martin Kohn, MD, IBM's chief medical scientist for care delivery systems. In fact, Watson will likely allow PAs and NPs, as well as physicians, to expand their clinical roles.

"Watson is not a decision maker," Kohn said. "It works based on the information given to it, and it understands the nature of the question being addressed. It scans through the evidence available - which could include the patient's electronic medical records or genomic information. It brings back detailed, prioritized suggestions based on evidence." Just as it is now, the decision on a clinical course of action will be made by the clinician and the patient.

It could be some time before Watson is integrated into day-to-day patient care. The technology is currently being "trained" to function in a clinical environment. It is being given the same information and standardized tests that are given to clinicians, and IBM is negotiating with the publishers of clinical journals and other information sources to build Watson's knowledge.

Last September, WellPoint announced plans for the first commercial application of Watson to assist clinicians in developing evidence-based, personalized cancer treatments. WellPoint is working with the Cedars-Sinai Samuel Oschin Comprehensive Cancer Institute to build Watson's knowledge of oncology treatment and research.

Kohn pointed out that Watson is not a search engine in that it does not find its answers based on key words. Over time, it will "learn" that some sources are more reliable than others, and it will recognize evidence-based information. For example, Kohn said, Watson may find one article that's relevant to the question asked. The technology is aware that no journal is correct 100 percent of the time, but it can tell based on past experience when an article is from a journal that tends to be reliable. If it finds a single article from a journal it judges to be reliable, it would likely offer a diagnosis based on that one article at a lower rate of confidence, Kohn said.

While Watson will put its ability to process enormous volumes of information nearly instantaneously into the hands of clinicians, it won't be able to duplicate observation and other uniquely human elements of patient care. Kohn used the example of a patient who claims to be suffering from pain equivalent to a 10 on the pain scale - a level of pain normally assumed to be immobilizing. If the patient is relaxed and moving normally, the clinician will know that the patient's idea of severe pain is different from the accepted clinical definition.

"Nobody will be replaced [by Watson]," Kohn said. "It's not a good idea, and it can't be done."

The concept behind Watson is not new, and other programs have existed for several years, said healthcare futurist Joe Flower, CEO of The Change Project and author of the book Health Care Beyond Reform: Doing It Right for Half the Cost. Clinicians have historically been resistant to anything that interrupts their work flow, but that could be changing, Flower said.  "Doctors are also very skeptical about being second-guessed by a computer program. But the information [provided by Watson] is not provided by a programmer. It's the latest information that can be broken down into decision-making algorithms. "It's very interesting to see a power like IBM deciding to take on this part of health care. It's probably a sign that we're going to see a lot more of it."