IBM supercomputer Watson made headlines a few months ago by beating two “Jeopardy!” champions on national television. Now, Watson has a new gig: doctor.
On Monday, health insurance company Wellpoint announced a partnership with IBM to “develop Watson-based solutions to help patient care through the delivery of up-to-date, evidence-based health care for millions of Americans.” Patients are unlikely to have a Dr. Watson, rather, their human physicians would consult with the computer’s database in treating patients.
The partnership may feel a little odd - a former computerized game-show champion now heading to medical school - but the skill set is actually pretty similar. “You want to rapidly process information and provide the right answers,” Sam Nussbaum, Wellpoint’s chief medical officer, explained in an interview.
The computer can analyze about 200 million pages of data in less than three seconds, which could allow physician to more accurately diagnose and treat complex cases. Physicians could, for example, use Watson to consult medical records and the latest research findings for recommendations on treatment.
Nussbaum met Watson last April at IBM headquarters in Yorktown Heights, N.Y. The session, a few months after the “Jeopardy!” victory, included Watson’s animated avatar, IBM officials and Wellpoint executives (Watson’s hardware, stored elsewhere in the building, was not in attendance.) Nussbaum very quickly saw how Watson could fit into the increasingly data-driven business of delivering health care.
“Historically, if you look at our use of artificial intelligence, we use a lot of claims data to identify gaps in care and have done analysis on that,” Nussbaum says. “This really is an extension of that.”
Watson will make its debut, in small clinical trials, in 2012 -- after it goes through some training. “Watson can go to medical school, residency and fellowship in a matter of months,” says Nussbaum. “If you think about how much is happening in research, about 1 million new articles coming out every year, Watson can be so much more widely read than any physician.”
But some physicians do worry about a downside to computer-generated diagnoses. As Stanford University’s Abraham Verghese wrote in a recent New York Times op-ed, the practice can undermine some of the necessarily human aspects of medicine. Verghese worries that converting a patient into a set of data creates a sort of “iPatient,” which “threatens to become the real focus of our attention, while the real patient in the bed often feels neglected, a mere placeholder for the virtual record.”