WellPoint counting on Watson to ask the right health care questions

After being crushed by IBM's supercomputer Watson on the game show "Jeopardy!" past human champion Ken Jennings said he welcomed the new computer overlords. Insurance giant WellPoint soon will find out whether physicians will lay out the same welcome mat when Watson enters their practices.

WellPoint has partnered with IBM to develop the first commercial version of the Watson technology. It plans to conduct two major pilot programs using the Watson technology, one for clinical purposes and the other for care management.

The Indianapolis-based insurer believes Watson's ability to process human language, sift through the equivalent of 1 million books and analyze key data points to produce the most probable diagnosis or effective treatment options for complex cases in three seconds or less, has the potential to improve patient care and reduce costs significantly. IBM, which announced in February that Watson's next move would be to health care, believes the same thing.

"With medical information doubling every five years and health care costs increasing, Watson has tremendous potential for applications that improve the efficiency of care and reduce wait times for diagnosis and treatment by enabling clinicians with access to the best clinical data the moment they need it," said Manoj Saxena, general manager of Watson Solutions at IBM Software Group.

Sam Nussbaum, MD, WellPoint's chief medical officer, said the rules-based decisions being made by the insurer would be no different with Watson. They would just be faster, more efficient and developed from a much wider-reaching data set. He downplayed possible concerns that Watson would take away from a physician's autonomy.

"Doctors make decisions for patients," he said. "We want to work with our physician partners to inform those decisions, if the decision is not necessarily based on the most current scientific knowledge."

Greater use of decision support tools

Physician use of clinical decision support tools in their practices has risen significantly in recent months, mostly due to requirements under the federal meaningful use incentive program that requires decision support to be a function on the electronic medical records systems used to qualify for the bonuses.

A study of health IT adoption in California, published in May by the state's Healthcare Foundation, found 72% of physician practices in California had implemented decision support tools, including 58% of solo practices.

Physicians can use their own clinical decision support tools. But they also are receiving communication from health plans based on the insurers' use of their decision support tools. WellPoint, for example, will alert physicians to identified gaps in care that are based on claims data. But the promise of Watson is that a health plan could deliver even more information to physicians, more quickly.

Dr. Nussbaum said insurers are in the best position to offer the technology, not only because of their ability to make the large investment, but because of the amount of data a health plan can access to develop evidence-based best practices. WellPoint did not disclose the financial details of its arrangement with IBM.

Jeffrey Kellams, MD, president of the Indianapolis Medical Society, thinks an insurer that has the patient base of WellPoint is a great place to start with a technology like Watson. WellPoint has more than 33 million members from which data can be drawn.

"I don't think reasonable physicians will look at it as being an intrusion into practice," Dr. Kellams said. "I think they will look at it as being an assistant."

He said physicians are accustomed to receiving letters from insurers alerting them to possible drug interactions, side effects and alternative therapies if the insurer notices that the doctor recently prescribed a patient a new drug or made a diagnosis.

"It's always presented in a professional manner, and it's up to the physician whether they want to take what's conveyed in the letter to heart or not," he said. "I think to a large extent it will probably be regarded in the same way with Watson ... and it will be much more immediate. There's no doubt about that."

Eliot Siegel, MD, professor of diagnostic radiology and nuclear medicine at the University of Maryland Medical Center, said it would be very easy for physicians to become concerned if WellPoint planned to use Watson to make all decisions on what treatments patients should receive. "But I really don't believe that's the case," he said.

Dr. Siegel worked with IBM to prepare Watson for health care by helping the company understand the medical domain and important sources of information for Watson. He has not been involved in developing the Watson application for WellPoint.

Dr. Nussbaum said Watson's clinical application will be tested in pilot programs conducted among oncologists.

"You have rapidly advancing science, complex decisions, new biological therapies, emerging scientific clinical evidence in what's working and not working," Dr. Nussbaum said. Watson can bring all of those things together and match them to specific patient circumstances to come up with the best probable treatment options for the patient, he said.

Administrative uses

The other use case being tested is Watson's ability to streamline health plan administrative functions and processes. Watson can ensure that treatment is meeting care guidelines and standards in a more timely and efficient manner than what is happening today, Dr. Nussbaum said, which could affect WellPoint's pay-for-performance programs. WellPoint hopes Watson will drive better patient outcomes, which the company said could result in higher outcome-based pay for physicians.

When it comes to rules-based decision making, an intelligent application like Watson can remove plenty of administrative waste, said Dan Mendelson, CEO and founder of Avalere Health, a Washington-based health care business strategy and public policy advisory firm. The evaluation of claims, for example, is currently done by physicians employed by the health plans. "If an automated system can get you 95% of the way, that's pretty good," he said.

How the daily work flow processes would change with Watson is still unclear, Dr. Nussbaum said. Although Watson can process human language, Dr. Nussbaum does not envision physicians using that function. Instead, Watson probably would run in the background of an EMR system. The pilot programs with oncologists, expected to begin in early 2012, will help WellPoint understand how the technology can be used in a clinical practice and how it will fit into existing workflows.

Dr. Siegel said he expects a lot to be learned from the collaboration between IBM and WellPoint. "The implications of all this incremental work, whether it's done for an insurance company or whether it's done in the clinical arena or research arena, I think all those will be helpful for absolutely everyone," he said.

Back to top