Dec 29, 2011 (03:12 AM EST)
WellPoint Taps IBM Watson For Cancer Decision Support
Read the Original Article at InformationWeek
5 Key Elements For Clinical Decision Support Systems(click image for larger view and for slideshow)
WellPoint and IBM announced in September that they would jointly create commercial decision support tools, using the Watson supercomputer's ability to understand language and analyze vast amounts of information in the blink of an eye.
Cedars-Sinai will help develop the clinical content in WellPoint's decision support applications for breast, colon, and lung cancer. The institute's experts will also evaluate and provide feedback on tests of the decision support tools, but the institute has no plans to use those solutions in patient care.
Dr. William Audeh, MD, medical director of the cancer institute, said in a press release that he viewed the effort as a way to provide oncologists--especially those outside of academic medical centers--with increased understanding of the evolving body of knowledge in cancer research. In an interview with eWeek, Audeh emphasized the importance of new developments in genomic medicine, and he said the hope is that Watson will help clinicians tailor treatments to individual patients.
[ See what else Watson is doing to improve healthcare. Read IBM, Georgia Tech Point Data Modeling At Kids Health. ]
The growing complexity of cancer diagnosis and treatment is a prime reason why WellPoint chose to address this disease first, said Andrew J. Lang, senior VP and chief information officer of WellPoint, in an interview with InformationWeek Healthcare.
"There's wide variability in cancer treatment and a lot of new developments are occurring there," he said. Watson could help physicians keep up with the latest evidence and select the best solutions for individual patients, he argued. And if useless or less effective treatments were avoided, it could lower the costs of treating this very expensive condition, he noted.
Clinical decision support tools of various kinds have been around for decades, but physicians have not used them very much. Why should Watson-based solutions be any different?
"Previous decision support tools were much more static and were very rules-based, and the rules had to be clearly mapped out and programmed," Lang replied. "They didn't allow for a learning-type engine because the technology wasn't there yet."
Watson, on the other hand, can process unstructured information and can learn from its mistakes, he noted. "This change in the use of technology puts us on a path that enables us to take in a lot of the unstructured information that surrounds the medical field and allows us to process that in a way that will be much more relevant to physicians."
Only a small portion of medical studies are relevant to a particular case. But with the help of clinical experts such as those at Cedars-Sinai, Lang said, "Watson can start to recognize which evidence it should give greater weight to and how it should apply that," based on patient records and other factors, including genomics.
Depending on the pace of development, WellPoint hopes to take its decision support tools to some oncology groups for field testing within the next year, Lang said. The company also plans to branch out to other conditions, including additional cancers, cardiovascular diseases, and other chronic ailments.
In the long run, WellPoint and IBM also want to apply Watson across the continuum of care, Lang added. This could involve intervening earlier in the process of treating a condition. "You could also use it to help with care coordination, to move further through your treatment protocols and make adjustments as new information comes in about how a patient is responding."
WellPoint has no current plans to use the Watson decision support tools in evaluating the performance of network providers, Lang said. "But clearly as the [Watson] engine gets more intelligent, and we're able to move to evidence-based decision information for physicians, naturally physicians are going to migrate to using the better information that's tied to the right evidence. And we'll want to have a continued dialog with our providers on how best to apply and use this solution."
When are emerging technologies ready for clinical use? In the new issue of InformationWeek Healthcare, find out how three promising innovations--personalized medicine, clinical analytics, and natural language processing--show the trade-offs. Download the issue now. (Free registration required.)