@font-face { font-family: "Cambria Math"; }@font-face { font-family: "Cambria"; }@font-face { font-family: "ヒラギノ角ゴ Pro W3"; }p.MsoNormal, li.MsoNormal, div.MsoNormal { margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: "Times New Roman"; color: black; }.MsoChpDefault { font-size: 10pt; }div.WordSection1 { page: WordSection1; }
ARMONK, NY – The ease with which IBM’s Watson computer dispatched its human opponents on Jeopardy! this past February was something to behold. Indeed, for those inclined to wonder darkly about the eventual obsolescence of homo sapiens, it may have been a bit disconcerting.
But Watson only wants to help mankind. Or would, were he actually a sentient being. Which, of course, he’s not. But the advanced technology at work in the supercomputer could have a transformative impact on healthcare delivery.
That’s why so many people are excited that IBM and Burlington, Mass.-based Nuance Communications are partnering to explore how Watson’s deep question answering, natural language processing and machine learning capabilities – combined with Nuance’s speech recognition and clinical language understanding technology – could help support clinicians in the diagnosis and treatment of patients.
Having enlisted the Columbia University Medical Center and the University of Maryland School of Medicine to help identify Watson’s usefulness, the companies hope the first commercial offerings from their collaboration will be available within two years.
Watson’s ability to glean meaning from written or spoken information and put it into context could be revolutionary.
Because, while all computers are good at dealing with numeric variables, “this is technology that can leverage unstructured data,” said Josko Silobrcic, MD, associate partner at IBM Research and a professor at Harvard University School of Public Health, in an interview at HIMSS11. “It can leverage text and 'human form' communication, which is huge. That’s what data look like in healthcare. Tremendous amount of it – some of the most important stuff – is in textual form.”
That capability allows Watson “to take a tremendous amount of baseline information” – such as medical textbooks and periodicals – “that directly impact healthcare” and digest and contextualize it.
It can then combine that knowledge with the acquired experience of a particular healthcare delivery organization and, most importantly, a particular patient’s medical history – potentially providing clinical decision support for doctors.
“The volume of healthcare information is increasing tremendously and probably accelerating,” sayid Silobrcic. (Consider, for example, advances in genomics, which has led to a huge uptick in data complexity.) “That exceeds the training of healthcare providers, as well as their ability to keep up with it. We all know how busy clinicians are.”
Watson’s value lies in its ability to sift through this evolving knowledge base dynamically – shifting and refining its understanding. Medicine is fraught with ambiguity. Physicians, unsure of the precise diagnosis or treatment for a patient, routinely seek out ways to help inform their decisions.
“They call each other for consultations,” said Silobrcic. “Or they might look something up on the Web or open a textbook. But wouldn’t it be terrific if they had a tool that could sift through vast quantities of information and come back with suggestions – with options?”
Paul Ricci, chairman and CEO of Nuance, said the technology “will introduce unmatched clinical information and analytic technological advancements for healthcare.” He said “the solutions we are developing with IBM will transform the capture, flow and use of clinical data, empowering healthcare organizations to drive smarter, more efficient clinical and business decisions.”
“I’m very excited about Watson's potential,” said Eliot Siegel, MD, director of the Maryland Imaging Research Technologies Laboratory at the University of Maryland School of Medicine. Watching Jeopardy, he said, “I was really amazed, along with the rest of the world, at how accurately and rapidly Watson was able to respond to the complex questions that often involved wordplay or puns in a way that seemed to give the impression of an understanding of the meaning of the question.”
Siegel said Watson could first be used to flag safety errors or inconsistencies in the EMR, such as patients on wrong or conflicting medications. As the technology evolves, he said it could “begin to suggest diagnostic tests or even therapies” or “perform real time interaction with the physician while she or he is seeing a patient – in the ER, for example.”
Silobrcic said Watson's healthcare applications are “still in development” and at least a year or two away from “supporting clinicians from day to day.”
As researchers fine-tune its uses, “there’s talk of focusing Watson on certain smaller domains,” such as oncology or cardiology, he said. “By virtue of limiting the domain, one might be able to more quickly exercise all the capabilities that Watson has in a more constrained space.”
Silobrcic said it’s “very conceivable that even a physician office in a rural area might be able to access Watson as a resource through cloud computing” or that “a very large integrated delivery network might decide to have a version of their own.”
“Watson has the potential to help doctors reduce the time needed to evaluate and determine the correct diagnosis for a patient,” said Herbert Chase, MD, professor of clinical medicine at Columbia University College of Physicians and Surgeons. “We also believe that Watson also has the ability to help doctors provide personalized treatment options that are tailored to an individual patient’s needs.”
Ultimately, said Siegel, “I have little doubt that technology like Watson’s will be routinely used for safety, surveillance and diagnostic assistance in the next 10 years and am excited to see the renaissance in ‘artificial intelligence’ in medicine.”
