Issue link: http://viewer.e-digitaledition.com/i/60305
++++++++++++++++++++ TECHNOLOGY +++ +++ ++++++++++++++++++++ + + CONNECTING THE DOTS IN A DYNAMIC, COMPLEX AND MULTIFACETED WORLD OF CLINICAL INFORMATION W AT C H ING I BM ' S "JEOPARDY!"-winning supercomputer deliver answers by considering complex, language-based clues raised questions about how a computer named Watson might be used by physicians and other clinicians. While Watson's capabilities made the imagination and expectations of many observers run wild, the answer to whether this will lead to a Watson-like supercomputer eventually replacing physicians and other clinicians is a simple and resounding, "no." However, Watson is expected to emerge as a new kind of tool for caregivers — albeit more dazzling and advanced than any clinical decision support tool we have today. What clinician has not felt the frustra- tion of an increasingly losing battle to keep up with the global, exploding volumes of medical information? What clinician hasn't wished for additional, current and to-the-point advice amid time constraints and complex and ambiguous patient findings? Watson's breakthrough understanding of natural language and relationships among complex, text-based terms and its ability to consider vast volumes of information — including textbooks, medical research, and data about patient populations and individuals — could significantly enhance medical education and the delivery of care. In recent years, caregivers have made limited use of computerized systems to assist their clinical decision making — but the technology has been awkward and intrusive, and working with check- lists and questionnaire-based systems has been too inflexible. Even automated clinical information in the form of clinical reminders, triggered by various data entered into an information sys- tem, can seem rigid, annoying and superfluous in routine practice. Watson, on the other hand, can "interact" in natural language rather than forms and dynamically "layer" every new finding — whether from lit- erature or aggregate or individual patient data sources — into its explora- t ion of complex pat terns and relationships among pieces of informa- tion. It "learns" as it goes. It can then display the results of such exploration, indicating varying degrees of likelihood of possible answers and offering sugges- tions for additional information that will most effectively increase certainty and decrease ambiguity in clinical decision making. A few years from now, consulting Watson could become a routine part of medical practice. It could become a natural choice a caregiver makes, simi- lar to turning to a trusted, expert colleague. And like speaking to a col- league, the suggestions and advice might be expertly well-informed and to the point, but the final clinical decision would remain with the caregiver himself. Consider the process of reaching a diagnosis. Watson might help "uncover" a hidden diagnostic possibility lurking in reams of data about similar patients and conditions or in recent scientific publications, or even in years of records for a complex, chronic patient with multiple comorbidities. Or it may sug- gest the next piece of information needed to resolve a difficult diagnostic ambiguity in a case of atypical symp- toms and findings. In either case, the diagnosis would, with Watson's assis- tance, be reached by the caregiver with greater confidence, precision and speed, translating into greater safety, quality and efficiency of care. Josko Silobrcic, M.D., M.P.H., M.S., is an associate partner at IBM Research, where he serves on a team of physicians developing IBM's next-generation Watson health care capabilities. He has more than 25 years' health care experience serving as a clinician, researcher, educator and industry consultant. He also serves as Course Director in the Department of Health Policy and Management at Harvard University. Dr. Silobrcic holds both a Master of Science in Healthcare Management and a Master of Public Health from Harvard University. He received his Doctor of Medicine degree from the University of Zagreb Medical School. ■ 16 | Greater Kansas MD NEWS ■ MDNEWS.COM