Editor’s note: This is a guest post by Anna Fredricks, IBM Healthcare Industry market manager.
IBM’s Watson will analyze natural language questions, posed as Jeopardy! clues, on U.S. television, February 14-16. What else would be possible in a Watson-enabled future? What if Watson could help physicians find evidence and scientific support for diagnosis and treatment decisions? What if Watson could act as a physician’s assistant during patient consults, crunching data from similar cases as the physician takes voice notes?
While medical centers don’t have a Watson to interact with, as Jeopardy! host Alex Trebek will, the same Smart Analytics technologies that Watson relies on to answer the quiz show clues is available today and is being applied to solve healthcare problems.
Creating value from health data
While poor information still hinders the ability of healthcare providers to consistently and accurately diagnosis patient conditions, today’s health analytics solutions are giving care providers, administrators and insurers new ways to understand what works, what does not work and why.
For example, the University of Ontario Institute of Technology saw a need to better detect neonatal babies’ subtle warning signs of complications – but doctors needed greater insight into the moment-by-moment condition of the babies.
IBM health analytics solutions (in this case, a first-of-a-kind stream-computing platform) provide the insights that give caregivers the ability to proactively intervene in the case of complications, such as detecting infections in premature infants up to 24 hours before they exhibit symptoms. This capability provides an unprecedented ability to glean insights from vast amounts of data, to spot trends in real-time and to take action.
According to the April, 2010 Harvard Business Review, doctors inaccurately or incompletely diagnose patient illness an estimated 15 percent of the time.
In another example, the University of North Carolina Health Care System decided to connect the complex and diverse information from across its operations.
Sources of UNC’s information range from patient admissions data, to lab results, to radiology images. For this information to be useful across all of its operations, UNC needed to first synchronize these disparate sources into a “single version of the truth” – in the form of a data warehouse later called the Carolina Data Warehouse for Health (CDWH).
“Our key challenge is that every part of the UNC system – the academic medical center, the school of medicine and schools within the university – as well as outside institutions, wants the data that the data warehouse offers,” said Dr. Donald Spencer, Associate Director of Medical Informatics for the UNC Health Care System.
Now, physicians to researchers can feed data, from patient records to X-ray images into the CDWH. The data warehouse’s single-point of access and healthcare informatics capabilities allow researchers, clinicians and administrators from other areas of UNC to analyze the data – leading to improved patient care.
A medical Watson
And in the future, doctors may be able to couple IBM’s currently available health analytics solutions with the Natural Language Processing (NLP) technology in Watson to discuss a patient’s exam results. A medical Watson could dissect information from similar cases, producing diagnosis suggestions, in real time, backed by evidence for the doctor’s review.
Health analytics solutions that exist today could benefit from advances in analytics that come as Watson’s team of researchers push the boundaries of the technology in this Grand Challenge. While I’ll be watching Watson compete on Jeopardy! on February 14 and cheering for the team to win, I’ll actually be cheering for the progress this challenge will provide in terms of improved physician assistance and patient care.