Watson Hiding in Your Data Center
IBM Content and Predictive Analytics for Healthcare offers Watson-caliber analytics without the supercomputer
Illustration by Otto Steininger
Unless you were living on a desert island, you no doubt heard about the IBM Watson supercomputer’s appearance on “Jeopardy!” Maybe you even watched as Watson accomplished a computing first—it beat two top human competitors at their own game. Sure, computers have long demonstrated the capability to quickly retrieve stored data, but only via rigidly structured database queries. On the game-show set, though, Watson did something unprecedented: It demonstrated the capability to understand natural-language queries, digest a broad range of unstructured sources of information in search of answers, and respond with speed, accuracy and confidence.
Maybe you were one of the viewers who immediately began to imagine how such a resource might benefit your business. And maybe when you heard that IBM is now planning to deploy Watson to solve real-world problems, you even wondered how many years you’d have to wait for these research technologies to become commercially available, or if you’d have to spend millions to get them.
Bet you never imagined that a range of companies across the globe is leveraging Watson-caliber performance right now.
Watson-Like Capabilities Now
With IBM’s current analytics portfolio, Power Systems* clients can have Watson-like capabilities without the Watson. And they can have it almost immediately.
“We need to change the perception that it takes several years of effort, a virtually unlimited portfolio of research assets and a supercomputer to build a thoughtful application of these DeepQA and natural-language processing technologies,” says Richard Talbot, director, Power Systems product line management. “We’re learning that we can respond to our clients’ interests using our current analytics offerings and much smaller scale platforms.”
Consider healthcare, for example. The U.S. healthcare industry is shifting away from a fee-for-service model to evidence-based medicine in which provider compensation will be based on patient outcomes and operational efficiencies. This shift will drive new clinical practices and business models, and support better management of patient diagnostics, treatment and outcomes data. Many providers are realizing that the capability to harness decades of medical records has the potential to change the profile of complex disease therapy and improve patient outcomes around the world.
Of course, that’s easier said than done. The volume of healthcare data swells by a staggering 35 percent per year (www.ibm.com/press/us/en/pressrelease/34610.wss). Worse than the volume is the fact that more than 80 percent of stored health information today consists of unstructured data such as physician notes and discharge summaries, email and satisfaction surveys, reference journals, possibly even text messages (http://bit.ly/tDVOeL). This information is stored inconsistently, and is often incomplete and difficult to access.
In October 2011, IBM announced one of the first “Watson ready” offerings, IBM Content and Predictive Analytics for Healthcare (ICPA). This solution enhances clinical and operational decision making by analyzing both structured and unstructured data and past, present and future scenarios in context. Using ICPA, organizations can leverage their own patient diagnostic and treatment records, clinical policies, medical literature, and other sources of qualified information in ways they never thought possible.
“It doesn’t require a roomful of computers and an army of guys in lab coats,” says Craig Rhinehart, director, ECM product strategy, IBM. “You can get immediate value and be on the same path to Watson, both from an infrastructure and a software point of view.”
”You don’t have to build everything from scratch because a lot of integration is done already.” —Craig Rhinehart, director of ECM product strategy for IBM