Dr. Watson? How IBM system can improve diagnosis, treatment.
Connecting state and local government leaders
The system's deep search and analysis ability can help doctors quickly sift through mountains of data in multiple databases.
IBM’s revolutionary Watson technology now has its first commercial application. The sophisticated data retrieval and analysis system, which was seen in February handily defeating the top human contestants of the "Jeopardy!" quiz show, will be helping doctors and clinicians at WellPoint to treat cancer patients.
The implications of Watson’s technology for the commercial and government medical sectors was discussed at a press briefing in Washington D.C., on Sept. 13. A panel of medical experts from academia, government and business discussed the data challenges facing modern medicine and how a capability such as Watson could dramatically change the speed and efficiency of medical treatment.
Watson is designed to answer human questions written in natural language quickly, accurately and confidently. Unlike a search engine, the technology analyzes the meaning and context of a question and then processes specific data, such as medical records, to suggest available options to support a diagnosis.
Related stories:
Why IBM’s Watson is good news for government
IBM’s Watson charts new frontiers for computer science
Until the last few years, there had been little progress or development of health care informatics systems to move medical records from paper to digital formats, said Dr. Eliot Siegel, professor and vice chairman of the University of Maryland’s Department of Diagnostic Radiology. But that change in itself has problems because the digital records are like their paper counterparts in that clinicians must rely on keyword searches. Current systems cannot adequately search across individual medical data nodes to gather information. “We don’t have the capability to do that,” he said.
Watson’s ability to perform a variety of patient analytic services — such as referencing medical health records against known drug allergies — could dramatically change how doctors diagnose patients, Siegel said. The IBM system's deep question-and-answer capability, coupled with its analysis functions, allow it to serve as a doctors’ assistant by performing patient chart reviews, the process of suggesting potential courses of treatment, and citing related reports and documentation to back up its conclusions.
One of the major problems facing doctors is the sheer amount of medical information available, said Basit Chaudhry, a medical scientist with IBM Research. This data overflow is also a challenge to health reform because it affects productivity. He added that as an industry, health care has not been able to increase productivity for 20 years. In fact, productivity — measured in the time and effort it takes to diagnose and successfully treat patients — has either been negative or flatlined primarily due to issues revolving around managing and analyzing massive amounts of data, he said.
As the first organization to use Watson, WellPoint will use it to support its clinical staff, said Lori Beer, the company’s executive vice president for enterprise business services. She said that the technology will help streamline the entire process of helping patients.
The initial deployment of Watson at WellPoint will occur in 2012 and focus on supporting the health care management company’s oncology department. Watson’s search and analytical capabilities will allow clinicians to provide more personalized treatment, she said.
An area where Watson could readily be applied to is health care for the Defense and Veterans Affairs departments. Data is key to health care, and DOD and VA are currently working on efforts to share medical data, said Capt. Michael Weiner, the Navy’s chief medical information officer. The challenge for the government is how to use data to drive better care and lower costs, he said.
Government clinicians can now pull up a great deal of information from DOD and VA medical databases to provide uniform care. However, what if this information could also be used to prompt patients to take more active steps to avoid or treat diseases such as diabetes, Weiner said. A capability like Watson could be useful to both clinicians and their patients. Clinicians would be able to call up patient records and use Watson to review individual health records and compare symptoms and potential drug interactions and other treatments. For patients, Watson could offer health care and dietary/lifestyle suggestions, he said.
One of the possibilities of technologies such as Watson’s is personalized medicine, Siegel said. But to provide such care, clinicians need access to medical databases to tailor treatment. However, to optimize care, Watson must be able to draw data from a variety of medical databases, such as clinical trial studies, he said.