While Barack Obama might be having a hell of a time trying to reform healthcare in the United States, Americans perhaps shouldn’t worry too much. IBM’s supercomputer Watson is now being used to fix America’s shortage of doctors.
Watson has had plenty of jobs along the way: as answering phones. But for the past year the team behind Watson has been honing its medical skills, having it scan exam books to learn the basic principles of diagnosis, so that it can make sense of the data held in electronic health records.
The system has already been trialled on the business side of Sloan-Kettering hospital — mainly because there are fewer litigation threats in that environment. But the supercomputer is now going to be unleashed on the Cleveland Clinic. Over the course of three years, the goal is to have the computer trawl patient records, point doctors to important data and perhaps even have it diagnose people.
It’s a smart idea: doctors are overworked, often unable to sift through information in as much detail as they’d like. At the same time, health records are increasingly digital, and more easily analysed by computer than ever before. So, it seemingly makes sense to apply Watson’s skills to the problem. So far, Watson has been put to work summarising medical records, then giving doctors a summary of the patient’s medical history. With some success, too, according to Dr. Neil Mehta from the Cleveland Clinic:
“I’ve had a couple of patients where Watson found things that I had missed. It doesn’t work every time, but it’s getting better.”
There will, of course, be problems. IBM openly admits that medical records are fairly dirty as data sets go: there’s often missing data or strange alternative terms in use, which don’t confuse humans but really hinder a computer. And then, of course, there’s convincing doctors and patients that having Watson do a physician’s job is a good idea. But if there’s benefit to be gained by applying the supercomputer to the problem then surely — surely — it’s worth a try. [Verge]