Today we continue on the healthcare theme in the build up to our Miami event on the 15th May. This post is focuses on Watson which captured the imagination of the public last year by winning the quiz show Jeopardy! and exploring how this could transform the healthcare industry.
Watson is an AI system capable of answering questions posed in natural language. In 2011, as a test of its abilities, Watson competed on the quiz show in the show’s only human-versus-machine match-up to date.
Watson beat Brad Rutter, the biggest all-time money winner on Jeopardy!, and Ken Jennings, the record holder for the longest championship streak (74 wins) to received the first prize of $1 million which was donated to charity. Watson had access to 200 million pages of structured and unstructured content.
We were delighted that Manoj Saxena, (General Manager of IBM Watson Solutions) was able to present at our SmartCamp Global Finals in Feburary. Manoj Saxena is an serial entrepreneur who joined IBM in 2006 through the Webify Solutions acquisition (a company he founded in 2002 and led as its chairman, president and chief executive officer). His new task is finding, creating and scaling new markets for Watson’s
On obvious area of opportunity is patient care and public health. How can this technology improve our healthcare systems and what opportunities are there for entrepreneurs working in these areas?
Watson has skills that could be applied to patient care and to public health. It could improve doctor-patient visits as a skilled physician’s assistant. And it could improve the payments process as a clever administrative assistant. It could even help spot epidemics at very early stages by monitoring blogs and Twitter feeds.
The US federal government is making huge strides in computerizing health care. The commitment of billions of dollars has spurred 41% of doctors offices to plan to adopt electronic health records within two years, up from 25% at the beginning of this year.
More than 80% of hospitals plan to use electronic records within two years, allowing them to participate in electronic health exchanges where information about patients can be shared rather than hoarded. The human genome project has harnessed computer power to decode the basic instruction set of human life.
Now it’s time to start using computers to make health care smarter. Watson isn’t ready to be “Dr.” Watson yet. But its core capability of parsing complex text and retrieving meaningful answers to questions should be adaptable to many medical situations. Its technology embodies natural language recognition, information retrieval, multiple layers of analytics, and even advanced betting strategies.
IBM is currently working with university medical experts and partnering with insurers and other health care businesses to develop applications. At University of Maryland’s medical school professors are taking Watson through the same course of study that they give a first-year medical student. They teach it what a symptom is, some basic anatomy and how to read a medical chart and spot abnormal values provided by laboratory tests. From there, they will move on to teaching it how to make diagnoses and how to watch out for drug interactions.
One promising application of Watson will be as a reader of electronic medical records. Once it understands what to look for, Watson could tirelessly proofread EMRs, looking for abnormalities like prostate cancer screening for someone who has had several pregnancies. It could serve as the equivalent of a grammar-checker in a word-processor. Researchers think that capability could be commercially implemented within 24 months.
Within five years, researchers expect Watson will be able to help doctors with diagnosing patients and creating treatment plans. That’s no small benefit. One quarter of all medical errors involve misdiagnosis or delayed diagnosis.
Research shows that 50% of cancer diagnoses change from the first opinion to the second opinion. If a patient receives the wrong cancer treatment, the results can be serious, if not fatal. But cancer research is among the fastest growing bodies of medical literature, and even specialists can’t keep up with the volumes of reports that appear each month. Watson can be relied on to read all the literature and make suggestions based on the latest evidence.
Better diagnosis early also has big cost implications. Cancer treatment costs are growing at a rate of 23% per year compared to regular health care costs which are growing at a rate of 8%. Getting it right quickly saves money.
Within ten years researchers hope that Watson might seamlessly participate in doctor-patient conferences, putting reminders or suggested questions on a doctor’s computer screen. Since many medical terms are predictable, it’s possible that Watson could use speech-recognition to understand what a patient says and use the information to prompt a doctor’s follow-up questions. Would doctors and patients ever be comfortable with having Watson interject its own spoken queries? Maybe.
Watson could also have another role in health care: helping doctors deal with health plans. This is a big issue for doctors. The average practice spends $60,000 a year dealing with payers. Studies show physicians typically spend four hours a week, in addition to five hours a week of staff time, coping with unnecessary complexity in payment systems. Doctors can’t remember whether a patient’s insurer covers a particular treatment or requires pre-approval. Watson could.
Watson would be much too expensive to put one in every physicians’ office. But a cloud-based Watson could become an aid that many doctors would use, if usage was paid for by insurers and Medicare. Watson is designed to be a collaborative tool, and reprogrammed for medical care it could become a standard part of any doctor visit.
Better, faster diagnoses, adherence to payer guidelines and advice on evidence-based medicine would all contribute to controlling medical costs.