Two years ago IBM Watson competed and won on Jeopardy against two of the shows most successful contestants. Watson was able to achieve this feat by using natural language processing and big data to comprehend the questions and then come up with the correct answer. Since this initial historic achievement IBM has been working on making Watson work in the real world. Now Watson is working with Doctors at the Memorial Sloan-Kettering Cancer Center (MSKCC) to help the proliferation of medical information and improve health care efficiency and quality.
Last week I was able to attend a briefing where IBM showcased how Watson is proliferating medical information and improving health care efficiency and quality. While at this briefing I kept thinking about how this was a perfect real life case study of using big data and how it fit in with SIIA’s recent white paper Data-Driven Innovation A Guide for Policymakers: Understanding and Enabling the Economic and Social Value of Data.
The briefing was led by Dr. Martin Kohn, the Chief Medical Scientist of IBM and Dr. Mark Kris, the Chief Thoracic Oncology Service at MSKCC. During the briefing they showed us how Watson is able to use a patient’s record and look at relevant data to come up with a list of potential treatment plans and their odds of being successful. If important information is missing Watson lets the doctors know what information it needs in order to make a decision. Over time as the patient has new symptoms or gets back the results of certain tests or treatments or expresses preferences on treatment Watson takes all of these things into consideration when coming up new treatments and their probabilities of success. Additionally based on the information Watson has received it can diagnose or change the diagnosis of a patient.
Dr. Kris believes that Watson is successful at diagnosing and offering treatments because it looks at everything not just what people believe are important. The other reason he believes Watson is successful is because it goes about things the way a doctor would such as giving a list of possibilities not one definite solution and the likelihood of various treatments being successful. Watson has the added ability to look at information collected by doctors in the field around the world and use their cumulative knowledge instead of just relying on what a few specific doctors at that hospital know. Just like with people Watson is able to learn and remember things so the more patients it works with the better it is able to do in the future.
While these initial results of transforming Watson from a games show winner into a doctor have been promising there are still many problems they have to work on fixing before using Watson at the hospital becomes a common occurrence. The two biggest of which are that for Watson to come up with diagnoses and treatments requires it to analyze and store massive amounts of data which is very costly to do at the moment. The second is that at the moment they need to figure out how to best maximize the use of Watson as it is only capable of working in a narrow field at the moment such as cancer instead of in the broader field of healthcare. Both Dr. Kohn and Dr. Kris stressed that Watson at the time is a tool that can be used to support or come up with a second opinion on things but is not a substitute for an actual doctor.
At the moment Watson is a useful tool at the MSKCC but there is a still lot of work that needs to be done before it is able to potentially revolutionize the healthcare industry. The most important thing is to remember the use of big data to create Data Driven Innovation to create real world benefits is still in the early stages and the best thing we can do is to not put restrictions or limitations on how or why it is used or collected so that we don’t accidentally prevent monumental changes in how we do things from happening.
Denys Emmert is the Public Policy intern at SIIA. He has a degree in marketing and political science from Florida State University.