An HIT Moment with ... is a quick interview with someone we find interesting. Nick van Terheyden MD is CMIO of Nuance Communications.
IBM is hyping Watson after what amounted to one big commercial for it on Jeopardy!. Does it really have immediate usefulness in healthcare?
Anyone who watched Watson outperform its game show counterparts in the original Jeopardy! challenge would agree that its potential in healthcare is both evident and enormous. As with many new technologies, however, there is still much to be done. In fact, it is quite likely that some of the applications for this technology have not even been imagined yet. But either way, it is clear that Watson represents a springboard to revive the initiatives behind artificial intelligence and its application to medicine.
While our vision for this is clear, getting there will involve many additional components and steps that were not part of the Jeopardy! challenge. If Watson is to enter the medical setting, it must first be integrated into the clinical workflow, offering caregivers more complete clinical knowledge that is contextually relevant and immediately available at the point-of-care.
What makes Watson better than the many other analytic tools out there?
Traditional expert systems use forward and backward reasoning, which follows rules from data to conclusions and from conclusions to data. Creating a system around these principles requires detailed logic statement construction and understanding, and needs to include every aspect of the domain knowledge. The process is time consuming and difficult to achieve and maintain in domains with large knowledge.
Watson, however, uses natural language processing, a wide range of search methods, data association, and statistical linking to create hypotheses from data. In the Jeopardy! challenge, Watson was able to consume data and create a knowledge base that exceeded the reigning champions in general knowledge.
In healthcare, we can load Watson with large quantities of clinical source data and rank patient-specific information against a vast matrix of values and identifiers. These observations can then be used to create a ranked list of clinical knowledge relevant to that one unique patient.
Nuance and IBM are working with Columbia and University of Maryland to determine where Watson can contribute to healthcare. How will that process work?
Actually, Nuance entered into a three- to five-year research partnership with IBM and will employ a combined staff of some 30 to 50 dedicated experts, researchers, and engineers from both companies. IBM and Nuance continue to explore ongoing clinical research with a range of partners, including Columbia and University of Maryland. These clinical sites are highly important in capturing the active clinical perspective and to ensure that what ultimately is introduced to the clinical setting aligns with what is needed for successful adoption.
How will Nuance’s speech recognition and Clinical Language Understanding (CLU) be integrated with Watson’s analytic capabilities?
Nuance’s speech recognition and Clinical Language Understanding (CLU) technologies can enable natural interaction and exchange with Watson, and will ultimately eliminate the need for keyboard interaction. Additionally, Nuance’s CLU technology will help to assign additional detail to knowledge that Watson consumes and preprocess patient data making the Watson responses more relevant and accurate.
You’re presenting at RSNA. Can you provide a preview of what you’ll be talking about?
I am excited to be presenting at RSNA this year. I will provide an update on Watson in healthcare — particularly as it relates to the world of radiology — covering key aspects of the underlying technology and what differentiates Watson from other reasoning engines and expert systems. I’ll outline some of the Watson use cases currently under consideration.