Watson, M.D.


“Snakes are sometimes perceived as evil, but they are also perceived as medicine. If you look at an ambulance, there’s the two snakes on the side of the ambulance. The caduceus, or the staff of Hermes, there's the two snakes going up it, which means that the venom can also be healing.”

-Nicholas Cage

For the greater part of the past 2000 years, medicine and technology have had an at times fickle relationship. Technology is often a harbinger of change, and as with most large institutions (which includes the realm of medical science), change is often looked upon with skepticism, and brought about in a slow and methodical manner, if at all. But what if the change in question was prompted by an innovation, insight, or advancement so profound (think Einstein’s Annus Mirabilis papers, Darwin’s Origin of Species, or Edison’s incandescent light bulb) that it would be a fool’s errand to ignore it? A seismic event on the level alluded to occurred in February 2011 and has paved the way for an explosive increase in the efficacy of medical provision worldwide. I am talking of course about the public debut of IBM’s Watson supercomputer.

Watson, the brainchild of IBM visionaries Paul Horn and Charles Lickel (former president and vice president of software research at IBM), captivated audiences on an even grander scale than its predecessor platform for public relations, Deep Blue (the computer designed to compete in chess against Russian grandmaster Garry Kasparov), by going toe to toe with the best the show Jeopardy had to offer on national television (and in the end winning handily). Watson, at its basal level, possesses and performs three core functions: natural language recognition, hypothesis generation, and dynamic learning. This is accomplished through the cohesive integration of advanced IBM software and hardware in conjunction with a vast array of data.

This is all well and good, but if you are a physician and have read this far, you are probably thinking at this point, “Why should I care? What’s in it for me? This isn’t the link to ICD-10 updates!  (You’re welcome).”

Watson has the potential to provide a broad range of medical applications both in the immediate future and in the long term. David Ferrucci, the man charged with leading IBM’s Watson design team detailed several of these applications in an interview he conducted in late 2011. It was conducted a half year after the computer’s breathtaking media splash. The core goal explicitly stated by Ferrucci was to “combine patient data symptoms, office visits, background databases, [and with these metrics] computing these profiles, providing these profiles to the medical teams and helping them to see how the evidence supports the hypothesis.” While this serves as a founding tenant around which Watson’s capabilities in the medical field will initially be geared, there are no doubt a multitude of individuals pondering their own notions for customization of the Watson computer system. One example of this touched upon by Dr. Nathan Edwards, a computational biologist at Georgetown University, was the possibility of customizing Watson’s capabilities to conduct automated image analysis, in effect further supplementing and potentially altering the roles and tools at the disposal of present day pathologists.

Now that you have either been exposed to or reminded of the potential to be found in Watson’s cutting edge capabilities, the pragmatist in you is probably starting to think “what’s the cost?” “Is there any way I could actually save money with this technology?” The answer is unequivocally yes.

IBM recently partnered with Wellpoint Inc. and has been adapting Watson technology for use by this insurance company, a subsidiary of Blue Cross Blue Shield, for the past few years. The overarching aim of this collaboration is to help facilitate the provision of empirical data for or against the use of various procedures or treatment regimes. This capability is essential, with over 1 million scientific articles that are published each year, a number that most certainly prohibits comprehensive review without sophisticated technology. By combining the indoctrinated intuition of the world’s practicing physicians with the multitude of useful functions that have been developed for IBM’s latest and greatest technology, medical administration will become not only more streamlined, but also more comprehensive. This in turn, will most likely lead to easily accessible and readily shared medical insights, as well as decreased costs borne by both the medical infrastructure and the patient.

While it is prudent to question the motives for change, to deny it entirely, especially when a net benefit can be gleaned, would be tantamount to an ostrich burying its head in the sand when frightened by a new threat. To those who recognize the benefits of evolution and correctly weigh them against their perceived risks and limitations (and these are certainly not absent from the attempted integration of Watson technology into the medical field), they can truly expect to make the world a better place. One need not look further than the likes of Dr. Barry Marshall with his H. pylori research, Galileo with the field of astronomy, Louis Pasteur with vaccinations, or Steve Jobs with the computer industry to see this ethos in action.

Since the debut of Watson three years ago, several groups have already tried to position themselves to be a part of this trendsetting mold by investing sizable financial and academic resources geared towards tapping Watson’s potential. These include major collaborations with University of Maryland Medical School and Columbia’s College of Physicians and Surgeons to discuss, identify, and develop potential applications. Meanwhile, IBM has recently committed 1 billion dollars to further drive innovation via its Watson applications (many of them geared towards the medical realm). This on top of partnerships with companies like Nuance Communications  and Welltok to develop and commercialize personalized health platforms, Sloan Kettering Memorial Cancer Center, Cleveland Clinic,  and the Maine Center for Cancer Medicine which have customized aspects of Watson’s three core capabilities mentioned earlier in the article ( natural language recognition, hypothesis generation, and machine learning) for medical application. These substantial investments by both IBM and the medical field will spur both collaboration and competition in the years to come, which will no doubt lead to significant amounts of innovation in a variety of fields. Hopefully this momentum will lead to even more entities, both public and private, committing R&D funds to spur advancement.

Now this all seems well and good; but since you, the readers are presumably smart, savvy individuals (except for this guy), you no doubt have considered a few potential causes for concern with Watson’s potential new role in the medical field. Let me attempt to address what many might perceive as some of the most common hang ups with Watson and its implementation. The first is the notion that this new technology will be too advanced for physicians; especially more seasoned ones, to effectively grasp. (Let’s go ahead and read that last sentence again…physicians, especially the more wizened leaders in their field, by nature relatively smart people, cannot handle the intellectual rigors inherent with embracing a new technology. You have my permission to laugh now…go ahead, I’ll wait…)

Now that we have that out of our system, let’s go ahead and address this problem in a proven,  somewhat stereotypical manner. Readily available and easily digestible training for any Watson based technology applications (provided by IBM or some subcontracted company), is a quick and easy solution to this conundrum. Another concern you intellectuals may have pondered is that with this transition to more automated functionalities, might we have opened ourselves up to more IT corruption? The answer, most obviously, is yes. However, the same question can be asked and answered in the affirmative to the use of the internet in general, and most would agree the reward of its usage outweighs its risk (though one might choose to err on the side of caution by beefing up IT security). The more kind hearted readers might also be concerned with some of their staffers losing their jobs, however, this can lead to less overhead cost for practices, which in turn can lead to more effort and training put in the staff that remains.

Watson certainly has the ability with the appropriate modifications to lessen their workload.  Those of you with a macroeconomic background might think that IBM will develop an unhealthy monopoly in the marketplace/medical field, however, as discussed earlier in the article, this would inspire competition to provide similar services at a discounted rate from another company (or from the public sector), just look at how well things turned out with the human genome project with two separate entities, one public one private, striving towards the same goals. This sentiment is echoed with the implementation of electronic health records over the past decade. Anyways, these concerns and many others like them lend themselves to being quashed by a myriad of different and completely logical solutions and explanations.

To the readers, I hope I have painted a pretty clear picture for you (even if it was paint by numbers) of some of the benefits and potential pitfalls that come along with embracing IBM’s Watson technology. If properly implemented, it is my belief that this technology will radically change the face of medicine as we know it, and my hope is that you will become a patriot in this revolution, and potentially be remembered as a founding father or mother in this movement when your descendants read about this new age of discovery in medical informatics and analytics.

John is a senior staff writer for MedicalGroups.com specializing in the fields of Biomedical and Healthcare IT. He is a graduate of Boston College and Georgetown University, and has spent time as a lab rat in both the academic and corporate worlds.