Elementary, My Dear Watson

I have previously confessed that one of my guilty pleasures is reading People Magazine.  Another is my willingness to devote time to watching TV game shows.  They fascinate me–everything from the badly coiffed host to the goofy jumping up and down of the winner.  It’s a little embarrassing to admit how much I like to be the first to get the Wheel of Fortune clue or the only one who knows the answer to some obscure Jeopardy question.   So naturally I was one of the people who tuned in to watch IBM’s Watson computer kick serious smart guy butt on Jeopardy last month.

For the 8 people who missed the public relations blitz, Watson, brainchild of IBM, is an artificial intelligence supercomputer that can directly and precisely answer natural language questions over an open and broad range of knowledge. It is programmed to have more information resident within it than any army of people could have and it has the means of accessing, culling through and integrating that information faster than the speed of light, or at least faster than the average Jeopardy champion’s reflexes.   According to IBM, the secret sauce of Watson is its ability to perform deep analysis of large volumes of content to deliver answers characterized by high accuracy, confidence and speed.

Watson was named after Thomas Watson, original founder of IBM.  However, it could just as easily have been named after Dr. James Watson, physician sidekick of Sherlock Holmes, given IBM’s plans to transition Watson from Jeopardy champion to healthcare revolutionary. Of course, in the Sherlock Holmes-Dr. Watson partnership, Watson was the intuitive one who could read and relate to people, whereas Holmes was the emotionless analytical half of the team, capable of sifting through massive amounts of information to find the hidden nugget of actionable data.

And it is this anomaly that really stuck in my mind as I thought about the whole Watson thing as wave upon wave of articles appeared after the Jeopardy slaughter.  Most of these articles, when not reminding the world that even Ken Jennings is fallible, discussed how Watson could revolutionize medicine by improving diagnosis and treatment and potentially replacing imperfect human decision-making with technology.  One article said,

How about this scenario: a patient is rushed to a hospital with an unidentified illness. A loved one explains to Dr. Watson the symptoms, and the good doctor searches terabytes of information in a split-second to narrow down the possible ailments, along with possible courses of action. “This patient would appear to have X, please prepare a drip of Y to begin treatment.”

Those who know me know that one of my stock jokes is to say that the biggest problem with medicine is the doctors, but seriously, I was only kidding.  I can imagine nothing worse than a healthcare system where real caregivers were replaced by unfeeling analytical machines, further reducing the time patients spend with their doctors.  In fact, that is already many patients’ biggest beef with their healthcare experience in a world where the average physician visit is less than 10 minutes.

IBM has been quick to say that it is not their intent or even belief that Watson will replace all or part of the doctor-patient interaction, but the substance of some of the discussion gives me pause.  Some out there have suggested that Watson could help make up for the vast shortage of primary care physicians we are experiencing by creating a surrogate master diagnostician.  Perhaps, but I fear that this line of thinking may lead to not only to short-cuts in the diagnostic process, but shortcuts in patient-provider interaction that might actually make treatment worse.

In a NY Times article about Watson’s healthcare application ,  Columbia University clinical medicine professor Herbert Chase (Columbia is partnering with IBM on the medical application of Watson) said:

“Watson has the potential to help doctors reduce the time needed to evaluate and determine the correct diagnosis for a patient,” He and others in the field noted that a device like Watson could prove infinitely helpful in the practice of medicine, where clinicians have to juggle facts and analysis virtually every minute of the day.  Chase added, “I have been in medical education for 40 years and we’re still a very memory-based curriculum.  The power of Watson- like tools will cause us to reconsider what it is we want students to do.”

Very true, but I’m a little worried that today’s crop of physicians, brought up on Facebook, Farmville and World of Warcraft are at risk of missing the lesson on the human part of medicine. While it would be amazingly valuable to have a system to augment the physician’s own knowledge base, particularly when applying the world of evidence-based medicine in conjunction with an individual patient’s detailed medical record, it is equally important for physicians to do some of that good old-fashioned laying on of hands, giving patients thorough physical examinations and really listening to their presenting complaints.

I recently read a story that really brought this message home in a very tangible way.  The story was written by a woman who volunteered to let medical students learn by listening to her unusual heart murmur.  18 students lined up to take their turn.  The first 9 students that listened to her heart never made any significant connection with the woman, treating her not much differently than they might a cadaver.  The 10^th student bothered to introduce himself and make a personal connection.   Naturally, this was the first time in the process that the woman felt at ease and responsive to the student’s comments about her condition.  The senior physicians teaching the students said to the woman, “There are some doctors who do not view you as a patient. They view you as someone with a heart murmur who they’re going to listen to.”   While technology may afford us a better way to detect and treat the heart murmur, it is the intuitive human connection between patient and doctor that will lead the patient to acknowledge and comply with the treatment necessary to make a difference in outcome.

Furthermore if Watson or a system like it is intended to be the know-all integrator of medical knowledge, how do we ensure that all the right information goes in?  Who will be the arbiter of what aspects of medicine should be included?  While Watson will be undoubtedly be chock full of the latest and greatest from the Journal of American Medical Association and the New England Journal of Medicine, will it also be able to distinguish whether a patient’s presenting complaints are actually stress-related and not truly “medical” in nature?  Will Watson be able to recognize that the reason the patient isn’t improving is because they can’t afford their prescription and thus have been cutting their pills in half, rendering them useless?  If someone just plugged the symptoms into a computer system, this critical fact might be missed.  It is only by really listening and querying a patient that the doctor can know the answer to the deeper questions that lay beyond symptomatology.  Watson may make medical diagnosis faster and more comprehensive in certain cases, but it can never be more than one tool in the hands of a real professional who also applies his/her listening and sensitivity skills to get a full picture of the situation.  Plus, I am suspicious of any computer that can be outfoxed up by a member of the U.S. Congress.  I don’t care if he is a rocket scientist.

As I was pondering all this, I happened to catch Dr. Andrew Weil on the Dr. Oz show and he and a panel of his colleagues spoke about the powerful potential of integrative medicine, which takes the importance of the physician-patient interaction to the next level entirely.  In this approach, physicians take a holistic view of each patient, taking into account their stress level, financial and psychological challenges and social support systems, in addition to their physical complaint in order to effectively diagnose and treat.  The integrative medicine model suggests that treatment should incorporate the best of both Western and Eastern medicine, food science, psychology and other clinical disciplines to ensure the best outcome.  Equally important, Weill said, is the physician’s duty to move the conversation with patients from one focused on treatment to one focused on keeping them well and preventing disease.   Obviously this takes a lot more finesse than what could be delivered by a computer, even one as slick as Watson.  It is this change in patient and provider perspective that is essential to the economic health of our nation.  If we can transition people to acting on their health before it’s impaired, we can avoid a significant amount of healthcare cost, reducing unnecessary spending and leaving valuable medical resources available for those who really need them.  Watson may well be a great tool for figuring out what to do when a patient is truly ill, but only real people are going to convince other real people to make fundamental lifestyle changes that encourage wellness and prevention.

Clearly it is the combination of characteristics present in both Sherlock Holmes and Dr. Watson that makes for an optimal healthcare provider.  This is probably why the two characters are virtually inseparable when it comes to their detective story exploits.  While both Watson and Holmes had their strengths, the combination of their complementary attributes was far more effective.  Analytics and emotion, logic and intuition, science and sensitivity…together they make for powerful fiction and even more powerful medicine.