Dr. Watson, I Presume…
The practice of medicine is all about one person caring for another. Plans to use a computer to decide the course of patient treatment remind me of the futuristic, space-aged world described by Douglas Adams in his book The Hitchhiker’s Guide to the Galaxy. According to a recent press release, Watson, the computer that made mincemeat of the flesh-and-blood contestants on the television game show Jeopardy, has gone to medical school. It seems that Watson may be as proficient at processing medical information as he is at processing eclectic trivia. (I refer to Watson as "he” only because this computer was named after Thomas J. Watson, the founding father of IBM). Watson is a "third generation” computer; the first computer able to answer questions that are posed in "natural language” (i.e., plain English sentences). The first generation of computers consisted of calculators (an adding machine used to tally patients’ charges). The second generation of computers were programmable (software generated exam notes and billing statements). You can view a special PBS Nova program about Watson at http://video.pbs.org/video/1786674622/, which presents a novel case study for ethical analysis – one that will require a bit of affective as well as cognitive wherewithal. In fact, it requires you to literally "put yourself in the patient’s shoes.”
Case: Imagine it is YOU, and not some John Doe, 56-year-old Caucasian male bus driver or other fictional patient, who recently suffered chest pains while climbing a flight of stairs. Past Medical History: Fill in your own here. Current Meds: List yours here. Recent Test Results: Copies of your EKG, blood tests, treadmill and heart scans are sent to your private doctor and to Watson. Your private doctor refers you to a cardiovascular surgeon to perform a bypass procedure tomorrow morning. Watson refers you to a cardiologist for treatment with modern pharmaceuticals. Whose advice do YOU follow?
The choice depends on a quick head-to-head, or, in this case, "head-to-CPU” comparison. Watson can process over 200 million pages of information in less than three seconds (that’s the equivalent of reviewing all the textbooks and cardiology journals on your doctor’s bookshelf in about 1 nanosecond). Watson bases his medical decision on processing information from your medical chart, the insurance company’s history of your past prescriptions and treatments, as well as peer reviewed research, population health data from a library of textbooks and journals, and, last but not least, the recommended treatment protocols that are uploaded by WellPoint, Inc., the nation’s largest health benefits company. See the "WellPoint and IBM Announce Agreement to Put Watson to Work in Health Care” press release at: http://ir.wellpoint.com
Your doctor’s decision is based on knowing you for [fill in] number of years. He or she [select one] bases the medical decision for your care on [fill in] years of experience and an intuitive skill to recognize patterns in your signs, symptoms, test results and behavior. Your doctor confides in you that, considering the benefit of immediate surgical intervention, "If I were you, I’d have the surgery.” You don’t want to harm the relationship you’ve developed with your doctor, who has always demonstrated genuine friendship and love for the profession, but you must admit he (or she) is not infallible.
When posed with this very same question about trusting a machine to make our medical decisions, here are some of my students’ responses:
- Watson increases the options available to the patient, which ultimately would serve to improve health outcomes – sort of like the Internet.
- The physician-patient relationship is a very important bond. Health care providers try to establish a safe environment and trustworthy relationship with their patients.
- Although Watson has tremendous capabilities, it does not have a heart.
- How does Watson take the Hippocratic Oath?
- A concern I have regarding Watson being fed with a lot of information and data is that some of the data may not be accurate.
- Too much information can actually be detrimental to decision-making.
- I worry about potential conflict of interest between treatment options that are covered by the insurance company and treatment options that are programmed into Watson.
- I think Watson would be better suited for work in a research lab or university where it can advise young medical students.
- I feel sorry for the individual who sits down at a computer, accesses Watson, types in their medical history and symptoms, and then waits for Watson to spit out a poor prognosis or fatal diagnoses. Computers simply cannot replace physicians.
- I’d like to admit that I would pick the medication treatment plan…not because it was chosen by Watson, but because I prefer less invasive procedures.
All of the responses listed above express the hesitancy of placing one’s trust entirely on a machine (especially when it comes to our own lives), suggesting that any information provided by Watson may best serve to confirm or enhance the doctor’s plan, rather than replace it. Even if Watson can demonstrate statistical superiority over most doctors (as he did over game show contestants on Jeopardy), patients still base decisions on intangibles such as trust and loyalty. How do contextual features such as faith in modern technology and doctor loyalty, patient preferences, values, beliefs and other circumstances enter into the equation except via the doctor acting as a patient advocate? Perhaps there will be a fourth generation of machines that will be capable of affective computing.
Anyone who has read The Hitchhiker’s Guide knows that the ultimate answer produced by the supercomputer, Deep Thought, was "42.” In fact, if you type "the answer to life the universe and everything” (without quotes) into Google, the Google Calculator will give you "= 42.” The ultimate answer is pretty meaningless; so, Deep Thought suggested a way to design an even more powerful computer that would produce the ultimate question. This ultimate computer turns out to be our planet Earth.
The problem is that computers are subject to the "garbage-in, garbage-out” (GIGO) rule of programming. Futurists predict that the GIGO process cannot be avoided (just look at how wrong electronic spelt checkers can be); instead, computers like Watson can only increase the speed at which GIGO-type errors occur. Despite winning on Jeopardy, Watson’s bloopers were hilarious only because lives didn’t hinge on the correct answer. For example, Watson responded "How tasty is my little Frenchman” to a clue about Louis Pasteur, and beat everyone to the buzzer casually saying the "F" word in response to the clue "A German four letter word for ‘no.’ But, Watson is capable of learning, and perhaps with time, will develop a reputation that rivals even the infamous Dr. Gregory House. Maybe future doctors will end like airline pilots who rely on computers to fly planes – a future where doctors will be glorified computer operators. Until then, the practice of medicine is still left to those of us who realize that medicine not an exact science, but a caring profession, which uses science as a tool to help understand the disease process.
Holub P. Dr. Watson, I Presume…. The Internet Journal of Allied Health Sciences and Practice. 2011 Oct 01;9(4), Article 2.