According to Money magazine, (March 2011 pages 98-106), the Association of American Medical Colleges has projected that in 2015 there will be 63,000 fewer physicians than it needs. Furthermore, an article in The Atlantic, “Artificial Intelligence Is the Next Killer App” referenced a statement by a Microsoft executive in a New York Times article speculating that future applications might include a “medical doctor in a box” that could help with basic medical issues.
Martin Ford told NPR’s Linda Wertheimer “Radiologists basically focus on looking at visual images from medical devices — things like X-Rays or CAT scans, that type of thing. Now, machines are getting much better at analyzing that type of visual information.”
How concerned or thrilled should I be that the intelligence designated to my future health care decisions will be potentially limited or artificial?
What does the practice of medicine mean? What makes a “good doctor?” What is it you want in your physician? For me, I want him/her to listen to my complaints/concerns. Not all patients know what symptoms to prioritize or what signs and symptoms might be significant or related. Personally, I need a physician to ask appropriate and sometimes probing questions. For instance, a complaint of being tired and unable to sleep should prompt questions such as are you going to the bathroom all the time? This additional information might mean the difference between getting a B12 shot or being evaluated and treated for diabetes or a prostate condition.
In the past, a good physician knew his/her patients. He/she was there at your birth and then at the birth of your children — the Marcus Welby, M.D. or Dr. Kildare who did it all. As medicine and science and technology evolved, the medical specialist was born. That was actually necessary given the accelerated influx of information and research discoveries. To know all areas of medicine thoroughly is virtually impossible. The problem with the specialist scenario, however, is making sure you are going to the “right” one. It is more than a matter of competence. Specialists tend to be very focused and may “listen” only to those symptoms/signs relative to their specialty and assume some other specialist is dealing with “everything else.” In most instances, triage from an astute general internist or primary physician or other health care provider is required.
In addition to listening, there is also the laying on of the hands, palpation, and the physical examination. How many doctors these days even know how to take a good history or perform an accurate physical exam? The typical visit usually consists of an EKG, a stress test, an X-Ray, and a bill.
As identified in the February 26 New York Times article, “Treat the Patient, not the CT Scan,” patients are not “looked at” anymore. As a patient, either in the office or in the hospital, does your doctor look at you, shake your hand or is he/she fixated on the intake information, the chart and/or the monitors?
Medicine has become dependent on lab tests and imaging examinations; the “art” of medicine is being lost. Even more disturbing, is that most laboratory examinations highlight when the results are too high, too low, or abnormal in some way. Medicine is being dumbed down at the same time that technology and communication of results is becoming faster.
Compound all of this with the growing list of non-physician health care providers each with varying levels of education, experience, licensure and credentials, and then add to this list, Artificial Intelligence (AI). Who (or what) has privileges to order and analyze lab and imaging examinations? Can AI be taught to evaluate, e.g., interpret, an “abnormal pattern” on an X-Ray? Will the computer be able to differentiate between “abnormal” and an “abnormal finding” that is artificially created by technique or positioning?
Currently, 4-5 years of training after medical school is typically required for radiologists to be able to make this differential accurately. Before any treatment plan is determined I want to know that a trained experienced physician specializing in what is being clinically questioned analyzes the lab and/or image examination, and is not being compensated or rewarded in some way for ordering the examinations in the first place.
An academic subspecialized radiologist is an unbiased patient advocate. With regard to “Watson” and AI, I cannot help but think of the frustration I feel when I am speaking to a computer regarding directory assistance, a credit card company, a bank, etc. and the computer does not understand my query. I often find myself yelling into the phone that I need a “representative” and the computer responds “sorry, I did not get that, did you say “lost card?” I hope the AI “medical doctor in a box” will understand while I am coughing and saying “I can’t breathe.” I hope the frustration of being misunderstood will kick in enough adrenalin to either help relieve my symptoms or kill me and put me out of my misery.
So where is medicine going? Is it better to sit back and assume that things are changing for the better? More non-physician healthcare professionals and AI to tell me how I am doing? It will be hard for me not to question decisions about my health and wonder who and how the ultimate decision for my treatment is being made.
Add to these concerns, the issue of the decision-making involvement of the insurance companies. It is the insurance company’s employee who will determine what treatment I can have or test that will be authorized for payment. It is hard not to worry that many insurance companies are focused on their financial bottom line. Their executives are very aware that morbidity is much more costly than mortality. I would hate to think the goal of health care reform is being focused not on decreasing morbidity but on increasing mortality (e.g., death squads).
Radiologists have been the earliest physician adapters to technology and I am not anti-technology or opposed to non-physician assistants being involved with patient care but I am concerned with appropriate and inaccurate information being communicated and acted on without validation and appropriate oversight. In my opinion and for me personally, I want trained experienced personnel to listen, examine and treat the patient, not the lab test or the X-Ray result.
So is knowledge power or would it be better to know less and think less about these things. How sad to feel that “ignorance may be bliss.”
Fellow, American College of Radiology; Radiologist in Chief, Hospital for Special Surgery