One of the most established rituals in medicine, the annual checkup, may soon be a memory. Within ten years, the United States will have a shortage of primary care physicians numbering in the tens of thousands, and the problem is set to increase. In advance of a visit to Zócalo by David Lawrence, former CEO of Kaiser Foundation Health Plan and Hospitals, we asked some authorities in the field of medicine to reflect on whether the problem might be alleviated, even made outmoded, by technology. Will technology render the primary care physician obsolete?
Sometimes yes–and sometimes yes and good riddance
Technology will not render the physician’s profession obsolete. Technology will, however, render obsolete the notion that there are things that only doctors can do, as advances like molecular diagnostics and clinical decision-support software continue to enhance the capabilities of other individuals–including patients–to manage increasingly sophisticated levels of care in more convenient and affordable settings. But our health care system, like any other sector, will always need experts on the cutting edge, and doctors will constantly seek out new and complicated medical problems that demand their skill and attention. On the other hand, individually, there indeed will be doctors that are made obsolete by technology; in particular, those that make futile attempts to defend their turf and who refuse to adapt to a changing competitive environment. As Harvard Medical School’s Dr. Warner Slack famously professed, “Any doctor who can be replaced by a computer should be.”
Dr. Jason Hwang is an internal medicine physician and executive director of healthcare at Innosight Institute.
Not obsolete–just different
Technology won’t make doctors obsolete, but it will require them to change in order to provide the best care to their patients. Those that don’t learn to make effective use of these tools may be unable to provide the care patients expect.
The core of what primary care physicians do is diagnose, treat, and coordinate with other clinicians and service providers. Computers can’t diagnose, or do it badly, because they bet on the most likely diagnosis. They can get it right only 90 or 95 percent of the time, and that’s not good enough.
On the other hand, while well-trained physicians can bring more sophistication to the process of diagnosis, using that training requires keeping all the possible causes of the presenting symptoms in mind, all the information needed from the history, physical and basic lab tests to assess each cause, and a systematic way of assessing each possible cause using the data from the history, physical and tests. This is beyond the capacity of the human mind, House notwithstanding. With the right computerized tools, however, collecting the right information and analyzing against all known causes of the patient’s symptoms might become routine. Computers are not ordinarily used in this manner today, but models of how they might be are already in use.
For some conditions, treatment is straightforward. There’s one way to proceed. For others, treatment choices can be complex, and decisions must draw on past experience (does the patient have an allergy to a specific drug, for example) and patient preferences. Computers can assist in making better treatment decisions by making relevant sections of the patient’s history known and by serving as a platform for educating patients and sharing the experience of other patients so treatment choices are well-informed. The expansion of electronic health records may make the integration of patient’s history easier. While there are examples of effective use of computers to improve patient decision-making, this area is still in its infancy.
The most widespread use of computers today is in coordinating care. Orders for tests and prescriptions can be sent electronically and clinicians can communicate electronically and share information. The principal obstacles to more widespread use of these methods are lack of training, cost, limited numbers of user-friendly apps, and payment systems that don’t support this type of coordination. All these can be changed.
Computers and new monitoring technology offer the prospect of routine monitoring of patients away from the physician’s office or hospital, distant diagnosis using monitoring and video technology, and real-time consultation with specialists and other experts. Here, too, changing payment to allow the routine use of such technologies will be critical to their adoption.
Dr. Jack Needleman is professor of health services at UCLA School of Public Health. His research focuses on the impact of changing markets and public policy on quality and access to care.
No way. The primary care physician will still be the critical point of care for the patient. Technology helps, but it cannot replace the doctor when it comes to managing complex chronic illnesses, which more and more Americans are developing as we age (“old parts wear out”) and as we continue with behaviors of overeating, leading sedentary lifestyles (the two leading to obesity), suffering chronic stress, and (among 20 percent of us) smoking. Coordinating the care of these patients is critical for quality, safety and control of expenditures.
We are blessed with enormous innovation and creativity in the medical arena. There is an ongoing revolution as a result of genomics; a promise for the future in stem cells for regenerative medicine; multiple new vaccines on the drawing boards not only for infections but also for chronic illnesses such as atherosclerosis, Alzheimer’s and cancer; multiple new medical devices that are smaller and smaller yet more powerful (witness pacemakers for heart failure, stents and now catheter-based technologies to replace or repair damaged heart valves (wow!); imaging devices that show internal anatomy better than an artist can with a dissection plus imaging devices that can increasingly observe metabolic changes in real time.
These technologic advances are creating five medical megatrends: personalized or custom-tailored medicine; better preventive care; an expanding ability to repair, restore or replace damaged tissues or organs; digital information available anytime, anyplace; and much safer and higher-quality care.
Sounds great–and it is. But the use of these technologies all revolve around the primary care physician, who knows his or her patient and family well, can make the correct choices (in consultation with the patient), and can then direct or coordinate the patient’s care as he moves through various specialists, testing and procedures. This is especially critical for the increasing number of patients with one or more chronic illness such as heart failure, diabetes or cancer who will need a team of care givers, multiple specialists, many diagnostic tests, a myriad of drugs and frequent procedures over many years. These are all potentially expensive and should be used only as truly necessary so as to improve the quality of the patient’s care without excessive expense. It is only a well educated, well trained and committed physician who can do this level of coordination.
Dr. Stephen C. Schimpff is an internist, professor of medicine and public policy, former CEO of the University of Maryland Medical Center and consults for the U.S. Army, medical startups and Fortune 500 companies. He is the author of The Future of Medicine – Megatrends in Healthcare.
Nope–but with computers helping out, your doc will become awesome
One of the very first ‘expert systems’ was designed at Stanford University and designed to mimic an expert in infectious diseases. It was called Mycin. When given the right information about a patient’s symptoms, physical findings and lab tests, Mycin performed very well when compared to human doctors. Designed in the 1970’s, it was a rule-driven system, where the available data was run through a set of preprogrammed rules to arrive at a conclusion.
Today’s versions of expert systems have become familiar to us as we interact with commercial systems–like Amazon, Netflix and the Apple’s iTunes “Genius”–which recommend books, movies and music to us based on our prior purchases and those like us in their massive customer databases. So-called pattern-detection systems are also used to detect credit card fraud and stock trends.
These newer systems rely on ‘smart’ algorithms and the availability of electronic data in large enough quantities that we can reliably identify patterns that predict which books we’d like, anomalous credit card use or medical conditions of interest. These are data-driven pattern recognition systems, and they’re likely to have a dramatic influence on medical care by helping us diagnose patients, identify patients at risk for bad outcomes like readmission, and determine which medical practice patterns are the most effective and cost-efficient. In fact, IBM’s Jeopardy-playing supercomputer, Watson, will soon begin to work on health care information to do just this.
Electronic tools are proliferating rapidly in medical care. New electronic medical record and order entry systems allow us to collect data in machine analyzable formats, making data analysis much easier and more immediate. Surgical robots and telemedicine are possible only because of technological advances. New medical buildings will have sensors built into the walls to detect patient movement and track providers.
But while all of this might suggest that the doctor will soon become obsolete, that is not the case. As in other highly automated industries, like aviation, manufacturing and finance, the best medical care in the future will rely on a marriage of human and mechanical capabilities where each complements the other, making the whole greater than the sum of the parts.
Dr. C. William Hanson is professor of anesthesiology and critical care and chief medical information officer at the University of Pennsylvania Health System. He is also the author of the book Smart Medicine.
Yes, in theory–but in practice we love nice, caring doctors
I could give you great advice about choosing a physician.
I don’t necessarily take it myself.
I understand our country needs to shake loose from the fragmented, disorganized, expensive, proceduralist-oriented, fee-for-service health care system. We need primary-care-focused systems in which we reward doctors for working together to coordinate our care, not just for prescribing more and more stuff. Technology will be important. But it’s a tool, not an end in itself.
I would tell you to find a doctor who is part of a larger health care organization–a medical home in the making, a budding ACO, a multi-specialty physician group. Find someone who is excited, not frightened, by coming changes in health care .
I would recommend a physician who has installed and mastered health information technology, not only to streamline paperwork but also to understand, monitor and coordinate health. Health IT is not a panacea, but it sure would be nice to have all those records and MRIs in one place–and have that place not be a dusty shelf in my bedroom closet.
Yet both of my docs are solo practitioners without health IT. I chose the primary care doctor when I moved to D.C. in 1994 by looking on a list–paper back then–for a female physician near a Metro stop who had gone to a good med school. I found my OB/GYN through a friend.
I stuck with my primary care doctor after she had found a minor thyroid problem and I asked whether fixing it would mean I wasn’t going to be tired anymore. She asked how old my son–then six–was. “You’ll be tired for 12 more years,” she replied.
I stayed with my OB–even after she stopped taking insurance–because she saved the life of my second son during a high-risk and utterly miserable pregnancy.
I assume both of these physicians will, over time, migrate to more technology, and both will adapt to change. In the meantime, they provide me with something that a machine can’t provide and that our system needs more of. They listen. And they care.
Joanne Kenen is a journalist and author who has covered health policy in and out of Congress since 1994. She was the senior writer for New America Foundation’s Health Policy program from 2007 through 2010.
*Photo courtesy of a.drian.