Integration of artificial intelligence into our personal lives is increasingly becoming a reality. Not only will it help us with daily tasks, but it will also manage our health. Healthcare providers are welcoming the new technology into their practices as a way to provide better quality care to more patients. However, some believe that artificial intelligence may eventually take over physicians’ jobs.
The medical community is no stranger to technology. Machines such as the electrocardiogram, which reads and interprets the heart’s electrical activity as lines and waveforms, have aided providers in caring for patients [1]. Hospitals are now adding artificial intelligence as another layer to current technologies to produce better outcomes. For example, University College London Hospitals (UCLH) and the Alan Turing Institute formed a partnership that will use artificial intelligence to perform tasks such as diagnosing cancer on CT scans that are traditionally performed by physicians. The goal is to not replace physicians, but rather to “allow time for things the human expert is best at”, according to Professor Chris Holmes, director of health at the Alan Turing Institute [2].
Radiology will be one of the first few specialties targeted by artificial intelligence. Radiologists often examine up to 1000 images per CT and PET scans. The work can become tedious and make it prone to error. Jim Brink, radiologist in chief at Massachusetts General Hospital, expects artificial intelligence to act as a “diagnostic aid” in which the technology would flag specific images for the physician to examine [3].
Google is also investing in this area of research. In a recent study published in Nature, Google researchers fed electronic health record data into a deep learning model. It was found that the algorithms predicted the length of a patient’s stay, time of discharge, and also time of death. To the best of the researchers’ knowledge, their “models outperform existing EHR models in the medical literature for predicting mortality (0.92-0.94 vs 0.91), unexpected readmission (0.75-0.76 vs 0.69), and increased length of stay (0.85-0.86 vs 0.77).” Similar to the studies being conducted at University College London Hospitals, the goal of Google’s study is to increase the quality and efficiency of patient care and to decrease the amount of data-oriented work for healthcare providers [4].
With the vast array of research and technologies out there, some may think that doctors may eventually be replaced by these all-knowing machines. However, human doctors may still have a few advantages over robots. A recent study conducted by MIT computer scientists found that a doctors “gut feelings” about particular patients determined how many tests were ordered for those them. After analyzing doctors’ notes on intensive care unit patients, researchers found that doctors take into consideration their own experience, a patient’s history of symptoms, illnesses, lifestyle, and more that are not evident in the data. A doctor’s sentiments predicted how many tests they would order for a patient. When they felt more pessimistic about a patient’s condition, they ordered more tests up to a point. “They’re tapping into something that the machine may not be seeing,” said Mohammed Ghassemi, a research affiliate at MIT’s Institute for Medical Engineering and Science. Researchers hope that this data could eventually develop artificial intelligence systems that could learn to evaluate patients the same way as doctors do [5].
Another area that artificial intelligence may fail in is, ironically, the human aspects of healthcare. No machine as of yet can convey the necessary behaviors and emotions needed in caring for patients. Empathy is necessary for doctors and healthcare providers to connect with patients. Dr. Atul Gawande, author of the New York Times bestseller Being Mortal: Medicine and What Matters in the End [6], recalled an encounter he had with a patient in his third year of medical school during his commencement address at the U.C.L.A. Medical School on June 1, 2018. “We’d been summoned to see a prisoner who’d swallowed half a razor blade and slashed his left wrist with the corner of the crimp on a toothpaste tube,” he said. “The first thing out of his mouth was a creepy comment about the chief resident, an Asian-American woman. I won’t say what he said. Just know he managed in only a few words to be racist, sexist, and utterly menacing to her.” Gawande then proceeded to perform a physical examination of the patient and sutured the lacerations. “I tried to summon enough curiosity to wonder what it had taken to push him over the edge, but I couldn’t. I only saw a bully,” he said. “And I suddenly remembered a lesson a professor had taught about brain function. When people speak...they are, even more, expressing their emotions...So I stopped listening to the man’s words and tried to listen for the emotions.” Gawande continued, “I didn’t understand or like him. But all it took to see his humanity - to be able to treat him - was to supply that tiny bit of openness and curiosity.” [7]
Artificial intelligence has great potential in the healthcare industry. The technology is already being applied in various settings and will no doubt improve patient outcomes in the future. Although these machines have proven themselves powerful enough to potentially replace healthcare providers, they still lack in providing the human touch to care. Human doctors and artificial intelligence will be key to higher quality and efficient patient care. “The foreseeable future is not going to be human vs. machine, but human plus machine vs. human without a machine. The human plus machine is going to win,” said Keith Dreyer, vice chairman of radiology computing and information sciences at Massachusetts General Hospital [3].
References:
Mukherjee, S. “A.I. Versus M.D.” The New Yorker. 3 Apr 2017.
Devlin, H. “London hospitals to replace doctors and nurses with AI for some tasks.” The Guardian. 21 May 2018.
McFarland, M. “What happens when automation comes for highly paid doctors.” CNN. 14 Jul 2017.
Rajkomar, A., et al. “Scalable and accurate deep learning with electronic health records.” Nature. 8 May 2018.
Trafton, A. “Doctors rely on more than just data for medical decision making.” MIT News. 20 Jul 2018.
N.a. “Atul Gawande: About.” Atul Gawande. N.d.
Gawande, A. “Curiosity and what equality really means.” The New Yorker. 2 June 2018.