With increasing frequency, the use of artificial intelligence is becoming commonplace. From Alexa to self-driven cars, machines that perform human-like tasks are invading all parts of our daily lives. In medicine, AI has opened a whole new frontier of potential uses that has the capacity in the future to be involved in nearly every aspect of health care.
While AI assistance to the medical profession is new and not widespread, the general concept has been around since the 1950s. Artificial intelligence is not the Hollywood science fiction depiction of robots doing manual chores, but rather it is a program generated by computers that develop intellect, formal reasoning and learning through analysis of mounds of data for use in a specific application.
These computations can produce software that improves specialized medical tasks. AI is better suited for some functions more than others. It is easier to apply AI to those clinical circumstances that have established criteria and are based on complex, observational data. That is why AI has found early success in radiology and pathology.
Other physician-led assessments like diagnostic pathways, prognosis and therapy are more difficult. AI, for these aspects of patient care, is not ready to be used by clinicians.
Historically, physician acceptance of new technological advances has been mixed. In 1819 Rene Laennec invented the stethoscope. It was the first real tool that helped in the physical evaluation of the sick patient. Because of its uniqueness it achieved rapid international popularity.
However, other 19th-century diagnostic instruments like the ophthalmoscope and otoscope initially had only limited penetration into general medical practices. Even Joseph Lister’s life saving concept of surgical antisepsis took decades before it was embraced by the conservative surgeons of the time. On the other hand, the discovery of X-rays in 1895 had its medical utilization immediately understood and put to work to help patients.
Today, some practitioners are reluctant to consider integrating AI operating systems into their medical practices. Recent journal articles suggest that this skepticism may be based upon a concern that AI may harm the doctor-patient relationship. But not all authors espouse that sentiment. They suggest that AI eventually will assume routine tasks previously performed by the physician. This will give additional time for patient interaction.
But before AI can be successfully coalesced into health care on a universal level, three obstacles need to be overcome. To do that will take time and significant financial investments.
First, there are often multiple divergent but appropriate pathways to diagnose and treat particular clinical syndromes. Because AI requires consistent similar data to be evaluated, a consensus standard for the evaluation and treatment of specific entities needs to be agreed upon.
Second, as with other medical innovations, AI-driven tasks must undergo testing. Demonstrating improved outcomes, clinical effectiveness and that there are no unanticipated consequences is mandatory. In addition, the economic impact of the new tool needs to be considered.
Third, to be accepted by both providers and their patients the technology industry has an obligation to recognize that AI integrated into clinical care will not always be perfect. A misjudgment by a computer program playing chess has little importance. But like the computer-enabled driverless car that errs and hits a pedestrian, a fatal mistake made by an AI-empowered clinical determination would be devastating.
Commingling AI into medicine will not reduce the need for physicians. It will improve productivity and precision. But evaluating the infirm involves more than making data-driven decisions. It includes empathy and understanding that only a doctor, standing by the patient’s side, can bring to the process.
So, when will artificial intelligence be available in your physician’s office? Ask Alexa — she will know the answer!
Stolz is a retired physician and author of the book “Medicine from Cave Dwellers to Millennials.”