More than 70 years ago, Alan Turing, a mathematical genius who helped decode messages sent from the German military command during World War II through the Enigma coding machine, proposed the imitation test to assess machine intelligence. This subsequently became known as the Turing test and involved an examiner questioning and being questioned by a person and by a computer. To pass the Turing test, the computer’s inquiries and responses would need to be indistinguishable from the questions and answers offered by the human respondent. The computer would need to be able to fool the average person into thinking it was in fact another average person.
Over the past few decades, computers that can pass the Turing test have appeared. One area in which this type of artificial intelligence (AI) is still being developed is medicine. Telemedicine is the early form of computer assisted healthcare, and it will probably be replaced in the near future by the android [with a lower case ‘a’] systems envisioned by Turing.
Telemedicine is health care provided without the physical presence of the patient or the doctor or both. Phone contacts or Zoom calls between physicians, nurse practitioners, physician’s assistants and patients are routine now. The advantages to the patient are numerous and self-evident. The patient does not need to dress to leave the home, travel to an office, and wait to be seen in a room with other sick people. Electronic medical records and drug prescribing allow for sharing of histories and test results from a variety of sources and allow medications to be ordered, stopped, or otherwise modified without relying upon easily lost or destroyed paper.
The advantages to the medical professional are also mostly obvious, but there are a few features that may not have occurred to the patient. With Zoom or other imaging programs, the healthcare provider can see the patient in his or her natural setting. The patient is likely to be more relaxed and informative, and the view of the patient’s surrounding, even though that view is usually quite limited, is also informative. One colleague noted that on a Zoom call with a well-respected university professor she was surprised to find the professor surrounded by clutter and the interview frequently interrupted by some of the professor’s thirteen cats.
Imaging studies, biopsies, and invasive procedures are currently being reviewed and interpreted by agencies and individuals far from the sites where they are being performed. This has the potential of improving the care available in areas with limited health resources or providers. It also should raise the ‘standard of care’ in all subspecialties of medicine by making more intellectual resources available in all communities.
Even though telemedicine is widely available and unarguably helpful in providing patient follow-up, healthcare provider conferencing, and even international cooperation, it has shortcomings that are routinely overlooked. The most obvious is the lack of contact between the patient and the healthcare provider. Being in the same room as the person requesting medical help and being able to see in 3 dimensions, hear without distortion, smell, and touch that person is invaluable. It is also inconvenient, and, amongst the reasons it is being done less and less frequently, it is costly.
Obviously, reducing the cost of healthcare is desirable because it increases access and, paradoxically, increases quality. This was illustrated recently by the reduction by an insulin manufacturer [Eli Lilly] of its insulin vials. By capping its charge to many consumers at $35, which is still nearly five times that charged by some foreign manufacturers, it made its insulin more available to those patients dependent on it.
The next obvious cost-saving measure in healthcare is to eliminate to the extent possible the healthcare provider. Insurance companies, hospitals, clinics and other agencies dispensing medical information and making referrals will inevitably introduce machines that pass the Turing test. In fact, they may have already replaced those humans who demand reasonable hours, health and disability insurance, vacations, and other such perks with computers that only demand a little electricity and occasional upgrades. These machines will have much more information concerning your life and problems than any single healthcare provider.
One of the provisions of the Health Insurance Portability and Accountability Act of 1996 (HIPAA), an act alleged to increase patient privacy, was that insurance companies could exchange information concerning your health. The Turing machines would have all of this information available as soon as your healthcare avatar reviewed your caller ID to tentatively identify you, even before the machine said, “Hello, my name is Mary. How may I help you today.” The machines could handle millions of calls simultaneously, and no one would ever again be put on hold. No longer would you hear the absurd notification that “Your call is very important to us.”
That we have any privacy under the current healthcare system is arguable, but with the advent of the Turing machines, privacy will be another item we make reference to as an aspect of the “good old days.” Another likely consequence of the Turing telemedicine computers is the purchase of endorsements. If a patient complains of typical migraines, the computers can all be programmed to make the same recommendations for treatment, recommendations that in a profit driven medical sector (i.e., pharmaceuticals) would inevitably be for sale. Healthcare professionals could of course choose alternate treatments not suggested by the machines, but the effectiveness of drug advertising in directing treatment choices indicates that the drugs listed as the top choices by the Turing machines will be prescribed most frequently. Simply put, the machines programmed to imitate ordinary people are as corruptible as ordinary people.
Many of the multimillionaires who emerged during the first half of the twentieth century supplemented the income of congressmen with what today would be considered bribes. Some of those supported by Henry Ford allegedly suggested that the Federal government pay him ‘subsidies’ for each motor vehicle he manufactured. Ford, being more sensitive to appearances than the legislators on his payroll, advised them to simply allocate more money to build roads and the sales of Model Ts would climb in relation to the miles of roadways available for them. With the rise of the Turing computer healthcare provider imitators, money will inevitably move from advertising to purchasing recommendations. That person helping you on the phone or on your computer screen may be no more real or independent than the Pillsbury Doughboy assuring you that his crescent rolls are the best in the business.
Dr. Lechtenberg is an Easton resident who graduated from Tufts University and Tufts Medical School in Massachusetts and subsequently trained at The Mount Sinai Hospital and Columbia-Presbyterian Medical Center in Manhattan. He worked as a neurologist at several New York Hospitals, including Kings County and The Long Island College Hospital, while maintaining a private practice, teaching at SUNY Downstate Medical School, and publishing 15 books on a variety of medical topics. He worked in drug development in the USA, as well as in England, Germany, and France.