Before I was old enough to attend school, I received medical treatment from a general practitioner (G.P) on at least three occasions. He made house calls and trudged up the three flights of stairs to our apartment without complaining. He wore a suit and carried a black bag that was filled with instruments and medicines. On each of those three memorable days, I had a high fever, limb pains, a sore throat and nausea. On each occasion, he looked at my throat, listened to my chest, pressed on my belly and injected my buttock with something he called ‘penicillin’ for what he called ‘strep throat.’ He took no cultures, ordered no blood tests, recommended no follow-up, and got paid in cash. I believe he got $10, which was equivalent in purchasing power to more than $125 in 2023 dollars. In 2023, a similar assessment and treatment would be considered woefully inadequate and, if any complications occurred, would have justified a malpractice suit.
Several years later a surgical resident [doctor in training] explained to me that a surgeon was a general practitioner who operated on people. General practitioners and surgeons would both treat people with belly or chest pain, sore throats, irregular periods and all other sorts of complaints. Those seeking treatment from a surgeon would, not surprisingly, often end up having an operation. Those treated by a G.P. were likely to be prescribed a variety of medications targeting their complaints. The frequency of unnecessary or inappropriate operations, such as appendectomies, was high. The frequency of adverse drug effects or interactions was also high. As innumerable subspecialties in surgery and general medicine developed, the elimination of the ‘generalist,’ whether he or she be a dispenser of medicines or operations, was inevitable. The explosion of medical and surgical information and treatment options in recent decades has effectively replaced the general practitioner with what is best described as the general manager, a physician, physician’s assistant, nurse or other health professional qualified to triage and direct sick people to the appropriate specialists but not directly involved in the delivery of care.
Current technologies allow for more comprehensive patient assessments remotely than were practical just a few years ago in person. Implanted devices can record heart activity for days, weeks, or even months while the patient is at home or work. Computers can analyze the tens of thousands of heart beats in seconds and identify a worrisome discharge in the midst of thousands of unremarkable ones. Brain waves can be similarly monitored and assessed. Blood sugar can be measured and insulin dispensed as needed without the diabetic person needing medical visits. Medications previously requiring a physician’s prescription are becoming increasingly available over-the-counter.
Even procedures requiring surgeons in the past are becoming less dependent on physicians trained in the esoteric arts of cutting, cauterizing, stapling, and sewing. Cataract surgery has become virtually hands-free. Arterial stent placements and vascular anomaly embolizations have replaced surgeries on clogged arteries and the dissection of hazardous collections of arteries and veins. Kidney stones are blasted out of existence with sound waves delivered through the skin. Unsightly facial wrinkles are dissolved with Botox injections, often provided without medical supervision.
Of all the technological advances that are changing the character and delivery of medicine, the most important is, of course, the computer and, more specifically, artificial intelligence. No individual can retain all the information needed to practice medicine in 2023. No individual can even recognize all the signs, symptoms, imaging study results, blood test abnormalities, etc. that define many of the currently recognized diseases or surgical emergencies. Artificial intelligence provides more information in a few seconds than a human being can acquire in years of training and study. What one computer learns in a microsecond can be shared with and acquired by ten thousand computers in a few milliseconds. One of the many advantages of these computer systems is that they do not get irrationally invested in untenable theories, such as the value of bloodletting to relieve a fever. Simply put, they do not get stuck on stupid.
The main disadvantage of such systems is that they, like human physicians, can be corrupted by fraudulent information. One of the original developers of artificial intelligence likened this vulnerability to counterfeiting. If an individual, group, or nation wants to destabilize a currency, they can flood the market with bogus money. The more authentic the counterfeit bills appear to be, the more dangerous their impact on the target economy. Similarly, the more authentic bogus information appears to be, the more it undermines the reliability of the information or conclusions coming from the system.
Misinformation has been used over the centuries to justify wars, starve populations, enrich individuals, commit genocide, and undermine democracies. Medical information systems are no less vulnerable to manipulation by misinformation, but the speed of information dispersal and the impact of that misinformation is considerably greater than any we have experienced in the past.
One might consider a ‘general manager’ more important than ever in maintaining the integrity of our medical systems. Unfortunately, even an individual familiar with the rational practice of medicine can be blinded to malicious interventions. For example, settings on a ventilator can be revised to ensure the death of a patient without the equipment itself exhibiting any anomalies. Drug delivery systems can be compromised to provide lethal, rather than therapeutic, doses of medicine. There is a reasonable fear that the artificial intelligence machines might themselves initiate this assault on our healthcare system, but it is more likely to originate with malicious human operators. Hate groups can use information from genetic testing results to identify individuals with specific traits commonly seen in a group they decide to exterminate and intervene in medical treatments of those individuals to launch a ‘shadow’ genocide.
Consequently, we face an evolving dilemma: systems that take us beyond what the G.P. of the 1950s could have imagined are vulnerable to corruption or malicious manipulation. There have always been some evil or simply bad doctors, but the unavoidable adoption of artificial intelligence in medical care is fraught with hazards that we lack the mechanisms to recognize, let alone address.
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.