Ethics and their influence on the use of intelligent clinical decision support systems
Many are familiar with the name of Hippocrates and his association with the field of Medicine. Hippocrates was a Greek physician born on the island of Kos around the year 460 BC during Greece’s “Classical Period.” His fourth century B.C. contemporaries included Plato and Aristotle. He is considered by many throughout history as the founder of the practice of medicine as we understand it today.
The “Hippocratic Corpus” refers to a large volume of writings, in part credited to Hippocrates. The Corpus is an amalgam of many contributors over the course of several centuries and examines various aspects of the practice of medicine. Specifically, the corpus addresses abandoning dogma and superstition and using observation, the study of nature , and an understanding of cause and effect as a means to healing. It counsels physicians to focus on the nature of illness and disease, the science and art of healing, and, importantly, the ethical considerations of the practice of medicine.
Beyond establishing an “empirical” approach to diagnosis and treatment, Hippocrates, to this day, continues to represent the ethical and humane elements of medicine. The “Hippocratic Oath,” one of the treatises in the Hippocratic Corpus, lays out a series of ethical principles and guidelines for physicians to follow as they practice the “art and science” of medicine.
Various modernized versions of the oath are still taken by medical students around the world. At its core, the oath can be reduced to some basic and universal principals that remain relevant over 2500 years after its original writing. Perhaps the most famous and enduring part of the oath is an admonition to physicians to first “do no harm.” Further, the oath goes on to address a number of matters related to the practice of the profession of physician and what we would today call “professional ethics.”
In the most modern and widely used version of the oath, written by Louis Lasagna in 1964, the text identifies several key elements derived from Hippocrates original instructions to the physician. Specifically, to “…respect the hard-won scientific gains of those physicians in whose steps I walk, and gladly share such knowledge as is mine with those who are to follow.” That is, do not ignore the knowledge that has come before and share and further that knowledge for the future generations. Further he goes on to say that physicians should “…apply, for the benefit of the sick, all measures [that] are required.” In other words, exhaust all possible knowledge and skill for the benefit of the patient. He also notes that one should “… not be ashamed to say ‘I know not’, nor should (a physician) fail to call in colleagues when the skills of another are needed for a patient’s recovery.” Simply put, a physician should accept when their knowledge or skills are not enough to help the patient and they should seek help.
The three elements of the modern Hippocratic oath that I have highlighted are of particular interest for physicians and patients today, particularly in the context of the development and application of advanced computing methods such as machine learning (ML) and artificial intelligence (AI) and its direct application to improving the speed, efficiency, and accuracy of diagnosis as well as recommendations as to the most effective therapeutic path for a given patient and diagnosis.
In 1959, Robert S. Ledley and Lee B. Lusted, published a prescient article in the Journal Science; “Reasoning Foundations of Medical Diagnosis – Symbolic Logic, Probability, and Value Theory Aid Our Understanding of How Physicians Reason.” This article attempts to understand and to define and diagram the very nature of human reasoning related to the diagnosis of disease. The authors make three very important and observations. The first was that physicians and scientist understand that there exists an “…increasing interest in the use of electronic computers as an aid to medical diagnostic processes.” The authors conclude that the developing technology of digital computing will make possible “…the ability to make a more precise diagnosis and a more scientific determination of the treatment plan”. Yet they still make clear their concerns related to the interplay between the physician and the computer in stating that “This method in no way implies that a computer can take over the physician’s duties.”
The authors note the value that digital computers will ultimately bring to medicine, they also innately fear the prospects of computing technology taking away the physicians near absolute
and final say on matters of diagnosis and treatment and it seems they grudgingly admit, based on their research, that computing technology will ultimately be of benefit to the patient via better diagnosis and treatment plan decisions. As early as the late 1950s, clinicians saw not only the promise of digital technology and computer aided data analysis in improving patients’ lives beyond what humans alone could do, but they also articulated their reflexive insecurity in terms of a fear of machines “taking over” their decision-making power.
65 years on, we are finally experiencing the promise of computer technology in medicine that Ledly and Lusted predicted. Advances in ML and AI, coupled with developments such as the Internet of Medical Things (IoMT), where connected patients, clinicians, devices, and care environments generate massive amounts of data that feed and drive the development of ever more sophisticated disease modeling and therapeutic decisioning, are creating the very concept of “Digital Health”. The availability of massive amounts of data and nearly unlimited computational power have advanced digital health concepts across the globe. This is certainly reflected in an ever-increasing number of publications related to medicine and computers. Medline returns over 40,000 citations for the search term “Clinical Decision Support” with the earliest paper in published in 1966 (just one citation for that year) and exponential growth, reaching 2,836 citations in 2013 and a stunning 81,000 in 2022.
Today, clinicians are not questioning whether expert decision support systems are a part of modern medicine, rather they are discussing their value and utility. From academic discussions of analytic methods to practical considerations of how such system can provide “health care.” Further they are debating the ethics of complex clinical decision making in the presence and absence of computational aids. With all the progress made since 1959 there still exists some level of mistrust and fear in giving up control of clinical decision making to “machine s.” Yet, there exists a more fundamental question beyond how best to use algorithmic-based clinical decision support, or what is also called ‘software as a medical device.’ Perhaps the bigger question is not whether or how to use intelligent clinical decision support systems, but why we are not using these systems more, and whether NOT using digital technologies to support human decision making fundamentally violates the most basic ethical principles of medicine articulated more than 2600 years ago by Hippocrates.
In 2014, Cary Oberije from the Department of Radiation Oncology at the Mastricht University Medical Center published a paper that compared “predictions of doctors versus models for treatment outcomes in lung cancer patients.” The work clearly demonstrated that “models substantially outperformed Radiation Oncologist’s predictions and guideline -based recommendations currently used in clinical practice.” Dr. Oberije commented on the work she and her colleagues did with the following statement: “If models based on patient, tumor, and treatment characteristics already out-perform the doctors, then it is unethical to make treatment decisions based solely on the doctors’ opinions.”
In 2013, Dr. Robert Koker, in an article published by The Atlantic entitled “The Robot Will See You Now,” was quoted saying, “In Brazil and India, machines are already starting to do primary care, because there’s no labor to do it. They may be better than doctors. Mathematically, they will follow evidence—and they’re much more likely to be right.” A decade later, there still exists a robust debate on the routine integration of such technologies in daily clinical practice. The Medical Internet of Things and Digital Health will be critical in solving the many challenges of healthcare today and in creating many opportunities for improvement in care delivery, access, costs, quality, and personalization. The result will not be whether we should be taking advantages of such technologies; rather, the question will be why are we not using them more.
If the medical community were to add a chapter to the Hippocratic Corpus today and Hippocrates himself could look at the progress that has been made in computer models and algorithmic decision support and the role the Internet of Medical Things plays in the addressing challenges of diagnosis and treatment and how model-based decision making can outperform human decision making alone, I think he would conclude that not taking advantage of such technologies would be profoundly unethical.
Clinicians take an oath to first, do no harm, to respect the hard-won scientific gains of those physicians in whose steps they walk, to gladly share such knowledge with those who are to follow, and to apply, for the benefit of the sick, all measures [that] are required, and not be ashamed to say “I know not”, nor to fail to call colleagues when the skills of another are needed for a patient’s recovery.” Surely, to fail to take advantage of model-based and algorithm-based clinical decision support violates the most fundamental principles of the Hippocratic Oath.
BY: Mark Wolff, M.S., Ph.D.
Global IoT Division , SAS Institute
Visiting Fellow University of Miami, Institute for Data Science and Computing
About SAS in IoT
SAS empowers organizations to create and sustain business value from diverse IoT data and initiatives, whether that data is at the edge, in the cloud, or anywhere in between. Our robust,
scalable, and open edge-to-cloud analytics platform delivers deep expertise in advanced analytics – including AI, machine learning, deep learning, and streaming analytics – to help customers reduce risk and boost business performance. Learn more about our industry and technology solutions at www.sas.com/iotsolutions
