Artificial Intelligence and Medical Education. Two challenges for the new Code of Conduct

by Antonio Panti

09 JAN –

Dear Director,
last autumn, the FNOMCeO launched the process of renewing the Code of Conduct, after the pandemic and in the face of continuous scientific and technological innovations, with an important conference in which experts in various disciplines discussed the most relevant ethical issues that doctors must face in the profession; a cultural basis for drafting the new text.

In this extensive comparison, some points deserve further study, two in particular, the responsibility of the doctor in the age of Artificial Intelligence and the relationship between ethics and training, that is, how to evaluate the training of doctors on a deontological level.

In the discussion on the problems posed by AI to medicine and healthcare, addressed in the conference in its many aspects, including the liability profiles of the producer and the user, one aspect deserves further attention.

There are already many applications of telemedicine and, within a few years, the profession will be immersed in this technological field with the risk of distorting the “physicality” of the medical visit into virtuality. How does the relationship with the patient survive within a sensory limited contact and little producer of empathy?

ICT cannot fail to influence the methodology with which the doctor has so far arrived at the diagnosis, prognosis and therapy. In the smartphone, the patient’s biological data, his medical history and the narration of his experience will be compared with the results of the literature: the app will propose the solution.

Will it be like following a guideline? No. It’s not the same thing: the guideline provokes a mental confrontation with ample room for freedom, the app offers a written result, the result of the big data entered into the tool, which it is not easy to oppose. It is highly probable that the doctor will rely on it by decreasing the degree of freedom and autonomy of choice.

But the AI learns deep learning starting from the data entered and there are already examples of distorted information that have led to serious errors of assessment: if male hypertensives, usually over 60, are often bald, the app could give advice for baldness as antihypertensive therapy.

Let’s imagine this fantasy in court during a lawsuit for medical negligence, based on the search for objective responsibility, that is, for the “truth” by a magistrate thirsting for certainties. The meeting of law – now unmanageable due to the sum of confused laws and the ever more tenuous concatenation of facts passed off as causality – with medicine, a probabilistic science, means that, in the face of an alleged professional fault, the judgment is aleatory if not random.

AI imposes a drastic rethink on professional liability: moving from the culture of sanction to that of no blame. If damage has occurred during a health procedure, established that there is no responsibility for the injured party, it is simply necessary to compensate him.

In modern medicine, professional error must be dealt with on a training and contractual level, without penal sanctions, without prejudice to any deontological infractions, and there is no shortage of examples in many countries.

However, limiting the concept of responsibility could affect the freedom of choice of doctor. A culture of the heuristic potential of error (training versus sanction) transferred to the contractual level, towards a restorative and non-vengeful justice, would respond to modern juridical demands and would keep professional autonomy unchanged.

The evaluation no blame of medical error and therefore of “professional guilt” is recognized in all texts on clinical risk and in many ministerial documents. It is clear that a law is needed to modify the penal and civil aspects. But the Deontological Code could move in this ambit both on the level of the relationship between error and professional fault and on that of the deontological intertwining between professional error and training.

It is precisely the deontological value of training that deserves further reflection. The current Code of Conduct, having reaffirmed the doctor’s obligation to constantly update and continuously train technical and non-technical knowledge and skills, chooses a precise path: “the Order certifies the credits acquired by each member and evaluates any defaults”.

It was then decided to submit to a public regulatory regime. The Order will evaluate the level of training of the members on the basis of the credits obtained according to the guidelines and standards issued by the National Commission chaired by the Minister. A unique choice in a Code that is based on absolute autonomy with respect to the laws, albeit within the framework of the values of the Constitution.

In this context, two questions cannot be avoided: on the one hand, a reflection on the relationship between continuous training and deontology, on the other, the imminence of the sanctions that the Orders will have to inflict on members who are not in compliance with the credits provided for by the directives ministerial. I think the reasoning to do is broader.

The doctor must adjust his knowledge and improve behaviors and skills. Continuing education has cognitive and professional aspects. The first could be measured with the classic mechanism of the exam, the second evaluate on problem solving ability. In the free profession, the client is the judge and the unprepared doctor will leave the market; if the doctor practices as an employee or under contract, the evaluation is up to the employer who has an interest in providing clients with quality assistance. In this case, the training obligation is contractual and must be evaluated in the forms agreed between the parties.

From a professional point of view, training falls within the vast theme of evaluating results, a theme on which the Code of Conduct has not found a concrete drop point. The CME credit system is more aimed at acquiring notions even if it does not neglect the professional part. The assessment of the training of doctors, which the Orders are preparing for, while it has an evident contractual footing, it is not clear how it can be mechanically linked to the possession of a number of training credits established administratively by a body of the NHS.

What about ethics? One sense has an optative norm, which pushes the doctor to maintain adequate, indeed to improve his own professional performances, another has a dispositive norm which measures training with the meter of credits and on which he bases a disciplinary evaluation. From a purely deontological point of view, it is not so much the mechanics of credits that are of interest as the ability of each member to know how to deal with the problems that arise according to the most modern scientific acquisitions, in a correct human relationship, without wasting resources, following a methodology recognized.

But how to evaluate this competence? And, in any case, what evaluation tools does the Order possess on the basis of the law in force, the 3/18, should one wish to give effect to this reasoning? The connection between training and deontology is not clear except on the exhortative level. Beyond a possible role of the Order, as a subsidiary body of the State, with regard to freelance professionals, a deontological evaluation of the state of professional adequacy of each doctor, deriving from the personal training continuum, would rather recall that institution, never materialized as far as discussed, which is the recertification of the degree.

Could such a task be entrusted to the Order in the Italian legal system? I really do not think so. The word “order” recalls a world of rules in which the absolute professional autonomy that the “Orders” claim is difficult to place. What will happen in a different political climate? Despite this reasoning, on which it would be good to reflect, the Orders will follow the indications of the ECM Commission, despite the great diversity of the training offer, I fear in the midst of difficulties and controversies in a framework of weak attachment to deontology.

In conclusion, both responsibility in times of AI and the deontological measure of training represent two points of discussion for the next Code of Conduct and for a vision of the profession more in line with the evolution of the times.

Antonio Panti

PS On 31 January, with the milleproroghe decree, the Government extended the expiry of ECM credits by one year; consequently the date of the order control should be postponed. As often happens, the experimental phase never ends and we find ourselves caught between an inconclusive practice and the interests that have developed in the meantime. I suggest a reflection table.

Other articles in Letters to the editor