A new review article in the open access journal Laws argues that medical schools are not teaching enough law, and that the gap is making patients less safe. The authors, a group of medical and legal academics at the University of Medicine and Pharmacy of Craiova, propose building legal literacy into medical training from the first undergraduate year through to continuing professional development. Their conclusion is that integrating law into medical education is “no longer optional but essential.”
The framing is useful. The evidence base is not Ontario’s. The article is written from an explicitly European regulatory perspective, built on European Union law, Council of Europe instruments, the Oviedo Convention and the GDPR, none of which has any force here. Its litigation-risk evidence is largely American, drawn from a body of malpractice research describing a system with different claim rates, different funding and different insurers. Nothing in the article is authority in Ontario.
That does not make it uninteresting. It makes the question a comparative one. Strip out the European regulatory scaffolding and the American claims data, and what remains is a claim about causation: that medico-legal risk arises less from clinical error than from a widening gap between what the law expects of physicians and what physicians have been taught about it. On the Ontario evidence, that claim is partly right, and it is right in narrower places than the authors suggest.
The Argument
The authors work through three domains where they say the gap shows up. In medically assisted reproduction, risk comes from regulatory fragmentation: patients travel between jurisdictions with incompatible rules, and clinicians pick up the continuity-of-care problem afterward. In digital health, risk comes from an asymmetry between responsibility and control, because physicians remain accountable for decisions made on data flowing through systems they do not operate. In artificial intelligence, risk comes from partial delegation of clinical reasoning to systems whose reasoning is opaque, without any corresponding shift in liability.
Their proposed answer is structural. Legal content should be integrated horizontally across the undergraduate curriculum rather than isolated in a single course, and vertically across residency and continuing professional development, so that legal competence grows alongside clinical autonomy. It is a sensible framework, and the authors are candid that it is conceptual and has not been tested empirically.
Ontario Is Not Starting from Zero
The framework describes something Ontario already has, at least on paper.
CanMEDS, the competency framework the Royal College uses for postgraduate training across the country, already treats legal obligation as a named competency rather than an afterthought. Under the Professional role, physicians are expected to adhere to the professional and ethical codes, standards of practice and laws governing practice. Under the Communicator role, documentation is framed expressly as a regulatory and legal requirement rather than a clinical habit, and consent is tied directly to the documentation obligation. The Medical Council of Canada examines a domain covering the legal, ethical and organizational aspects of medicine. Canadian medical school accreditation runs through curricular standards on ethics, communication and professionalism, which is where legal content lives.
Then there is the College of Physicians and Surgeons of Ontario, which does something European regulators largely do not. It publishes detailed, operational policy on the precise points where law meets the clinical encounter: consent to treatment, medical records documentation, disclosure of harm, and protecting personal health information. These are not academic statements of principle. They tell a physician when express consent must be obtained and documented, what a record must contain, and what must be disclosed after a harmful incident.
The horizontal and vertical integration the article recommends is, in Ontario, largely built. If a preparedness gap exists here, it is a delivery and assessment problem, not an absence of architecture.
The CMPA Changes the Picture
The larger omission is institutional. The article’s European frame has no equivalent of the Canadian Medical Protective Association, and the CMPA is the single most important reason the diagnosis does not transfer cleanly.
The CMPA is a mutual defence organization rather than an insurer, and it covers the overwhelming majority of physicians practising in Canada, reporting more than 117,000 members. It is also, in practice, the country’s largest continuing medico-legal educator: a good practices guide, a medico-legal handbook, thousands of advice articles, and a telephone line physicians can call before they act rather than after they are sued.
The scale of that advisory function is the part worth noticing. In 2024 the CMPA reported assisting with more than 30,000 new matters, including 958 civil legal matters, 5,310 College matters and 2,275 hospital matters, and engaging with members more than 54,000 times. The CMPA is careful to say its case data should not be read as a measure of patient safety incidents generally, and that caution is well taken. But the shape of the numbers is instructive on its own. College matters outnumber civil legal matters by roughly five to one, which is a useful corrective to any account of medico-legal risk that treats litigation as the main channel. It is not the main channel, in Ontario or anywhere else in Canada. Whether the underlying volume is rising at all is a separate question, and one I have looked at in more detail in Is Medical Malpractice on the Rise in Canada?
None of this means Canadian physicians are uniformly well informed about the law. It means the mechanism the article proposes to build already exists here in a different form, delivered continuously and at the point of decision rather than in a lecture hall a decade earlier.
Where Law and Clinical Care Actually Meet
There is a real overlap zone, and it is narrower than “legal literacy.” It is consent, records and disclosure. These are the three places where a legal rule governs a clinical act directly, and they are, not coincidentally, where a great many Ontario claims are decided.
Consent is the clearest example. The disclosure obligation here is not a matter of professional courtesy. It comes from Reibl v Hughes, [1980] 2 SCR 880 and Hopp v Lepp, [1980] 2 SCR 192, and it is worked out against the modified objective test for causation, which asks what a reasonable patient in this patient’s position would have done if properly informed. The elements of a valid consent are set out in the Health Care Consent Act, 1996. And the duty keeps expanding at the edges. In Denman v Radovanovic, 2024 ONCA 276, the Court of Appeal addressed the treatment team’s duty to inform where treatment unfolds across multiple steps and multiple physicians, and held that the obligation to convey cumulative risk is not discharged by the operating physician alone.
Records are the second. A chart that does not record the risk discussion is not merely an administrative failing, it is an evidentiary one, and at trial the absence of a note is frequently the whole fight. Privacy is the third. The Personal Health Information Protection Act, 2004 governs who may look at a health record, and unauthorized access to a patient’s chart is prosecuted and disciplined rather than merely criticized.
On these three, the article’s thesis is basically correct. A physician who does not know what the consent obligation actually requires is at real risk of failing it, because here the legal rule and the clinical act are the same act.
Artificial Intelligence Is Where the Gap Is Real
The article’s strongest section is on artificial intelligence, and it is strongest because Ontario’s position right now is genuinely unsettled.
There is no comprehensive federal AI statute. The Artificial Intelligence and Data Act died with Bill C-27 when Parliament was prorogued in January 2025 and has not returned. Health Canada regulates machine learning enabled medical devices through device guidance. The CPSO has issued advice to the profession on using artificial intelligence in clinical practice, but advice is not policy, and its central message is that while no specific law addresses AI, “the core expectations of physicians remain unchanged.” The Information and Privacy Commissioner has published guidance on AI scribes in the health sector. The CMPA advises members to obtain and document consent before recording a clinical encounter and to review every AI generated note, on the footing that the physician remains responsible for the chart either way.
That is a great deal of guidance and no law. Which means the standard of care for AI assisted practice in Ontario will be set the way standards of care are always set here: retrospectively, case by case, on expert evidence, years after the conduct in question. Physicians adopting these tools today are working against a benchmark that does not yet exist and will be constructed after the fact. That is a real preparedness gap, and it is the place where the article’s diagnosis lands squarely.
It is also worth noting that the CMPA has itself pushed back on the idea that the burden of managing AI risk should fall entirely on individual physicians. That is a fair point, and it cuts against the article’s framing as much as for it.
What the Argument Gets Wrong
Three things.
First, and most importantly, negligence is not decided by what a physician knows about law. In Ontario the standard of care in a medical negligence action is established through expert evidence about what a reasonable physician in the same circumstances would have done, and absent narrow exceptions a claim without that evidence fails. I have written about the expert evidence requirement and how unforgiving it is. A legally sophisticated physician who departs from the clinical standard is liable. A physician who has never read the Health Care Consent Act but obtains a proper consent and meets the standard is not. Legal literacy is neither a defence nor, in its absence, a breach. Outside the consent, records and disclosure overlap, the connection the article draws between legal knowledge and liability is weaker than it asserts.
Second, teaching physicians to think about legal risk carries its own cost. The defensive medicine literature is largely American and should be treated cautiously here, but the mechanism is not exotic. A clinician primed to think about liability orders the extra test, writes the longer note and hedges the conversation. That is not obviously better care, and it is not obviously safer.
Third, framing medico-legal risk as an individual preparedness deficit shifts attention away from where a good deal of it actually sits. Understaffing, handover failures, overcrowded emergency departments, poorly procured records systems and communication failures between physicians are system problems, and they generate a substantial share of the harm I see. Training an individual physician to recognize a legal risk does not fix an institution that has made the safe course difficult.
What This Means for Patients
Very little directly, and that is worth saying plainly.
If you have been injured by medical care in Ontario, nothing about how your physician was educated changes what has to be proven. You still have to establish, on expert evidence, that the care fell below the standard, and that the failure caused the injury. Whether the physician took a health law seminar in second year is not a question a court will ask.
Where the argument does touch patients is indirect but real. Consent that is properly obtained and properly documented produces better records, and better records produce clearer answers about what actually happened. When a family comes to me and the chart contains a genuine note of the risk discussion, the case resolves faster in either direction. It either has merit or it does not, and we find out sooner. That is a modest benefit. It is not nothing.
The authors are right that law is not an external constraint on medicine but a feature of ordinary clinical decision making. They are right that artificial intelligence is arriving faster than the rules governing it. Where I would part company is on the remedy. In Ontario the difficulty is less that physicians have not been taught the law, and more that the law, in the areas moving fastest, has not yet been written.
Larisa Pătru, Oana Andreea Diaconu, Maria Cristina Bezna, Gabriela Boldeanu, Ciprian-Laurențiu Pătru, Adrian Bogdan and Elena Cristina Andrei, “Integrating Legal Education into Medical Training: A Conceptual Framework for Reducing Medico-Legal Risks in Healthcare” (2026) 15:4 Laws 81.
