Almost no country lets a doctor trained somewhere else simply start practising. Usually they must register with the local authority, and they may have to sit further exams or work under supervision until the system is satisfied they understand how medicine is done here, not medicine in the abstract. Even where doctors move most freely, as within the EU, qualifications are recognised through a formal process rather than simply accepted. We accept this without much argument, because graduating medical school in one country does not automatically make you safe to practise in another.
It is strange, then, how readily we assume the opposite of the tools we learn from. A study app built for another health system is treated as though it travels for free, as if clinical knowledge were the same everywhere and only the accent changed. It is not. The same instinct that makes us check a doctor at the border should make us look twice at the software we hand a student to prepare with.
Why do we check the doctor but not the tool?
We should, and for the same reason. A medical education tool is making clinical claims on every screen: which drug to reach for, which investigation is reasonable, what a good consultation looks like, what counts as a safe plan. Those claims are not universal. They are answers a particular health system has already settled in its own way, and a tool built inside another system carries that system's answers whether it intends to or not.
Used well, an overseas tool is still genuinely useful for the parts of medicine that do travel. Anatomy and physiology barely change from one country to the next, and a clear explanation of how the kidney handles sodium is a good explanation anywhere.
Clinical reasoning is where that stops being true. How a clinician weighs risk, what they choose to investigate, what they prescribe, and how they talk to a patient are all shaped by the system they trained in. The difficulty for a student is that the interface gives no signal about which is which: advice that travels and advice that quietly does not look identical on the screen. The place the mismatch surfaces is the exam they have to pass, or the patient in front of them.
What actually has to change between countries?
What is available to prescribe, the way risk is weighed, how clinicians and patients communicate, and the standard a student is finally held to all shift from one country to the next. Some of it is obvious. Units of measurement differ, and the medicines on the shelf differ with them. A drug that is first-line in one country may be unregistered, unfunded, or simply unavailable in another, so a regimen that is routine in one place is one a student elsewhere should never reach for. Practising against the wrong list is worse than unhelpful; it is misleading.
Beneath the surface, the shape of a health system changes the reasoning itself. Where specialists are easy to reach and the threat of litigation is ever-present, as in the United States, medicine tends to be taught in a more defensive, investigate-everything style. Where the cost and risk of care fall differently, the tolerance for uncertainty shifts with them. To take one concrete example, New Zealand's no-fault accident compensation scheme removes much of the litigation pressure that shapes practice elsewhere, which can support a higher tolerance for clinical risk. None of these stances is simply right or wrong, but each produces genuinely different clinical reasoning, and a tool built inside one system teaches its habits by default. How a system is funded and resourced pulls the same way: which medicines are available and paid for, and which investigations are within reach, decide what is reasonable to do, and therefore what is taught.
Communication is local too. How clinicians and patients speak with each other, what counts as respectful, and what is expected of the clinician all vary between countries, and between cultures within a single country. This is not simply etiquette. In many systems, culturally safe practice is a professional and regulatory requirement, and students are assessed against it. A tool built on the other side of the world has no reason to encode any of that, which is precisely why genuine localisation has to treat it as part of the clinical content, not a finishing touch applied at the end.
Finally, the standard a student is held to is itself local. What their own system expects them to demonstrate, and how it judges that, is defined in that country on its own terms, and knowledge shaped around another system's expectations will not line up neatly with it.
So does the tool have to be built locally?
Not necessarily, but it does have to have that country's knowledge and clinical understanding built into it: the medicines, the guidelines, the pathways, the norms. The real question has always been how you do that, and for the large international tools it has sat in the too-hard, too-expensive basket. The traditional way to build clinical content, commissioning large teams of experts to write and maintain it market by market, only pays back in the largest markets, and with economies of scale pulling the other way there was never a commercial case for them to do the deep local work. So a smaller health system sat at the back of the queue, and mostly stayed there.
What changes the maths is the form the knowledge is built in. Translation was never localisation, and relabelling drug names over content built for another system was never enough, because what has to change is the clinical knowledge underneath. Done by hand, country by country, that was the bottleneck. Built once in a structured, reusable form that can be checked, adapted, and extended for each new system rather than rewritten from scratch, local knowledge stops being a cost only the biggest markets can justify. That is what turns building it in properly for a smaller country from a commercial non-starter into a viable opportunity.
What does this mean for smaller health systems?
More than it first appears, because the logic only runs one way. A tool built so that it can be localised down to the smallest system will also work in the largest, because no large market is really a single market. The United States is often treated as one, but in practice it behaves more like fifty: licensing, scope of practice, and the standards a student is examined against vary from state to state, and what is funded and available varies again by payer and setting. Anyone who has done serious market segmentation knows this. Something that can be brought down to the level of a single small system can be assembled back up into a large one, but the reverse does not hold: you cannot take a tool built to one dominant template and trust it to fit a small place it was never shaped for.
Paradoxically, building for a smaller market rather than the largest may actually prove the more durable choice over the long run. There has always been a pull between global reach and local fit in education, and as teaching and professional training lean more heavily on software, the tools that last will be the ones that can be localised. The most immediate benefit falls to smaller markets that have never had access to this kind of localised content before, but larger markets gain in time too, served by tools that can be localised to the many smaller components those markets are really made of.
For any institution choosing a tool, the question is worth asking early: was this built so it can be localised to us, or are we expected to fit an established template? Because once the student becomes the doctor, their patients' needs are no less valid for the size of the system caring for them. The doctors who treat them deserve to learn on that system's own medicine, shaped by the people who practise it.