Medical school ends and medical education does not; it changes state. The curriculum dissolves, the lectures stop, the exams thin out, and learning fragments into a dozen informal channels that will carry a doctor's development for the next forty years. Understanding those channels, and the one structural problem they share, is the closest thing there is to a theory of postgraduate learning, and it explains both why experienced doctors know so much and why so much of what every doctor encounters evaporates.
The fragments
Patients: the primary curriculum, forever. Every unusual presentation, every diagnosis revised, every management plan that worked or did not is a lesson delivered at the bedside, unscheduled and unrepeatable. Colleagues: the corridor consultation, the registrar's aside, the consultant's heuristic, medicine's oral tradition, transmitting judgement no guideline captures. Guidelines: the formal layer, updating continuously, encountered mostly at the moment of need. Errors: the most expensive teacher and the most effective, when the learning is actually extracted rather than merely survived. Cases: the deliberate version of patient-learning, case reports, morbidity meetings, the discussed disaster. Courses: the periodic concentrated download, update days, congresses, life support, high in quality, low in frequency. Questions: the daily engine, the dozens of small uncertainties work generates, each answered, most forgotten. Teaching: the secret weapon, since nothing consolidates knowledge like explaining it. CPD: the formalised wrapper obliged by revalidation, at its best the structure for all the above, at its worst a certificate archive. Exams: the intermittent forcing function, brutal and effective, then absent for decades.
The structural problem
Every channel above delivers exposure; almost none delivers retention. The bedside lesson is real and gone by Friday; the guideline consulted is real and re-consulted, identically, in March; the course's brilliant day decays on the standard forgetting curve; the question answered at 3am is asked again in June. This is not a failure of diligence, it is the arithmetic of memory meeting the volume of medicine: unscheduled, unrepeated exposure does not consolidate, however vivid. Medical school solved this with structure, spacing, repetition, assessment, imposed from outside. Postgraduate life removes the structure and keeps the volume, which is why the honest summary of modern medical learning is: the challenge is no longer access to information; it is turning fragmented exposure into retained knowledge.
What the solution has to look like
The fix is not more exposure, doctors are saturated, but the reintroduction of structure around the exposure that already happens. Three mechanisms, borrowed from what made medical school work. Retrieval: being made to produce answers rather than recognise them, which is why questions beat re-reading everywhere they are compared. Spacing: returns scheduled as forgetting approaches, the single most robust finding in learning science and the one hardest to run on willpower. And correction at the level of the misconception: not the right answer restated but the wrong belief found and fixed, which is what turns an error channel into an education. Wrap those three around the daily fragments, the question captured at the bedside becomes a retrieval item next week, the guideline consulted becomes a spaced return, the error becomes a diagnosed and retested correction, and fragmented exposure starts compounding instead of evaporating.
What this looks like in practice
Concretely, a working doctor's learning system in 2026: capture the week's genuine questions as they arise; answer them from cited guidance at the moment of need; let an adaptive engine convert them, and the broader scope of practice, into short retrieval sessions; interrogate the misses until the misconception is named, the Tutor's Socratic job; space the returns automatically; and let the whole loop document itself as professional evidence, since the same records that prove retention also serve appraisal. Twenty to thirty minutes a week, attached to work that was happening anyway. This is, without disguise, the thesis iatroX is built on, but the argument stands independent of any product: the fragments are your curriculum, and structure is the difference between having experiences and having an education.
Frequently asked questions
Isn't experience itself the education, as the older generation says?
Experience is the exposure; the education is what survives. Experienced doctors are excellent partly because decades of repetition provided accidental spacing, the system above simply provides it deliberately, in years rather than decades.
Which fragment is most undervalued?
Teaching. Explaining a topic is retrieval practice at its most demanding, which is why the registrar who teaches the ward learns more than the ward does. Volunteer for it early.
Does this replace courses and colleges?
No; it makes them stick. The update course provides the download, the structure provides the persistence, and the combination is worth more than either doubled: /blog/nb-medical-vs-red-whale-vs-iatrox-gp-education.
How long before a structured loop shows results?
Weeks, measurably: repeat performance on returned topics is visible within a month, and the subjective marker, the same question no longer feeling new, follows shortly after. The compounding, though, is the point; the visible difference at one year is the argument.