Six weeks out is the point at which most candidates start taking the exam seriously, and it is a reasonable run-in: long enough to fix real weaknesses, short enough that nothing can be wasted. What sinks people is not the length of the plan but its shape. Most six-week plans are just "do lots of questions", accelerating towards a panicked final fortnight. A plan that works has phases, each with a different job, and gates between them. Here is the structure, driven by your own data rather than by a generic timetable, in a full-time version and a version for people doing this alongside clinical work.
Key takeaways
- Week one is measurement, not learning: build an honest map before you spend a single hour correcting anything.
- Weeks two and three are targeted correction, guided by the map rather than by what you feel like revising.
- Weeks four and five are mixed, timed, unfiltered practice, which is a different skill from topic revision.
- Week six is rehearsal and consolidation, and you should not be learning new material in it.
- Gate each phase: do not move on until the previous phase has produced what it was supposed to.
Week one: build the map
Resist the urge to start correcting things. You cannot target what you have not measured, and a week spent measuring pays for itself several times over.
Do three things.
Sit a representative diagnostic. Unseen, timed, unfiltered across the whole blueprint, at the real pace of the real paper. It will be uncomfortable and the number will be lower than your dashboard says. That number is the only honest baseline you have.
Run a coverage audit. For every domain in the official blueprint, record how many questions you have actually attempted alongside your accuracy. Domains with low exposure are unknowns rather than strengths, and the unknowns are where exam-day surprises come from. The method is set out in using Standard mode as a syllabus audit.
Analyse your errors by operation, not just by topic. Sort your recent mistakes into knowledge gaps, prioritisation failures, misreadings and timing casualties. If they cluster in one operation, you have found a cross-cutting weakness that a topic-based dashboard would never show you.
The output of week one is a single page: your weak domains, your unsampled domains, your weak operations, and your honest baseline. Everything after this is driven by that page.
Weeks two and three: targeted correction
Now spend your hours where the map says, not where your instincts say.
Attack the unknowns first, because they are the only category you cannot reason about at all, and because a domain you have never opened may be hiding either a strength or a catastrophe and you have no way of knowing which.
Then attack the known weaknesses, weighted by their share of the blueprint. A weakness in a domain worth thirty marks matters more than one in a domain worth four, and most candidates ignore the weighting entirely.
Review properly rather than passively. For every error, extract the rule, in one line, in your own words, and specifically identify the discriminator you missed. Reading explanations feels productive and is the commonest reason candidates work hard and stop improving.
And track your confidence as you go, because the questions you get wrong while feeling certain are the most dangerous items in your dataset and you will otherwise never revisit them.
The gate: do not move to phase three until every blueprint domain has cleared your exposure floor and your top three weaknesses have measurably moved on unseen questions.
Weeks four and five: mixed, timed, unfiltered
This is the phase candidates most often skip, and it is where the exam is won.
Topic-filtered practice tells you what you know when you have been told what to think about. The exam does not tell you. So switch decisively to mixed blocks, at the real pace, with no explanations until the block is finished and no filtering by subject.
Sit at least two full-length mocks under genuine exam conditions, including the breaks if your exam has them, and including the full day if your exam runs to two papers. This is uncomfortable and it consumes whole days, and it is the single most informative thing in the whole plan.
Then interrogate each mock twice. Once for content: which questions did you get wrong and why. And once for behaviour: did your accuracy fall by position within the paper, did you run short of time, did you leave anything blank, did you change answers and were the changes right.
Meanwhile, keep spaced repetition running underneath, returning the concepts you corrected in weeks two and three. A weakness fixed in week two and never revisited is dormant, not fixed.
The gate: your first-attempt accuracy on unseen questions in timed mixed blocks should now be at or above your target, and it is that figure, not your overall dashboard percentage, that tells you where you stand.
Week six: rehearse, consolidate, and stop
The final week has one job and it is not learning.
Sit your preserved official material, cold, timed, under exam conditions. If you have followed the advice to keep some back rather than exhausting it early, this is the moment it earns its keep, because it is the closest thing to an honest predictor you will ever get.
Consolidate the corrections you have already made. Return to your one-page list of extracted rules. Space the high-confidence errors. Revisit the operational weakness you have been training and check it still holds.
Do the logistics, deliberately and early: the booking, the identification, the route to the centre, the timings. Losing marks to a stressful morning is an avoidable and infuriating way to fail.
And taper. Do not attempt to learn new material in the last few days: it will not stick, and it will displace the sleep that will do far more for your performance than another fifty questions. Candidates who cram the final week reliably arrive depleted and perform below what they had earned.
The working-doctor version
If you are doing this around clinical shifts, the phases do not change but the arithmetic does, and the plan must be built for the life you actually have.
Halve your assumed daily capacity, and build the plan around the hours you will genuinely have on a typical day rather than your best day. An honest plan you complete beats an ambitious one you abandon in week two.
Protect the mocks ruthlessly. They are the least compressible part of the plan and the first thing to be sacrificed to a long day, so put them on your genuinely free days and defend them.
Use short sessions for spaced repetition and long sessions for anything that needs sustained reasoning, and match the task to your state: adaptive drilling after a night shift is fine, and the Socratic unpicking of a misconception is not.
Build in a catch-up day every week. Clinical work will take some of your study days. A plan with no slack breaks on first contact with reality, and a broken plan tends to become an abandoned one.
What not to do
Three things sink six-week plans with depressing regularity.
Starting a second pass of the question bank, which measures recognition rather than knowledge and produces a rising number that means nothing.
Leaving mocks until the final week, by which point their diagnostic value cannot be acted upon.
And postponing the small, learnable components, the statistics, the organisational content, the item style you find tedious, until there is no time left. Those components are finite and masterable in a way that clinical medicine is not, which makes them the best return on your hours and the ones candidates most reliably abandon.
Where iatroX fits
iatroX supports the whole arc: Standard mode for the week-one coverage audit, an adaptive engine that targets the weaknesses the audit exposes rather than the topics you enjoy, spaced repetition that keeps your week-two corrections alive into week six, and timed mocks at the real format of your exam so that the pacing is rehearsed rather than discovered. First-attempt accuracy on unseen questions is tracked separately from repeats, which is the metric the gates above actually depend on, and missed questions can be opened in the Socratic Tutor, which asks you to reason before it explains. Try it with free sample questions at iatroX. For what to do if the plan stalls, see the question bank plateau.
Frequently asked questions
Is six weeks long enough to prepare for a medical exam? For most postgraduate written exams, yes, provided the time is structured. What fails is not the length but the shape: a plan with a measurement phase, a correction phase, a timed mixed-practice phase and a rehearsal phase does far more than six weeks of undifferentiated question grinding.
What should I do in the first week? Measure, do not correct. Sit an unseen timed diagnostic, audit your coverage across the whole blueprint, and sort your errors by reasoning operation rather than by topic. The output is a one-page map that drives everything else.
When should I start timed mock exams? From week four, and sit at least two full-length mocks under genuine exam conditions. Leaving them until the final week wastes their diagnostic value, because you will have no time left to act on what they reveal.
Should I revise new material in the final week? No. Consolidate what you have already corrected, sit preserved official material as a benchmark, sort the logistics, and taper. New material learned in the last few days rarely sticks and displaces the sleep that would have helped more.