MCQ banks are the right tool for one part of RACP Adult Medicine and the wrong tool for the rest. They build and measure the written-knowledge base you need for the Divisional Written Examination, but they cannot assess the skills the Divisional Clinical Examination is designed to test: long-case and short-case performance, fluent Australian practice, and oral reasoning under an examiner's eye. This article names those skills precisely, maps them to the official format, and sets out how to train each one as a distinct competency — with iatroX positioned only as the written-knowledge and unseen-MCQ layer, not as a substitute for supervised case practice.
The direct answer: what MCQ practice cannot assess
A question bank tests whether you can select the best answer from five options with the stem laid out in front of you. It cannot tell you whether you can take a history and examine a real patient to time, synthesise a problem list out loud, defend a management plan to a sceptical examiner, or apply Australian-specific guidance, subsidised prescribing and local epidemiology rather than the practice you happened to train in. Those are performances, not recognitions — and the RACP clinical exam is built to observe performances. No amount of MCQ accuracy substitutes for them, which is why a high written score and a failed clinical exam so often coexist.
Official format map
The RACP basic-training assessment has two separate hurdles, and they test different things:
- Divisional Written Examination (Adult Medicine): two papers of single-best-answer MCQs sat on one day, covering Clinical Applications and Medical Sciences, computer-based at approved test centres. It totals roughly 170 questions across the day (widely reported as about 100 in one paper and 70 in the other) — verify the current question count and duration on racp.edu.au, as these are periodically reviewed. This is the knowledge hurdle, and it is where a bank legitimately lives.
- Divisional Clinical Examination: a separate performance exam comprising a long case (extended patient work-up followed by presentation and questioning) and short cases (focused, observed examinations against a rubric). Confirm the current number of long and short cases and their timing on racp.edu.au. This is the hurdle MCQs cannot touch. In the long case you are given real, undifferentiated clinical time with a patient and must then present a synthesised formulation and defend it under questioning; in the short cases you are watched examining focused systems and interpreting signs in a few minutes. Both are marked by examiners against published criteria, and both reward a fluency that only repetition on varied patients builds.
The written exam gates progression, but passing it proves knowledge, not clinical performance. Treat the two as separate projects with separate training methods, because time poured into the one you have already passed is time stolen from the one that will actually decide the outcome.
Separate knowledge from performance
It is worth being precise about what a correct MCQ answer proves. It proves that, given a structured stem and five options, you can recognise the best one — a real and necessary skill. It does not prove that you can elicit the findings that make up the stem in the first place, that you can prioritise among competing problems in a live patient, that you can express your reasoning in a way an examiner can follow, or that your plan reflects current Australian practice rather than a generic or overseas default. The clinical exam removes the scaffolding — no options, no tidy stem, a real patient and a clock — and measures exactly the parts the MCQ removed. Recognising this gap is the first step to closing it.
The under-tested skills, broken down
Long-case and short-case performance
- Observable behaviour: a complete, time-bounded history and examination; a synthesised problem list; a prioritised, justified management plan; and a clear, structured presentation that withstands questioning.
- Deliberate-practice task: repeated timed work-ups on real, varied patients, each followed by a spoken presentation to a supervisor who interrupts and probes.
- Feedback source: a consultant or trained examiner using the official RACP rubric — not self-assessment and not a generic score.
- Exit standard: you consistently complete the work-up to time and defend your plan against challenge, matched to the level the RACP rubric describes.
Australian practice
- Observable behaviour: management that reflects Australian therapeutic guidance, subsidised-prescribing realities and local disease epidemiology — for example the use of Australian Therapeutic Guidelines (eTG) and the Australian Medicines Handbook for medicines decisions, Pharmaceutical Benefits Scheme (PBS) considerations in prescribing, and awareness of conditions with distinct local relevance such as rheumatic heart disease in Aboriginal and Torres Strait Islander populations or region-specific infections.
- Deliberate-practice task: deliberately re-derive your management for common presentations using Australian sources, and flag every point where your default differs from local guidance.
- Feedback source: an Australian-trained clinician who can catch the subtle localisation errors an overseas-trained candidate will not notice.
- Exit standard: your plans default to current Australian guidance without prompting, and you can name the local source.
Oral reasoning
- Observable behaviour: thinking aloud in a structured, prioritised way — stating a differential, the discriminating features, the immediate actions and the disposition — clearly enough for an examiner to follow and challenge.
- Deliberate-practice task: rehearse spoken "differential → discriminator → plan → contingency" runs on unseen vignettes, out loud, against a timer, with a partner asking "why?" at each step.
- Feedback source: a supervisor who scores structure and defensibility, not just the final answer.
- Exit standard: you can reason aloud under interruption without losing structure, on cases you have not seen before.
A four-week modality ladder
Skills like these improve through graded exposure, not repetition of the same comfortable case. A workable ladder:
- Week 1 — isolated skill. Drill each component alone: history-to-time, examination technique, and spoken reasoning on vignettes. Keep the knowledge base warm with unseen MCQs (this is iatroX's job), but do not confuse that with case practice.
- Week 2 — coached case. Full long-case work-ups on real patients with a supervisor interrupting and correcting in real time. Volume of different patients matters more than polish on one.
- Week 3 — timed integrated case. Whole cases to exam time, presented under questioning, scored against the official rubric. Start logging which case types you have and have not done.
- Week 4 — unseen simulation. Full mock long and short cases with patients and examiners you have not rehearsed with, under exam conditions, to expose the gap between prepared and genuinely ready.
When AI feedback helps, when it does not
Automated feedback has a real but bounded role here. It is useful for the written-knowledge layer — generating unseen questions, checking factual recall, and drilling the reasoning skeleton of a case — and for low-stakes rehearsal of structure. It is unreliable for judging the lived quality of a bedside performance: whether your examination was slick, whether you missed a physical sign, whether your presentation landed with an examiner, and whether your plan is genuinely appropriate to this Australian patient. For those, a clinician or examiner is required, scoring against the official rubric. The general principle — calibrate any automated score against a human standard before you trust it — is set out in the iatroX AI-feedback-calibration pillar, and it applies with full force to clinical-exam preparation. Use AI to keep knowledge sharp; do not let it certify your bedside performance.
A balanced case and task matrix
Candidates drift toward the cases they find comfortable and arrive with blind spots. Build a matrix and fill it deliberately so you are not practising only familiar scenarios:
| Case type | Long case done | Short case done | Australian-practice check | Rubric-scored |
|---|---|---|---|---|
| Complex multi-system / multimorbidity | ☐ | ☐ | ☐ | ☐ |
| Cardiology | ☐ | ☐ | ☐ | ☐ |
| Respiratory | ☐ | ☐ | ☐ | ☐ |
| Endocrine / metabolic | ☐ | ☐ | ☐ | ☐ |
| Renal | ☐ | ☐ | ☐ | ☐ |
| Neurology | ☐ | ☐ | ☐ | ☐ |
| Gastroenterology / hepatology | ☐ | ☐ | ☐ | ☐ |
| Rheumatology / immunology | ☐ | ☐ | ☐ | ☐ |
| Haematology / oncology | ☐ | ☐ | ☐ | ☐ |
| Infectious disease (incl. locally relevant) | ☐ | ☐ | ☐ | ☐ |
An empty cell is a case type you have not performed — not one you have merely read about.
Red flags that you are training recognition, not performance
Watch for these signs that your preparation is drifting back toward the comfortable:
- Memorised scripts — reciting a rehearsed spiel that collapses the moment an examiner deviates from it.
- Repeated cases — practising the same three willing patients until you are polished on them and untested on everything else.
- Generic feedback — "good, keep going" that is not anchored to the official rubric and changes nothing.
- Uncalibrated scoring — self-assessed marks, or AI scores, never checked against a real examiner.
- No official-rubric check — practising against your own idea of the standard rather than the RACP's published one.
If several of these are true, your written score may be climbing while your clinical readiness stands still.
A worked example: reading your two dashboards
Take Dr B, sixteen weeks out. Their written dashboard is reassuring: roughly 80% across a large adult-medicine bank, with Clinical Applications and Medical Sciences both above their floors on unseen, timed blocks. On that evidence alone, Dr B feels on track. The clinical dashboard tells a different and more important story. The case matrix shows long cases performed only in cardiology, respiratory and one endocrine patient — the same three cooperative ward patients, repeatedly — with renal, neurology, haematology and complex-multimorbidity cells empty. No performance has been scored against the official RACP rubric; feedback has been warm but generic. Australian-practice checks have not been done at all, because the bank Dr B used was written for an overseas market.
Read side by side, the two dashboards give a clear instruction. The written project is essentially finished; pouring more hours into MCQs would lift a number that is already high enough and change nothing that matters. The clinical project has barely started. Dr B's most valuable next sixteen weeks are dominated by varied real patients, examiner-scored presentations and a systematic pass through Australian sources — not by more questions. The lesson generalises: when your knowledge dashboard is green and your performance dashboard is empty, the honest reading is that you are half-prepared, and the missing half is the one the exam is designed to fail people on.
Bottom line
MCQ banks, iatroX included, do one job well for RACP Adult Medicine: they build and measure the written-knowledge base, and a strong bank is worth having in the stack for exactly that. What they cannot do — assess whether you can perform a timed long case, examine to a rubric in the short cases, reason aloud under challenge, and manage a patient in line with current Australian practice — is precisely what the Divisional Clinical Examination exists to test. The candidates who come unstuck are usually not short of facts; they are short of performed, examiner-scored, Australian-contextualised case repetitions. Use the bank to keep knowledge sharp and honestly measured, then spend the bulk of your remaining time on the modality ladder, the case matrix and rubric-based feedback from clinicians. That is the division of labour this whole article is arguing for.
Frequently asked questions
How do I know whether I have covered the full RACP Adult Medicine blueprint? You have covered the written blueprint when every Clinical Applications and Medical Sciences area from the RACP curriculum has recorded questions attempted, unseen first-attempt accuracy and a recent review date, with no area below your floors — coverage is a property of the official RACP curriculum, not of any bank's contents, so map onto racp.edu.au. But "covered" for the exam as a whole also requires evidence from the clinical side: a filled case matrix and rubric-scored performances. A complete written audit with an empty case matrix means you have covered one hurdle and not the other.
Can one question bank be enough for RACP Adult Medicine? One bank can be enough for the written-knowledge hurdle if its coverage maps onto the RACP curriculum and you keep a protected pool of unseen items for honest measurement, and most candidates use a primary bank plus a second unseen source under the two-Q-bank rule for exactly that reason. No bank, however, is enough for the exam as a whole, because none trains or assesses the long case, the short cases, Australian-practice localisation or out-loud reasoning — those require real patients, Australian-trained supervisors and rubric-based feedback.
What should I measure instead of my overall Q-bank percentage for RACP Adult Medicine? On the written side, measure per-area first-attempt accuracy on unseen, timed, mixed blocks, your high-confidence error rate, retention on delayed re-testing, and pacing — not the blended headline figure, which hides your weakest area behind your strongest. On the clinical side, measure something a percentage cannot capture at all: how many distinct case types you have performed to time, and how your rubric-scored performances trend under examiner challenge. The overall percentage is the wrong readiness signal for both hurdles, as the iatroX percentage article explains.
When should I stop doing new RACP Adult Medicine questions? Stop adding new written questions when every curriculum area is above your coverage and accuracy floors, retention is holding and high-confidence errors are falling — at that point additional MCQ volume yields little, and your marginal hour is far better spent on coached cases, Australian-practice localisation and spoken reasoning. Do not stop because you finished a bank or because a percentage looks reassuring, and do not let a comfortable written score delay the clinical training that the exam actually turns on.
Which RACP Adult Medicine resource should I use for my weakest component? Match the resource to the deficit. If the weak component is written knowledge in a specific area, use a targeted bank plus the relevant Australian guideline, and let iatroX supply unseen questions to re-measure the fix. If it is long-case or short-case performance, the resource is supervised practice on varied real patients with rubric-based feedback — no software substitutes for it. If it is Australian-practice localisation, work systematically through Australian sources with an Australian-trained clinician; if it is oral reasoning, rehearse structured think-aloud runs with a partner who interrupts and scores you.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. Exam-format details are summarised from the Royal Australasian College of Physicians; the Divisional Written Examination question count and the number and timing of clinical cases are periodically reviewed and not fully published on a single accessible page, so verify the current structure on racp.edu.au before planning around any number. Disclosure: iatroX operates a competing question-bank and clinical-knowledge platform; this article confines iatroX to the written-knowledge and unseen-MCQ measurement layer and states plainly that it does not reproduce or replace the RACP Divisional Clinical Examination, supervised case practice or examiner assessment. Automated feedback should be calibrated against a human examiner before it is trusted; no proprietary-algorithm claims are made. Corrections are welcome via the feedback route on iatrox.com.
References: Royal Australasian College of Physicians — Divisional Written Examination and Divisional Clinical Examination pages and Adult Medicine curriculum (racp.edu.au); iatroX Australia exam hub (https://www.iatrox.com/australia); iatroX comparison hub (https://www.iatrox.com/compare); "Your Q-Bank Percentage Is Not Your Exam Score" (https://www.iatrox.com/blog/qbank-percentage-not-your-exam-score); "Calibrating AI-graded SAQs and OSCEs" (https://www.iatrox.com/blog/ai-graded-saqs-and-osces-how-to-calibrate-automated-feedback-before-you-trust-the-score).