An MCQ bank is necessary for the Final FRCA and insufficient for it. Selecting the right option from five proves you can recognise a correct answer; it does not prove you can construct one, defend it aloud, interpret an unfamiliar image under pressure, or criticise the evidence behind it. This article names the exact skills ordinary MCQ practice cannot assess, and gives each an observable behaviour, a deliberate-practice task, a feedback source and an exit standard — plus a four-week modality ladder to build them.
Official format map (last checked 20 July 2026)
The Final FRCA currently has a Written examination — a CRQ paper of 12 constructed-response questions in 3 hours, each marked out of 20 (240 total), all attempted; and a 90-SBA MCQ paper in 3 hours (approximately 45 general anaesthesia, 10 perioperative medicine, 10 regional, 25 other) — and a separate Structured Oral Examination (SOE). Only the MCQ paper is pure selected-response. The CRQ requires you to generate and prioritise written answers, and the SOE requires you to reason aloud and defend under questioning. That is the core of the point: two of the three components test performance an MCQ bank never touches.
Flag — 2027 changes. From July 2027 (subject to GMC approval), the SOE is scheduled to be replaced by a Final Clinical Performance Exam (FCPE) — a circuit of 12 stations — and the written knowledge test by a 100-SBA Applied Knowledge Test, with the current format running until June 2027 and a 2027–2028 transition year. The skills in this article become more central under the new format, not less: a station-based clinical exam raises the premium on judgement, communication and defence. Verify your components against the RCoA 2027 changes hub.
Knowledge versus performance
A correct SBA answer proves a narrow thing: that, given the options, you could identify the best one. It does not prove that you would have generated that option unprompted, that you could justify it against the alternatives aloud, that you could act on it under time pressure with incomplete information, or that you could defend it when an examiner pushes back. The Final FRCA is designed to test the second set — the performance — because that is what safe independent practice requires. Treating a high MCQ percentage as evidence of readiness for the CRQ and SOE is the central error this article exists to prevent; the general version of the argument is in Your Q-Bank Percentage Is Not Your Exam Score.
The four under-tested skills, each made trainable
For each skill, train the observable behaviour, not a vague aspiration.
Clinical judgement (prioritisation under uncertainty). Observable behaviour: you state a prioritised plan for an evolving scenario and justify the order. Deliberate-practice task: work timed CRQ-style prompts and unstructured cases, writing a ranked management plan. Feedback source: a consultant or senior trainee marking against the curriculum's expected points. Exit standard: your prioritisation matches the marking rubric on unfamiliar cases without prompting.
Evidence and critical appraisal. Observable behaviour: you interpret a study, name its principal flaw and quote the numbers that matter. Deliberate-practice task: appraise a paper a week and answer statistics prompts to CRQ time. Feedback source: a rubric or a colleague with critical-appraisal fluency. Exit standard: you can criticise a method and interpret the result cold, without notes.
Imaging and data interpretation. Observable behaviour: you read an unfamiliar image, trace or lab trend and state the finding and its implication under time. Deliberate-practice task: daily short sets of chest and airway imaging, ECGs, echo clips, blood gases and coagulation trends. Feedback source: model reports and senior review. Exit standard: accurate interpretation of unseen data at exam pace.
Structured oral defence. Observable behaviour: you deliver a structured, spoken answer and hold it under challenge. Deliberate-practice task: viva practice aloud, with a listener asking follow-ups. Feedback source: the listener — ideally an examiner or trained consultant. Exit standard: fluent, defensible reasoning to an unfamiliar prompt, unshaken by pushback.
The trap is to train these four in isolation and never integrate them. In the exam — and still more in the 2027 station-based Final Clinical Performance Exam — they arrive together: a case that demands judgement will also demand that you defend it aloud, interpret its data and justify it from the evidence. So once each skill clears its individual exit standard, rehearse them combined, in a single timed case, which is precisely the purpose of the upper rungs of the ladder below. A candidate who can appraise a paper, read a trace and run a viva separately, but cannot do all three inside one evolving scenario, has trained the components and not the exam.
A four-week modality ladder
Do not jump straight to full simulation; build in rungs.
| Week | Rung | Focus |
|---|---|---|
| 1 | Isolated skill | Drill each skill separately — appraisal, imaging sets, single vivas — at low stakes |
| 2 | Coached case | Work integrated cases with a senior coaching in real time |
| 3 | Timed integrated case | Full CRQ answers and vivas to exam time, marked against a rubric |
| 4 | Unseen simulation | Unseen mock papers and vivas under exam conditions, minimal help |
The ladder matters because a candidate who only ever does full mocks never isolates the failing sub-skill, and a candidate who only ever isolates skills never rehearses integration under pressure. Climb it in order.
Worked example: a strong-MCQ candidate on the ladder (illustrative)
Consider a registrar who sits comfortably above their working target on unseen SBA blocks but freezes in viva practice — a common and dangerous profile, because the MCQ score breeds false confidence. In week one they isolate: three short vivas a day on single topics, plus daily imaging sets, marked only for structure and accuracy. In week two a consultant coaches them through integrated cases in real time, interrupting to expose gaps in prioritisation. In week three they run full CRQ answers and vivas to exam time, marked against a rubric, and discover that their written prioritisation lags their spoken reasoning — a gap the MCQ never revealed. In week four they sit unseen mocks under exam conditions with minimal help, and the previously fragile viva now holds under follow-up. Throughout, the MCQ score barely moved; it did not need to, because the readiness gain came entirely from the modalities a bank cannot train. These figures are illustrative and are not a pass prediction — the pattern of where the improvement came from is the point.
When AI feedback helps, and when it does not
Automated feedback has a real but bounded role. It is useful for high-volume, low-stakes reps: checking factual recall, giving instant feedback on a data-interpretation set, or drafting practice prompts. It is unreliable for scoring nuanced clinical judgement and for grading a spoken defence, where a fluent but wrong answer can be scored generously and a terse but correct one harshly. And a clinician or examiner is required whenever the exit standard is calibration to the official rubric — CRQ marking, viva defence, and any judgement of prioritisation. Before you trust any automated score, calibrate it against a human mark on the same work; the method is set out in the iatroX guide to calibrating AI-graded answers. Do not let an uncalibrated AI score stand in for examiner judgement.
The concrete risk is worth naming. Ask an automated grader to mark a CRQ answer and it will reliably tell you whether you mentioned the expected facts; it will far less reliably tell you whether you prioritised them safely, which is what the marks actually reward. A well-structured answer that recommends the wrong first step can read as competent to a model and would lose heavily with an examiner. So use automation to widen the funnel — more reps, faster factual feedback, more prompts generated — and reserve the human for the narrow, decisive judgement of whether your reasoning would survive the exam room.
A balanced case matrix
Left to choose, candidates rehearse the scenarios they already like. Force breadth with a matrix so you practise across the blueprint, not just your comfort zone.
| Domain | Judgement case | Evidence prompt | Imaging/data set | Viva topic |
|---|---|---|---|---|
| Obstetric | ||||
| Paediatric | ||||
| Cardiac/thoracic | ||||
| Neuro/ICM | ||||
| Regional/pain | ||||
| Perioperative |
Fill every cell over your run-in. A blank row is a domain you are avoiding.
Red flags that your performance practice is hollow
- Memorised scripts delivered regardless of the question — fluency without responsiveness fails a viva.
- Repeated cases re-run until smooth — you are rehearsing recall, not judgement.
- Generic feedback ("good, be more structured") that never references the official rubric.
- Uncalibrated scoring — a number from a tool or a friend that has never been checked against an examiner standard.
- No official-rubric check — practising against your own idea of the standard rather than the RCoA's.
Any of these means your simulation is theatre. Fix them before you trust your readiness.
Reading your performance signals
Performance practice generates signals you must read differently from an MCQ percentage. A viva that feels fluent but earns terse feedback is a warning, not a success, because fluency can mask thin content. A CRQ answer that scores well on knowledge but loses marks on prioritisation tells you the gap is judgement, not recall. Imaging errors that cluster in one modality — echo rather than plain films, say — point to a specific, trainable weakness rather than a general one. And a viva that collapses under a single follow-up question reveals that your answer was memorised rather than reasoned. Log these qualitatively, recording what failed and why, because unlike an SBA score the signal here is diagnostic rather than numeric, and it tells you which rung of the ladder to repeat.
How much performance practice is enough?
There is no published magic number of cases, and any figure is a guide rather than a standard, but the useful targets concern consistency, not count. For the SOE, aim for enough vivas — realistically several a week across the run-in — that you can deliver a structured answer to an unfamiliar prompt without preparation; the exit signal is unfamiliarity handled well, not a tally reached. For imaging and data, short daily sets beat occasional long ones, because interpretation is a perishable skill that rewards frequency. For the CRQ, a smaller number of answers marked against a rubric is worth more than many written and self-assessed. The honest test is not how many you have done but whether your performance on unseen material has stabilised — the same readiness logic that governs the Final FRCA content-gap checklist.
Frequently asked questions
How do I know whether I have covered the full Final FRCA blueprint? Coverage means more than breadth of knowledge; for the Final FRCA it means you have trained each modality across the blueprint. Build a matrix with a row per domain and a column per skill — judgement, evidence, imaging, viva — and confirm every cell has had deliberate practice, not just that you have done questions in that domain. A completed knowledge map with empty performance columns is not coverage, and the companion content-gap checklist turns this into a tick-list.
Can one question bank be enough for Final FRCA? No single MCQ bank can be enough, because two of the three components — the CRQ and the SOE — test performance a bank cannot assess. A bank is the right tool for the SBA paper and for building the knowledge the other components draw on, but it cannot train constructed written answers or spoken defence. Treat a bank as one instrument in a set that must also include marked CRQ practice, imaging and data work, and viva rehearsal with a clinician.
What should I measure instead of my overall Q-bank percentage for Final FRCA? Measure performance, not just recognition: your CRQ marks against a rubric, your accuracy on unseen imaging and data at exam pace, and the quality of your spoken defence under challenge as judged by a clinician. On the MCQ itself, track first-attempt accuracy on unseen items by domain rather than a blended figure. A single overall percentage tells you nothing about the two components most likely to fail you, so it is the wrong headline number to watch.
When should I stop doing new Final FRCA questions? Stop adding new MCQs once your unseen, first-attempt SBA score is stable across sittings and no blueprint domain is under-attempted — and redirect that time to the modalities a bank cannot train. For most candidates near the exam, the marginal hour is far better spent on CRQ practice, imaging and viva defence than on additional selected-response items. Continue new questions only for a specific, measured knowledge gap, not as a default comfort activity.
Which Final FRCA resource should I use for my weakest component? Match the resource to the modality of the weakness rather than reaching for the bank by habit. Weak clinical judgement or CRQ construction calls for timed constructed-response practice marked by a senior; weak imaging calls for daily data-interpretation sets with model reports; weak oral defence calls for viva practice with a clinician who can challenge you; weak SBA breadth calls for an unseen bank. iatroX supplies the unseen SBA measurement and knowledge layer; it is explicitly not an SOE or station simulator, so pair it with structured-oral and imaging practice for those components.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. Exam-format facts are from the Royal College of Anaesthetists Final FRCA pages and the RCoA 2027 changes hub; because the format changes from July 2027, re-verify your components against the official pages before relying on this map. Any third-party figures elsewhere in this backlog are vendor-reported and dated.
Disclosure: iatroX operates a Final FRCA question bank and competes with other knowledge products. This article is candid that a bank — including iatroX — cannot prepare you for the CRQ or the SOE, and confines iatroX's role to the unseen SBA measurement and knowledge layer that the other modalities build upon. iatroX is not an SOE, OSCE or station simulator. Corrections are welcome via the feedback route on iatrox.com.
References: Royal College of Anaesthetists — Final FRCA examination and Final FRCA Written examination; RCoA 2027 Launch of the New FRCA Exams; iatroX — Final FRCA content-gap checklist; iatroX — calibrating AI-graded feedback; iatroX comparison hub.