This is for MRCEM SBA candidates revising around clinical shifts who already use, or are considering, FRCEMTutor. It addresses the written single-best-answer papers, and it opens with an honest correction: as checked on 19 July 2026, FRCEMTutor documents no adaptive difficulty algorithm. Its "personalisation" is a peer-comparison histogram plus the topic filters and test settings you choose yourself. So "following the algorithm" here really means following that signal and your own filters — and the principal limitation is that self-selected filters can hide blind spots behind a rising score.
Current-state box (checked 19 July 2026)
Figures are vendor-reported from FRCEMTutor and should be re-verified before purchase.
| Item | What the product states (19 July 2026) |
|---|---|
| MRCEM Intermediate SBA questions | Over 1,250, mapped to the RCEM specialty learning outcomes (vendor-reported) |
| Access and price | 2 months £20; 3 months £25; 4 months £30; 6 months £35 (vendor-reported) |
| Adaptive / AI engine | Not documented. No claim that difficulty adjusts automatically; tests are user-configured |
| Analytics | Performance tracking with a peer-comparison histogram updated daily |
| Content | Illustrated teaching notes, answer feedback, personal revision notes, question flagging |
| Practice modes | Unlimited timed tests and user-built mock exams; mobile app |
The strengths are real: a large, blueprint-mapped bank at a modest price, with illustrated notes and a peer histogram that many candidates find motivating. The caveat is equally real: because there is no adaptive engine, nothing corrects for the natural drift toward the topics you enjoy. The workflow below is built to supply that correction manually.
Exam anchor
The MRCEM SBA is 180 single-best-answer questions across two two-hour papers of 90, machine-marked with no negative marking, at a pace near 80 seconds per item, with the pass mark set per sitting by modified Angoff (RCEM, checked 19 July 2026). It maps to the 2021 RCEM curriculum and its specialty learning outcomes, and the authoritative calibration standard is RCEM's own sample questions. Distinguish that official requirement from any third-party claim: a vendor's "exam-style" mock is practice volume, not the blueprint, and only the RCEM material defines the standard.
Baseline week: measure before you personalise
Before you let any bank — or your own filtering habits — shape your feed, take a blueprint-stratified unseen baseline. Draw a small sample, perhaps 60 to 90 questions, spread deliberately across the six domains in rough proportion to their weighting: Complex or Common Stable Patient, Resuscitation, Injured Patient, Paediatric EM, Procedural Skills, and Complex or Challenging Situations. Sit it timed and unseen. The result is not a grade; it is a map of where you actually stand, and it stops you personalising your way into your comfort zone from day one.
First pass: set domain floors
Now set floors so no domain can hide. Decide a minimum number of attempted items per domain — scaled to the blueprint weighting — that you will hit before you allow yourself to over-invest anywhere. FRCEMTutor's topic filters make this easy to run and easy to abuse; used with floors, they guarantee breadth, and the peer histogram then tells you where you sit within each domain rather than being read as a single reassuring average across all of them. The floor is the antidote to a rising overall score that quietly leaves paediatric EM or procedural skills untouched.
Error taxonomy: label every miss
An unlabelled wrong answer teaches almost nothing. Tag each miss into one of six types, because the fix differs by type:
- Knowledge gap — you did not know it. Needs a source read and a spaced re-test.
- Misread stem — you knew it but misread the question. Needs slower reading, not more content.
- Premature closure — you committed before weighing the options. Needs a deliberate options-comparison habit.
- Guideline error — you gave a superseded answer. Needs a recency check against current UK guidance.
- Calculation error — you slipped on a dose or rate. Needs drilled calculation practice.
- Time-pressure error — you rushed near the end of the block. Needs pacing practice, not new topics.
Use FRCEMTutor's personal-notes and flagging features to record the type, not just the topic. After a week, the tally tells you what kind of studying you actually need.
Review interval: not everything deserves a repeat
Match the response to the error type. A knowledge gap deserves a new transfer question on the same concept a few days later — not an immediate repeat of the identical item, which only trains recognition of that item. A misread or time-pressure error deserves a pacing or reading fix, not content. A guideline error deserves a short source read against the current guidance and its date. Reserve immediate repeats for almost nothing; space the genuine gaps so you are testing retention, not short-term memory.
Mixed-block switch: when to reduce topic-filtered practice
Topic-filtered practice is for building a weak domain; it is not what the exam is. Switch toward timed random blocks when three objective conditions are met: every domain has cleared its floor; first-attempt accuracy on your weaker domains has stopped rising steeply; and you are within a few weeks of the exam. At that point, reduce filtered sets and increase full-length, mixed, timed blocks that mimic a 90-item paper, so you train domain-switching and pacing under realistic conditions rather than staying in the comfort of one topic at a time.
Exit criteria
You are ready to taper new questions when all of the following hold — not when the bank reports completion:
- Coverage floor cleared in every domain, in blueprint proportion.
- Stable first-attempt performance on unseen mixed blocks, not repeat accuracy.
- Pacing under control across a full 90-item block, with no late collapse.
- Retention confirmed by spaced re-testing of older material.
- Official-material calibration against RCEM sample questions holding up.
Bank completion alone is not on that list, and neither is your peer percentile.
A seven-day plan: one job each
Use FRCEMTutor for one defined job and a second unseen bank for transfer. This makes no claim to any proprietary algorithm; it simply divides learning from measurement.
- Day 1 — Blueprint-stratified unseen baseline (above); label every miss by error type.
- Days 2–4 — FRCEMTutor for one job: filtered practice in your two weakest domains, using the illustrated notes to close knowledge gaps, floors enforced, misses tagged by type.
- Day 5 — Recency and calibration: reconcile guideline-sensitive misses against current UK guidance, then sit a RCEM official sample set.
- Days 6 — Transfer: run a fresh, unseen, mixed, timed block in the free iatroX MRCEM SBA bank, which maps to the 2021 curriculum and surfaces weak topics, so you measure recall on items you have never seen rather than re-scoring FRCEMTutor questions.
- Day 7 — Re-audit the coverage table and decide the next week from the measured gaps.
This is the two-Q-bank rule: one bank to learn, a second unseen bank to measure, with no duplicated items to corrupt calibration. For the fuller coverage instrument this workflow feeds, see the MRCEM SBA content-gap checklist.
Decision checklist: continue, supplement, switch or stop
- Continue FRCEMTutor while it is still closing knowledge gaps in your weak domains and you have unseen items left.
- Supplement when your peer histogram looks healthy but your own coverage audit shows a thin or weak domain the filters have let you skip; add a second unseen bank and RCEM samples.
- Switch only for a measured reason — you need genuine adaptive weak-topic surfacing FRCEMTutor does not offer, or materially different content — not because a rival looks new.
- Stop adding new questions when the exit criteria are all met; then simulate and rest.
Every branch turns on a measured gap, not on novelty or the sunk cost of a subscription.
Frequently asked questions
Is FRCEMTutor enough for MRCEM SBA on its own? It can carry most of the learning load — the bank is large, blueprint-mapped and reasonably priced (over 1,250 items, vendor-reported, 19 July 2026) — but it should not be your only measurement, because you cannot test transfer on items you have already worked through, and it documents no adaptive engine to force your weak areas up. Pair it with RCEM's official sample questions for calibration and a second, unseen bank for transfer. As a single learning resource it is a strong option; as a complete readiness system on its own, it is not.
Which MRCEM SBA component does FRCEMTutor not reproduce well? Its main gap is not a topic but a mechanism: with no documented adaptive difficulty engine (checked 19 July 2026), it will not automatically surface the domains you avoid, so paediatric EM, procedural skills and the smaller challenging-situations domain can stay under-practised behind a healthy peer histogram. You must impose domain floors yourself. Verify on the product page whether time-critical image and ECG interpretation is well represented for your needs, and supplement with targeted resources and RCEM samples if it is thin.
How many FRCEMTutor questions should I complete per day for MRCEM SBA? Coverage and conditions matter more than raw count, but 30 to 50 items on a study day is a workable pace for a busy trainee, weighted toward your audited weak domains and with every miss tagged by error type rather than merely marked wrong. Around exam approach, convert some of that daily volume into full-length timed mixed blocks so you train pacing across a 90-item paper. Do not chase a daily target for its own sake; a smaller number of well-reviewed, correctly-labelled misses beats a large number skimmed.
When should I stop using FRCEMTutor and move to mixed mocks? Move to mixed mocks when every domain has cleared its floor, first-attempt accuracy on your weaker domains has plateaued, and you are within a few weeks of the exam — because beyond that point, filtered single-topic practice no longer resembles the paper, and pacing plus domain-switching become the binding constraints. You need not abandon FRCEMTutor entirely; you shift its role from topic drilling to occasional gap-filling while full-length, unseen, timed blocks become your main measurement.
How should I combine FRCEMTutor with iatroX without duplicating practice? Give each a single job. Use FRCEMTutor to learn — filtered practice, illustrated notes, error-type tagging — and use the free iatroX MRCEM SBA bank to measure transfer on unseen, mixed, timed blocks mapped to the 2021 curriculum, logging first-attempt accuracy by domain. Never re-answer the same items in both, because duplicated questions inflate your scores and wreck the calibration a second bank exists to provide. If both coverage audits flag the same weak domain, close it against RCEM sample questions and current guidance. The comparison hub shows which tool owns which job.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; question counts, prices and features are vendor-reported and change without notice, so verify current figures on the FRCEMTutor product pages before purchasing. Where this article notes an absence — such as no documented adaptive difficulty engine — it reflects what the product pages state on the checked date, not a claim about internal design. Disclosure: iatroX operates a competing, free UK MRCEM SBA question bank; this article confines iatroX's role to the unseen-measurement and transfer jobs the audited FRCEMTutor product does not claim, and directs calibration to RCEM's official sample questions. Corrections are welcome via the feedback route on iatrox.com.
References: RCEM — MRCEM exams and the 2021 curriculum with official sample questions (rcem.ac.uk/mrcem-exams); FRCEMTutor MRCEM Intermediate SBA product page (frcemtutor.com/imrcem); iatroX MRCEM SBA landing and comparison hub (iatrox.com/mrcem-sba, iatrox.com/compare); iatroX — "Your Q-Bank Percentage Is Not Your Exam Score", the two-Q-bank rule pillar and the MRCEM SBA content-gap checklist.