Lead finding first, because it changes the plan: MedRevisions does not currently offer an MSRA product. On 19 July 2026 its own site describes an adaptive question bank for PLAB 1 and the UKMLA Applied Knowledge Test, not the Multi-Specialty Recruitment Assessment. So if you are preparing for the MSRA, you cannot run this workflow inside MedRevisions. What you can do is take the adaptive first-pass, review and exit discipline MedRevisions is built around, run it on a bank that actually covers the MSRA, and reserve iatroX for the unseen measurement — with MedRevisions used only if, and where, PLAB 1 or the UKMLA AKT is separately on your path.
What MedRevisions offers right now
Current-state box, vendor-reported and last checked 19 July 2026 (verify the current figures on medrevisions.com):
- Exams covered: PLAB 1 and UKMLA AKT only. No MSRA question bank, no Professional Dilemmas content, no MSRA mocks.
- Question volume: the vendor states 5,400-plus SBAs mapped to the 2026 MLA content map, with 30-plus customisable mocks.
- Adaptive and AI features: an "AI Professor" tutor, an "Adaptive Insight Engine" producing a real-time readiness score, weakness-targeted mocks, and AI-generated study plans and flashcards aimed at low-scoring areas.
- Access and price: a small free sample tier of roughly 70 questions, with paid access vendor-reported from about US$10.83 per month; confirm the current currency and tiers on the page.
The adaptive features are real and, for their target exams, a reasonable offering. The problem is not quality — it is fit. An adaptive engine tuned to the MLA content map cannot generate the MSRA's primary-care-weighted clinical distribution, and it contains no situational-judgement paper at all.
Exam anchor: what the MSRA actually is
The MSRA is a computer-based assessment at Pearson VUE, 170 minutes long, in two independently timed papers. Professional Dilemmas runs 95 minutes and presents roughly 50 situational-judgement scenarios that are consensus-scored against a panel, using ranking and multiple-best formats rather than simple right-or-wrong marking. Clinical Problem Solving runs 75 minutes and presents about 86 single-best-answer and extended-matching items, weighted toward common and primary-care presentations, at a sub-minute pace of roughly 52 seconds per item. Official requirements come from the current NHS England applicant guidance; third-party mappings are approximations, and a bank built for a different exam is not one of them.
The workflow, and where it can and cannot run
The first-pass, review and exit discipline is exam-agnostic: it is how you turn any adaptive bank into a measured, honest revision engine. The only rule that matters here is that the bank must cover the exam you are sitting. For the MSRA that means an MSRA-covering adaptive or analytics bank for the day-to-day loop, and iatroX unseen blocks kept back as the assessment layer. Where you are an international graduate whose PLAB 1 or UKMLA AKT still lies ahead, MedRevisions can legitimately run this loop — but in that phase, for that exam, not inside your MSRA week.
Baseline week: measure before the algorithm personalises
Before you let any adaptive feed reshape your revision, complete a small, blueprint-stratified unseen sample: two or three timed Clinical Problem Solving blocks of 40 to 50 items spread across the domains, plus a short set of official Professional Dilemmas sample scenarios. That baseline gives you first-attempt accuracy by domain, a pace figure, and a Professional Dilemmas calibration read — numbers the adaptive engine will otherwise obscure once it starts feeding you your weak areas.
First pass: set domain floors so the feed cannot hide gaps
An adaptive engine over-samples what you are bad at, which is good teaching but dangerous accounting: your headline score can rise while whole domains sit unattempted. Impose domain floors — a minimum number of first-attempt items per domain before you trust any readiness figure. On the first pass, work every domain up to its floor regardless of what the algorithm queues, so the readiness score cannot flatter you by routing around your gaps. Record first-attempt accuracy separately from repeat accuracy; only the former approximates unseen performance.
Error taxonomy: sort every miss before you act
Do not treat all wrong answers the same. Sort each into one of six buckets: a genuine knowledge gap, a misread stem, premature closure (committing before reading the whole vignette), a guideline error (out of date with NICE or CKS), a calculation error, or a time-pressure error. The bucket determines the fix. A knowledge gap earns a short source read and a spaced re-test on a new item; a misread or premature-closure error earns a technique change, not more content; a guideline error earns a check against current UK guidance; a time-pressure error earns pacing drills, not revision.
Review interval: not everything deserves a repeat
Decide, per error, whether it warrants a fresh transfer question later, a spaced review, or a quick source read — rather than an immediate repeat of the same stem. Immediate repeats teach recognition of that item, not mastery of the concept. Space knowledge-gap items to 48 hours and again to a week on new questions that test the same idea; retire careless errors with a technique note rather than a re-test.
Mixed-block switch: from topic feeds to random blocks
Adaptive, topic-filtered practice is how you learn; timed random blocks are how you rehearse. Switch the balance toward mixed blocks when two conditions hold: your first-attempt accuracy has cleared your domain floors, and your single-topic sessions have stopped surfacing new errors. From that point, reduce filtered practice and increase full-length, mixed, timed blocks so that topic order stops cueing your answers.
Exit criteria that preserve unseen questions
Exit the bank — stop adding new questions from it — when five things are true at once, not when a completion bar fills: coverage floors met across domains; first-attempt accuracy stable at target for about two weeks; pace comfortably under 55 seconds per Clinical Problem Solving item; misses now mostly careless rather than knowledge-based; and an official-material calibration passed. Crucially, do not consume your unseen assets to get there. Keep a reserve of genuinely unseen mixed blocks — iatroX is designed for exactly this — as the final, uncorrupted read of readiness. A bank you have fully worked can no longer measure you; only unseen items can.
Worked example: a seven-day pattern
For a specialty-training applicant, keep tools in their lanes and make no claim to any proprietary algorithm. Days 1 to 5: run the adaptive first-pass and review loop on your MSRA-covering bank — domain floors, error sorting, spaced re-tests. Days 3 and 6: sit a fresh, timed, mixed, unseen Clinical Problem Solving block in iatroX to measure transfer independently of the bank that taught you. Day 7: work official Professional Dilemmas sample scenarios and reconcile every answer against the published rationale. If PLAB 1 or the UKMLA AKT is separately ahead of you, MedRevisions' adaptive engine belongs in that plan, on a different week, for a different exam.
Decision checklist: continue, supplement, switch or stop
- Continue using MedRevisions only if your current target is PLAB 1 or the UKMLA AKT, where its adaptive engine is on-blueprint.
- Supplement is the wrong frame for the MSRA: you cannot supplement an MSRA plan with a bank that has no MSRA content.
- Switch to an MSRA-specific bank for Clinical Problem Solving and to official material plus a structured approach for Professional Dilemmas.
- Stop any spend on MedRevisions made in the belief it covers the MSRA, and reallocate it to a bank that does.
Frequently asked questions
Is MedRevisions enough for MSRA on its own? No, and not because it is weak but because it does not cover the MSRA at all. On 19 July 2026 MedRevisions offers PLAB 1 and UKMLA AKT question banks only, with no MSRA clinical content and no Professional Dilemmas paper. It therefore cannot be any part, let alone the whole, of an MSRA stack. Verify the current exam list on medrevisions.com before you spend.
Which MSRA component does MedRevisions not reproduce well? It reproduces neither MSRA component. It has no situational-judgement product, so Professional Dilemmas is entirely absent, and its clinical SBAs are mapped to the MLA content map for PLAB and the UKMLA rather than to the MSRA's primary-care-weighted blueprint, so even the clinical overlap is calibrated to a different exam and a different pace.
How many MedRevisions questions should I complete per day for MSRA? None, because none of them are MSRA questions. If you are separately sitting PLAB 1 or the UKMLA AKT, follow a plan for that exam and set your daily count there. For the MSRA, set your daily Clinical Problem Solving quota on an MSRA-specific bank and measure transfer on unseen mixed blocks; MedRevisions has no role in that number.
When should I stop using MedRevisions and move to mixed mocks? For MSRA preparation the question does not arise, because you should not be using MedRevisions for the MSRA in the first place. For a PLAB 1 or UKMLA AKT candidate, the usual rule applies: move from adaptive topic feeds to timed mixed mocks once first-attempt accuracy on unseen items is stable at target for about two weeks and your domain floors are met.
How should I combine MedRevisions with iatroX without duplicating practice? For the MSRA you would not combine them, because MedRevisions offers no MSRA content to combine. If you are an international graduate on the PLAB-to-MSRA pathway, keep the tools in separate lanes and separate exams: any MedRevisions use belongs to your PLAB or UKMLA phase, while iatroX unseen MSRA blocks belong to your MSRA phase, so there is nothing to duplicate.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Product facts are vendor-reported and change between diets; verify the current exam list, counts and prices on medrevisions.com. Disclosure: iatroX operates a competing MSRA question bank; because MedRevisions does not cover the MSRA, iatroX's role here is confined to unseen MSRA measurement, which MedRevisions does not claim to provide. For the wider principles behind trusting an adaptive engine, see the iatroX guide to auditing an AI medical-exam tutor. Corrections are welcome via the feedback route on iatrox.com.
References: MedRevisions product pages (medrevisions.com, accessed 19 July 2026); NHS England, Multi-Specialty Recruitment Assessment structure and applicant guidance (medical.hee.nhs.uk); iatroX, "Your Q-Bank Percentage Is Not Your Exam Score" (iatrox.com/blog); iatroX MSRA hub and comparison pages (iatrox.com/msra, iatrox.com/compare).