Medibuddy is one of the few MSRA banks that explicitly covers both papers and drills them with an adaptive engine — a real advantage, and a dashboard that can still mislead you about readiness. This audit is for specialty-training applicants deciding how far to trust its Topic Mastery scores and percentages. The principal limitation: the MSRA is two very different papers, and an adaptive engine tuned on clinical multiple-choice tells you least about the Professional Dilemmas paper that carries roughly half your score.
What Medibuddy offers for MSRA right now
Medibuddy's MSRA bank holds over 4,000 questions, including 190+ Professional Dilemmas scenarios, and explicitly covers both the Professional Dilemmas and Clinical Problem Solving components. Its AI engine calculates a Topic Mastery Level per subject, re-tests earlier errors through spaced repetition, and offers adaptive (algorithm-driven) and traditional (manual-filter) modes. Access runs £30 for one month, £40 for three, £50 for six and £60 for twelve. That PD coverage is a genuine differentiator — many clinical banks ignore the paper entirely — but as we will see, a bank's PD questions are a rehearsal aid, not a substitute for consensus-scored official material.
The exam the dashboard must answer to
The MSRA is 170 minutes in two papers, per NHS England's published structure: a Professional Dilemmas paper of 95 minutes (around 50 situational-judgement scenarios in ranking and selection formats, scored against expert consensus) and a Clinical Problem Solving paper of 75 minutes (single-best-answer and extended-matching across primary-care-weighted medicine, in the region of 86 items at a pace under a minute each). Two consequences: CPS pace is the hidden examiner, and PD judgement cannot be captured by any clinical accuracy metric.
What each metric actually measures
First-attempt accuracy on unseen items is the closest thing to a CPS predictor. Topic Mastery Level is a modelled competence estimate based on the items the algorithm chose — a useful weakness signal, not a pass probability. Repeat accuracy, inflated by the spaced-repetition re-tests, measures recognition. Percentile against other users compares you with a self-selected population at every stage of preparation. Difficulty and time-per-item are the underused pair, and for the MSRA specifically, per-domain time data matters more than accuracy, because the CPS paper punishes slowness more brutally than almost any UK exam. And crucially: Medibuddy's PD questions generate their own accuracy numbers, but PD is scored against consensus about professional judgement, so a high PD "score" in a bank is a familiarity signal, not a validated readiness measure.
Adaptive selection bias, in one paragraph
Every number the dashboard shows was generated under conditions the algorithm or your filters chose. Adaptive mode concentrating on weak areas deflates recent accuracy; traditional mode on comfortable domains inflates it; spaced re-tests raise repeat accuracy while predicting little. None is comparable with a mixed, unseen, timed block — the only condition that resembles the day. The general version of this argument applies to MSRA preparation with extra force because of the split-paper problem: even a perfectly-read CPS dashboard is silent about half the exam.
Audit your distribution against the exam, not the app
Fortnightly, compare three things. Your attempted CPS mix against the specification's primary-care-weighted spread — hospital-flavoured comfort zones reliably over-represent themselves. Your CPS time-per-item against the sub-minute pace the paper demands. And your total PD practice — timed, consensus-reviewed — against the roughly half of the assessment PD represents. For most Medibuddy-centric candidates the last ratio is the uncomfortable one, and fixing it matters more than another few hundred clinical questions or another row of green Mastery bars.
What a credible readiness signal requires
For CPS: unseen items, timed at exam pace, mixed across domains, no assistance, and a sample of at least 100 questions across multiple sittings. For PD: timed official practice and consensus-scored scenario work — a clinical bank's PD accuracy does not reach it, however many PD questions the bank holds. If your readiness evidence is a Medibuddy overall percentage, you have evidence about part of one paper, collected under conditions unlike the exam.
When to override the algorithm
Force full-pace timed mixed CPS blocks in traditional mode (the habit the adaptive feed will not prioritise); force the low-volume clinical domains the CPS spec still samples; force data-interpretation and calculation items if your mix lacks them; and schedule PD sessions on official-style material that no clinical algorithm will ever generate for you. The adaptive engine is a good clinical teacher; it is not an MSRA examiner, and it has no view of the paper that carries half your marks.
A worked dashboard example
Six weeks out, suppose Medibuddy shows: CPS overall 71%; Topic Mastery high in women's health and dermatology, low in musculoskeletal and ophthalmology/ENT; adaptive feed currently weighted to your weak clinical topics; average pace 66 seconds per item; PD practice: 60 bank scenarios attempted, none timed, none consensus-reviewed. Read properly: musculoskeletal and the thin specialties take the largest CPS quota; pace at 66 seconds needs three timed blocks aimed under 55; and PD is the real gap — those 60 untimed scenarios become two timed, consensus-reviewed sessions this week, because a bank PD score is not a validated readiness measure. No pass prediction appears anywhere, because none of these numbers can honestly produce one.
A seven-day pattern for applicants
Monday: 50 Medibuddy CPS questions in adaptive mode, explanations read. Tuesday: a timed PD session — using Medibuddy's PD bank to rehearse format, then reviewing your ranking logic against official-style consensus rationale. Wednesday: a timed, unseen 50-question CPS block in iatroX's free MSRA practice, whose adaptive engine probes related weaknesses across domains — your uncontaminated CPS signal. Thursday: 30 forced-pace CPS questions plus error review. Friday: second PD session. Saturday: a full-length timed CPS simulation, alternating source weekly; same-day review by error type. Sunday: rest. Medibuddy drills CPS adaptively and rehearses PD format; official material anchors PD judgement; iatroX measures on unseen items. Nobody's dashboard is asked to measure what it cannot see.
The single number that misleads MSRA candidates most
If there is one figure to distrust on the MSRA, it is a Medibuddy overall percentage read as a readiness score. It misleads in three compounding ways at once. It is inflated by the spaced-repetition re-tests, so it drifts upward through a subscription regardless of durable learning. It is shaped by the adaptive feed's choices, so it reflects the questions the algorithm served rather than the exam's distribution. And it is silent about the entire Professional Dilemmas paper, which carries comparable weight and which no clinical accuracy metric can see. A candidate quoting "I'm averaging 74% on Medibuddy" is quoting a number that is simultaneously too high, unrepresentative, and about half the exam.
The honest replacement is a three-part statement: unseen-timed CPS accuracy over a real sample, CPS pace against the paper's sub-minute budget, and PD performance against consensus-scored material. It is less comforting and far more predictive. Note what the Medibuddy dashboard cannot supply even in principle: the third component. Because PD is scored against expert consensus about judgement, a bank's PD accuracy — however many of its 190-plus scenarios you attempt — is a familiarity signal, not a validated readiness measure. Building the habit of stating readiness in those three parts, rather than as a single Medibuddy percentage, is the fastest cure for the false confidence that sinks otherwise well-prepared applicants who discover the Professional Dilemmas paper too late.
Continue, supplement, switch or stop
Continue while unseen timed CPS performance climbs and your domain audit stays level. Supplement the moment bank percentages rise faster than unseen performance, and supplement PD immediately if your PD practice is untimed or unreviewed. Switch only for a measurable CPS gap — the MSRA market is crowded (we compare the options here). Stop accumulating clinical questions in the final fortnight; the last two weeks belong to timed simulations of both papers.
The bottom line for applicants
The honest one-line verdict on Medibuddy for the MSRA: it is one of the better single banks precisely because it covers both papers and drills clinical medicine adaptively — and its Topic Mastery scores still cannot tell you whether you are ready, because an adaptive engine measures the feed it gave you, not the exam. Two disciplines make it work. First, treat Professional Dilemmas as a separate, consensus-scored project from week one; its 190-plus PD scenarios rehearse format, but a bank PD score is familiarity, not readiness, and the candidates who leave PD until the end are the ones whose strong clinical scores do not save them. Second, keep your readiness signal outside the adaptive feed — an unseen source it never selected — so the number you trust in the final fortnight is cold first-attempt accuracy, not recognition dressed up as progress. Get those two right and Medibuddy is a genuinely strong core; get them wrong and its reassuring dashboard is exactly the false comfort that sinks well-prepared applicants.
Frequently asked questions
Is Medibuddy enough for MSRA on its own? It is one of the more complete single banks because it covers both papers, but its PD questions rehearse format rather than validate judgement, so you still need consensus-scored official PD material and unseen timed CPS blocks from outside your practice history.
Which MSRA component does Medibuddy not reproduce well? The Professional Dilemmas paper's consensus scoring — a bank can hold PD scenarios and still not tell you how the exam would score your judgement — and, within CPS, the sub-minute pacing pressure that untimed adaptive practice does not build.
How many Medibuddy questions should I complete per day for MSRA? Around 40–60 CPS questions on study days, distributed by your blueprint audit rather than the adaptive feed's preference, with PD sessions and timed blocks protected separately; composition beats raw count.
When should I stop using Medibuddy and move to mixed mocks? Once every CPS domain has meaningful coverage, timed-block accuracy is stable and your pace is under the paper's per-item budget — usually two to three weeks out — shift to full timed simulations of both papers and use the bank for error review only.
How should I combine Medibuddy with iatroX without duplicating practice? Use Medibuddy for adaptive CPS drilling and PD format rehearsal, and iatroX for unseen, timed, adaptively selected CPS blocks that measure transfer — iatroX's MSRA practice is free, so the two-bank design costs the price of one and keeps your readiness signal honest.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; Medibuddy figures (4,000+ questions incl. 190+ PD; £30–£60 tiers; both papers covered) are vendor-published — verify before purchase. MSRA structure timings are NHS England's. Disclosure: iatroX offers a free competing MSRA bank; this audit confines iatroX's role to jobs Medibuddy does not claim. Corrections via the feedback route on iatrox.com. References: NHS England MSRA structure (medical.hee.nhs.uk); Medibuddy MSRA product pages (medibuddy.co.uk); related reading: the best MSRA revision apps and why your Q-bank percentage is not your exam score.