This workflow is for candidates using MediWord for the UKMLA applied knowledge test — MSC AKT for UK students, PLAB 1 for IMGs — who want a disciplined topic-to-mixed progression rather than an open-ended grind through a recall bank. It covers the written applied-knowledge component only. The principal limitation to design around: MediWord's questions are recall-based, reconstructed from candidate memory, so their phrasing and blueprint balance are less controlled than an official paper, and you should calibrate against official material rather than trust recall wording as gospel.
What MediWord offers for UKMLA right now
The figures below are vendor-reported and last checked on 19 July 2026; confirm them on the MediWord site, as counts and prices change.
- Coverage: a vendor-reported 3,000+ recall-based UKMLA questions, plus 900+ revision notes and 5+ full mock papers, aligned to GMC guidance; MediWord's wider platform also spans MSRA, MRCP, MRCS and UCAT.
- AI features: a final-week "AI tutor for instant guideline-referenced answers"; the bank itself is a conventional topic-and-mock structure rather than a documented adaptive-difficulty engine, so do not expect it to protect your coverage automatically.
- Access and price: subscription tiers vendor-reported at £12 for one month, £25 for three months and £35 for four months, with identical content across tiers.
- Components: it addresses the applied knowledge test only and does not reproduce the separate clinical and professional skills assessment.
Because the questions are recall-derived, treat MediWord as an exposure and pattern-familiarity tool, and hold the official RCGP-equivalent material — here, the GMC's own UKMLA resources — as your source of truth on phrasing and standard.
The exam anchor
The UKMLA comprises an applied knowledge test (MSC AKT for UK students; PLAB 1, 180 single-best-answer questions in 3 hours, for IMGs) and a separate clinical and professional skills assessment. Both rest on the GMC's MLA content map, updated January 2026 and applying from September 2026, organised around areas of clinical practice, presentations and conditions, at roughly one minute per item. A recall bank samples what previous candidates remembered being asked; the content map defines what you are actually accountable for, and only the latter tells you whether your coverage is complete.
Build a blueprint inventory and protect an unseen pool
Start by mapping MediWord's topics onto the MLA areas of clinical practice and marking, per domain, whether you have an adequate volume of unseen items reserved for timed assessment. Deliberately hold back a protected pool of questions you will not touch during first-pass practice, so that at the end you have genuinely unseen material for a clean mixed mock. In a recall bank this discipline matters more than usual, because repeated exposure to reconstructed items produces recognition quickly, and recognition is precisely what you must keep out of your measurement.
First pass: topic blocks only where foundations are weak
Use topic-filtered blocks only for domains where your foundations are genuinely shaky; everywhere else, start mixed. The reason is topic cueing — a filtered renal block silently tells you the answers are renal, which flatters your accuracy and does not rehearse the real decision of recognising the system from the vignette. Weak domains earn a short filtered scaffold to build baseline knowledge; sound domains go straight into mixed practice.
Review each miss with a code and one action
For every wrong item, assign one error code — knowledge gap, misread stem, premature closure, guideline error, calculation error or time-pressure error — and commit to a single corrective action. Do not transcribe the whole explanation into notes; that is copying, not learning, and it eats the time retrieval needs. One miss, one code, one action: a short source read, a technique drill, a guideline check, or a fresh transfer item, depending on the code.
Use transfer practice before repeating
When you miss an item on a knowledge gap, answer a new question testing the same principle before you ever revisit the original. Repeating the identical recall item teaches you that item's answer; a transfer item tests whether you understand the concept in a different dress. This is the single most important habit with a recall bank, because recall banks are where "I've seen this one" most easily masquerades as competence.
Switch to mixed timed blocks when floors are met
Move to timed mixed blocks once your domain floors are met, even if your first pass through the bank is incomplete. Completion is not the trigger; coverage plus stable unseen accuracy is. Finishing every recall question is a poor use of the closing weeks compared with rehearsing full-length, timed, mixed papers under exam conditions and calibrating against official material.
Exit criteria
You are ready when coverage floors are met across the blueprint, your unseen timed performance is stable, your mocks finish inside time, previously missed items stay corrected on delayed re-testing, and you calibrate cleanly against official material — not when you reach 100% completion of the recall bank. Completion measures the bank; the criteria above measure you.
Worked example: an IMG's seven-day week
Take an international graduate using MediWord for exposure to UK-style items and iatroX for unseen transfer practice. MediWord's defined job is pattern familiarity and mock rehearsal; iatroX's job is fresh, unseen, NICE/CKS-grounded items that confirm transfer, run without any claim to its internal algorithm.
- Monday: MediWord topic block in the weakest MLA area; code every miss; one action each.
- Tuesday: short source reads on Monday's gaps against current NICE or CKS; a few iatroX unseen items on the same principles.
- Wednesday: MediWord mixed block across cleared domains; audit attempted-volume floors.
- Thursday: UK-conventions focus — prescribing norms, safety-netting, referral pathways — where overseas graduates are most exposed; iatroX unseen items to test transfer.
- Friday: one MediWord mock paper, timed; record completion and pacing.
- Saturday: spaced review of coded misses; transfer items before any repeat.
- Sunday: rest, or a short unseen iatroX block as a clean progress check.
Reading your results and three mistakes to avoid
With a recall-based bank the way you read your own results matters more than usual, because recognition accrues fast and disguises itself as knowledge. Judge yourself on first-attempt accuracy on items you have genuinely not seen before, broken down by MLA area, and on whether your mocks finish inside time — not on how much of the reconstructed bank you have cleared.
Three mistakes cause most trouble here. The first is trusting recall phrasing as though it were the official wording: candidate reconstructions drift, so calibrate pacing and standard against the GMC's own UKMLA materials and let those, not the recall bank, define what "right" looks like. The second is re-answering items you have already met and reading the inflated score as progress; a recall bank is where "I've seen this one" most convincingly imitates competence, which is why the protected unseen pool and the transfer-before-repeat rule exist. The third is transcribing whole explanations into notes in the belief that copying is studying — it consumes the time that active retrieval needs and produces a document you rarely reopen. One miss, one code, one action keeps the review lean.
The deeper point is that a cheap, high-exposure bank is a good exposure tool and a poor measurement tool, and the two jobs must not be confused. Use MediWord to meet a wide variety of item patterns and to rehearse pacing on its mock papers, and use a separate, unseen source to measure whether your understanding is real. If your in-app accuracy is high but your unseen accuracy lags, that gap is the single most useful number you will see, and it is invisible if you only ever look at the bank's own running total. None of this makes MediWord a weak choice at its price; it makes it a specific one — treat it as breadth and exposure, anchor to official material for standard, and keep one unseen source outside it for measurement, and the recall format becomes an asset rather than a trap.
Decision checklist: continue, supplement, switch or stop
- Continue MediWord if its recall items are surfacing genuine gaps and your unseen scores are rising.
- Supplement with an unseen measurement bank once recognition of recall items makes your in-app accuracy hard to trust — the two-Q-bank rule, for measurement not duplication.
- Switch if recall phrasing is drifting from current UK guidance or a domain stays weak despite work.
- Stop adding new questions when every exit criterion is met; review misses and rest.
Frequently asked questions
Is MediWord enough for UKMLA on its own? As an inexpensive exposure-and-mock tool it can support the applied-knowledge component, but a recall-based bank is best treated as one input rather than a sole source, because its phrasing and blueprint balance are less controlled than official material and it does not reproduce the CPSA. Confirm sufficiency with unseen, timed, domain-level performance and official calibration, not with completion.
Which UKMLA component does MediWord not reproduce well? It does not reproduce the clinical and professional skills assessment, the practical component; and because its items are recall-derived, it also reproduces official phrasing and standard less faithfully than the GMC's own materials, so calibrate pacing and wording against those.
How many MediWord questions should I complete per day for UKMLA? A sustainable one to two coded, reviewed blocks a day (roughly 40 to 80 items) beats a high-volume skim, because with a recall bank the review — coding, transfer testing, guideline checks — is where the learning happens. Scale to your runway, and do not treat clearing the bank as the target.
When should I stop using MediWord and move to mixed mocks? Move to timed mixed mocks once domain floors are met and unseen accuracy is stable, typically the final two to three weeks, regardless of how much of the recall bank remains. Use MediWord's mock papers to rehearse pacing, and calibrate against official UKMLA material before the sitting.
How should I combine MediWord with iatroX without duplicating practice? Keep the jobs separate: MediWord for exposure and mock rehearsal, iatroX for unseen transfer items that confirm understanding, and never answer the same item on both. The two-Q-bank rule protects your measurement bank from recognition effects, which matters most with a recall-based source where familiarity accrues fast.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; MediWord's counts, tiers and features are vendor-reported as at that date and should be re-verified on the product page before purchase. Disclosure: iatroX operates a competing UKMLA question bank; here its role is confined to unseen transfer measurement, which a recall bank cannot perform on itself. No proprietary-algorithm claims are made about either platform. Corrections are welcome via the feedback route on iatrox.com.
References: GMC — Medical Licensing Assessment content map and official UKMLA resources, gmc-uk.org; MediWord, mediword.com; iatroX UKMLA bank, https://www.iatrox.com/ukmla; "Your Q-Bank Percentage Is Not Your Exam Score", https://www.iatrox.com/blog/qbank-percentage-not-your-exam-score; the two-Q-bank rule, https://www.iatrox.com/blog/the-two-q-bank-rule-how-to-add-a-second-bank-without-duplicating-questions-or-destroying-calibration; and the iatroX comparison hub, https://www.iatrox.com/compare.