This workflow is for physician associates using PassMAP for the Physician Associate Registration Assessment (PARA) written exam. One honest caveat up front, because it shapes everything that follows: PassMAP advertises spaced repetition and performance analytics, but we could not confirm a true adaptive-difficulty algorithm at check — so "adaptively" here means you drive the personalisation deliberately, not that a proprietary engine does it for you. The principal risk of any personalised feed is that low-weighted blueprint domains disappear behind a rising overall score.
What PassMAP offers for PARA right now
Last checked 20 July 2026; figures are vendor-reported and change, so verify on passmap.co.uk before you buy.
| Item | What we could confirm (20 July 2026; verify) |
|---|---|
| Type | A PARA-style single-best-answer question bank for physician associates |
| Confirmed features | Spaced repetition, performance tracking by clinical system, timed tests and mock exams |
| PARA components covered | The written KBA (SBA) only — no OSCE |
| "Adaptive" difficulty engine | Not confirmed at check — do not assume it adapts item difficulty; verify on passmap.co.uk |
| Question count / price / access | Not published on the homepage (engagement metrics only, e.g. vendor-reported "150+ students", "25,000+ interactions in 28 days") — verify |
The important distinction: spaced repetition schedules when you see items again; an adaptive-difficulty engine changes which items you see based on ability. Those are different mechanisms, and only the former is clearly advertised. So build your workflow on the confirmed features — spaced repetition and system-level analytics — and treat any adaptivity as unverified. Doing so also keeps you honest: no proprietary-algorithm claims you cannot substantiate.
The PARA exam anchor — mind the low-volume domains
PARA is RCP-delivered and GMC-regulated (physician associates have been GMC-regulated since December 2024). The written KBA is 200 single-best-answer questions in four papers of 50, one hour each. The June 2025 blueprint weights clinical areas from Cardiovascular 8%, Acute and emergency care 7% and Gastrointestinal 7% at the top down to just 2% each for ophthalmology, palliative care, health promotion, healthcare service and research/statistics, with clinical haematology, ENT and medicine of older people at 3%. Those low-weighted domains are exactly what a personalised feed tends to starve, because the algorithm — or your own habit — chases your biggest weaknesses in the biggest topics. Two per cent of 200 questions is still around four marks; on a tight pass mark, four marks matter. The cognitive weighting is Clinical management 40%, the two diagnosis tasks 20% each, Biomedical knowledge 20%. Verify on rcp.ac.uk.
Baseline week: a blueprint-stratified unseen sample
Before you let PassMAP personalise anything, spend the first week establishing a baseline: a small, blueprint-stratified sample of unseen questions covering every domain, including the 2% ones, under timing. This gives you a true starting map — which areas are genuinely weak, where your pace sits, and which cognitive tasks trip you — rather than a picture distorted by whatever the feed decides to show first. Keep this baseline unseen and do not re-sit it; it is a measurement, not a study set.
Set domain floors
The single most important safeguard is a set of domain floors: a minimum number of questions (and a minimum recent-accuracy check) you require in every blueprint area before you consider it covered — regardless of what your overall score is doing. Floors stop the algorithm, or your own preferences, from hiding unattempted syllabus behind a rising headline percentage. Set them in proportion to the blueprint, but never let any domain reach zero: even palliative care and ophthalmology at 2% get a floor.
Error taxonomy
Analytics are only as good as the categories you feed them. Separate your errors into types so PassMAP's system-level tracking becomes a diagnostic rather than a scoreboard.
| Error type | What it signals | Typical fix |
|---|---|---|
| Knowledge gap | You did not know the fact | Short source read + transfer question |
| Misread stem | You knew it but misread | Reading discipline under timing |
| Premature closure | You committed before weighing options | Force a differential before answering |
| Guideline error | Wrong management step | Re-derive from current UK guidance |
| Calculation error | Arithmetic or unit slip | Redo the method; drill calculations |
| Time-pressure error | Right idea, too slow | Retest under a stricter clock |
Review intervals: not every error deserves a repeat
Decide, per error, what it earns. A knowledge gap deserves a short source read and then a fresh transfer question — not an immediate re-sit of the same item. A misread or time-pressure error deserves a retest under timing, not more content. A recurring guideline error deserves a spaced review scheduled days later. Reserve PassMAP's spaced repetition for the items where delayed recall is genuinely the point, and resist the reflex to hammer every miss immediately, which trains recognition of specific questions rather than transferable understanding.
The mixed-block switch
Move from topic-filtered practice to timed random blocks on objective criteria, not on feel:
- Every domain has met its floor, including the low-weighted ones.
- First-attempt accuracy on unseen items is stable across sessions.
- Your pace holds around 60 seconds per question.
When those hold, reduce topic-filtered work and increase mixed, timed, random blocks that resemble the real paper — where no topic is signposted and low-volume domains appear without warning.
Exit criteria
Exit on readiness, not on bank completion:
- Coverage floor met in every domain.
- Stable first-attempt performance on unseen items.
- Pacing sustained at exam speed.
- Retention across spaced intervals.
- Official-material calibration — the RCP sample sat once, unseen, timed — supporting your readiness.
Bank completion alone is not on this list; a finished bank can still hide a starved domain, which is the whole reason for floors.
Seven-day worked example
One week using PassMAP for spaced, analytics-guided KBA practice and iatroX for unseen transfer measurement. No proprietary-algorithm claims — you set the floors and the schedule.
| Day | PassMAP job | iatroX job (unseen measurement) |
|---|---|---|
| Mon | Blueprint-stratified baseline block (all domains) | — |
| Tue | Work the two weakest high-weight domains; error-tag | 15 fresh unseen items on those domains |
| Wed | Deliberate low-volume domain top-up (e.g. palliative, ophthalmology) | — |
| Thu | Timed mixed block; check domain floors | 20 unseen mixed items, timed |
| Fri | Spaced review of scheduled misses only | — |
| Sat | Mock-style timed block; read the system analytics | — |
| Sun | Update floors and error taxonomy; plan next week | 25 unseen mixed items as a readiness read |
Three mistakes this is designed to stop
First, trusting "adaptive" personalisation you have not verified, and letting the feed decide your coverage — set floors instead. Second, letting a rising overall score reassure you while a 2% domain sits at zero attempts. Third, hammering every miss with an immediate repeat, which builds recognition of specific items rather than transferable reasoning. Remember your percentage is not your exam score.
Decision checklist: continue, supplement, switch or stop
| Signal | Action |
|---|---|
| Floors filling, unseen scores rising, analytics improving | Continue |
| A domain persistently under floor despite the feed | Supplement with manual, targeted top-ups |
| Little unseen volume left; items feel familiar | Switch daily practice to a fresh bank |
| All floors met, pace stable, calibration holds | Stop adding items; consolidate and rest |
Decide on measurable gaps — an under-floor domain, an error-type pattern, a pacing figure — not on novelty or sunk cost.
Bottom line
PassMAP is a reasonable option for PARA written practice, with genuinely useful spaced repetition and system-level analytics. Because a true adaptive-difficulty engine is unverified, drive the personalisation yourself: baseline first, set domain floors so low-volume areas are never hidden, review by error type with sensible intervals, and switch to mixed timed blocks on objective criteria. Use iatroX for the unseen items that confirm transfer.
Frequently asked questions
Is PassMAP enough for PARA on its own? For the written KBA it can serve as a primary bank thanks to its spaced repetition, system-level analytics and mock functions, but "enough" is unconfirmed until you verify the total question count on passmap.co.uk, since the homepage showed engagement metrics rather than an inventory. It is not enough for the OSCE, which it does not cover. Most candidates should add a second bank for fresh unseen items and arrange separate OSCE preparation.
Which PARA component does PassMAP not reproduce well? The OSCE — PassMAP is a written single-best-answer bank and does not train the 16-station clinical exam. Within the written exam, its analytics are strong but can under-serve the low-weighted 2–3% domains unless you impose floors, and its "adaptive" positioning was not confirmable at check, so do not rely on an unseen engine to guarantee your coverage. Treat it as KBA practice and manage blueprint breadth yourself.
How many PassMAP questions should I complete per day for PARA? There is no official figure; 40–60 questions daily with tagged review suits most PAs near the exam, but weight the count toward filling domain floors rather than toward the topics the feed keeps surfacing. Let spaced repetition schedule your genuine review items and spend the rest of your quota deliberately on unattempted or low-volume areas. Scale to the weeks you have and protect unseen items for mocks.
When should I stop using PassMAP and move to mixed mocks? Shift the emphasis to mixed mocks once every domain floor is met, first-attempt unseen accuracy is stable, and your pace holds — not when the overall percentage looks good, because that can mask a starved domain. If items are becoming familiar and little unseen volume remains, move daily practice to a fresh bank and keep the RCP official sample for a single calibration sitting. Mixed, timed, random blocks should dominate the final phase.
How should I combine PassMAP with iatroX without duplicating practice? Give each a role: PassMAP is your spaced, analytics-guided KBA practice; iatroX is your fresh, unseen written-measurement bank. Do not re-answer the same items across both to inflate a score — that duplicates practice and corrupts calibration. When PassMAP's analytics flag a weak domain, verify it with new iatroX items on the same content rather than repeating PassMAP questions, following the two-Q-bank rule.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. PassMAP's question count, price and access period were not published on the homepage at check (only vendor-reported engagement metrics were visible), and a true adaptive-difficulty algorithm was not confirmed; verify all of this on passmap.co.uk before purchase. Disclosure: iatroX operates a competing UK question bank and covers the PARA written/SBA layer only — not the OSCE; this article confines iatroX's role to unseen written measurement alongside PassMAP and makes no proprietary-algorithm claims for any product. Corrections via the feedback route on iatrox.com.
References: RCP — PARA information for candidates and KBA blueprint (June 2025), rcp.ac.uk; GMC — PA registration assessment content map, gmc-uk.org; PassMAP, passmap.co.uk (vendor-reported). Internal: Your Q-Bank Percentage Is Not Your Exam Score; the two-Q-bank rule; the iatroX comparison hub.