This workflow is for physician associates preparing for the Physician Associate Registration Assessment (PARA). PA Practice is a low-cost, broad UK MCQ bank that fits the knowledge-based assessment (KBA) well — but its risk is memorisation, and it does not touch the OSCE. The principal limitation to plan around: an MCQ bank trains recognition unless you force transfer, and PARA has a substantial clinical component that no question bank reproduces.
What PA Practice offers for the PARA right now
Everything below is vendor-reported (papractice.com) and last checked 20 July 2026. Confirm each figure on the product page before you buy — counts and prices change.
| Item | What is published (vendor-reported, checked 20 July 2026) |
|---|---|
| Audience | UK physician associate students and candidates |
| Question count | 3,000+ bespoke MCQs (a footer figure of 2,500+ also appears — confirm current number) |
| Content | 350+ conditions from the physician associate clinical matrix; disease summaries with investigations and management |
| Personalisation | Fully customisable quizzes by specialty or specific condition; suggested reading on weak areas |
| Analytics | "Performance over time" graphs; weak-area prompts |
| Access and pricing | Advertised at under £5 per month; confirm access periods on the product page |
| Adaptive AI / OSCE | No proprietary adaptive engine and no OSCE component are described |
| Exam naming | Marketed for the national PA exam and the PA matrix; confirm explicit PARA/KBA alignment on the product page |
The honest read: a genuinely inexpensive, broad recognition bank aligned to the PA clinical matrix, with basic analytics — good for KBA breadth, silent on the OSCE.
The exam you are actually sitting
PARA is delivered by the RCP and is GMC-regulated — the GMC began regulating physician associates in December 2024. It has two components. The knowledge-based assessment is 200 single-best-answer questions, split into four papers of 50 questions, each one hour, sat with breaks between papers. The OSCE is a separate clinical exam of 16 stations — 14 clinical scenarios plus two rest stations — each with two minutes' reading and eight minutes inside the station. Content follows the RCP PARA blueprints and the GMC PA content map; PA Practice, like any MCQ bank, addresses the KBA layer only. Distinguish the official requirements from any third-party claim, and verify current counts on rcp.ac.uk, as blueprints were updated in 2026.
Build a blueprint inventory and protect an unseen pool
Before drilling, list the PARA KBA domains from the RCP blueprint and the GMC content map — the systems-based clinical presentations plus prescribing/therapeutics and professional practice — and set a coverage floor for each. Then ring-fence a pool of questions you will not look at during first-pass revision, so you have genuinely unseen items for timed mixed assessment later. Without a protected pool, every "mock" late on is really a memory test of items you have already seen.
First pass: mostly mixed, topic blocks only where foundations are weak
Start mixed. Topic-filtered blocks are useful only where a domain is genuinely weak; used by default they let the topic label cue the answer, which flatters your accuracy and hides the recognition problem. If your baseline shows a specific weak system, take a short topic block there, then return to mixed practice so the exam's unsignposted structure is what drives recall.
Review each miss with one code and one action
Do not transcribe PA Practice's disease summary into notes — that is copying, not learning. Give each miss a single error code and one corrective action.
| Error code | Meaning | One corrective action |
|---|---|---|
| K — knowledge gap | Fact not known | Short source read (NICE/CKS, SmPC/eMC), then a transfer item |
| S — misread stem | Vignette or lead-in misread | Re-answer under time; slow the first read |
| P — premature closure | Committed before weighing options | Force a differential before choosing |
| Rx — prescribing/therapeutics | Wrong drug, dose or monitoring | Check the SmPC/eMC and NICE/CKS; re-drill the class |
| T — time pressure | Right idea, out of time | Re-drill under a stricter clock |
For prescribing items, work from the SmPC/eMC together with NICE/CKS — not a memorised list — so what you learn transfers to any stem.
Use transfer practice instead of repeats
The single most effective anti-memorisation move: when you miss an item, answer a new item testing the same principle before you ever repeat the original. Repeating the original mostly trains recognition of that question; a fresh item on the same concept tests whether you actually learned it. This is the job an unseen second bank does best.
Switch to mixed timed blocks when floors are met
Move to predominantly timed mixed blocks — at PARA's roughly one-minute-per-item KBA pace — once every domain sits above its coverage floor, even if you have not finished the whole bank. Completion is not the trigger; coverage and stable unseen accuracy are.
Exit criteria: what "ready" means for the KBA
Stop adding new questions when you can show coverage above floor across every domain, stable first-attempt accuracy on unseen timed blocks, comfortable pacing across the four papers, retention of earlier misses on spaced re-test, and a clean run at any official RCP sample material as calibration. One hundred per cent bank completion is not on the list — and none of these criteria address the OSCE, which needs separate clinical preparation.
Worked example: a seven-day plan
A physician associate with six weeks to PARA, using PA Practice for one defined job — broad KBA retrieval — and iatroX for unseen transfer practice and measurement. No proprietary-algorithm claims are made about either tool, and neither replaces OSCE preparation.
- Day 1 (baseline): 40-item mixed unseen block, timed; record domain scores; set floors.
- Day 2: PA Practice topic block on the weakest domain; code every miss; short source reads for K and Rx errors.
- Day 3: answer a fresh iatroX item for each Day-2 miss rather than repeating the originals; log which principles still fail.
- Day 4: timed 40-item mixed block; review by error code; spaced re-test of Day-2 items.
- Day 5: second weak domain via a short topic block, then back to mixed; prescribing items checked against SmPC/eMC and NICE/CKS.
- Day 6: full-length timed block at KBA pace across a spread of domains; read analytics as context, not as a target.
- Day 7: an unseen mixed block in iatroX as the readiness signal; re-check floors; book separate OSCE practice for the coming week.
Three mistakes this workflow is designed to stop
First, mistaking a rising bank percentage for readiness — repeats and topic cueing inflate it, and your Q-bank percentage is not your exam score. Second, revising the whole exam through an MCQ bank and arriving under-prepared for the OSCE, which is 16 stations of applied clinical skill that no bank reproduces. Third, copying disease summaries into notes and calling it learning; the test is whether you can apply the principle to an unseen stem, not whether you can reproduce the explanation.
Decision checklist: continue, supplement, switch or stop
- Continue PA Practice if domain floors are filling, unseen first-attempt accuracy is rising, and its explanations are teaching you.
- Supplement with an unseen-measurement layer (iatroX) if you cannot separate real learning from item memorisation — and separately, always add dedicated OSCE preparation, which is a different modality; see our note on calibrating scored clinical-skills feedback.
- Switch primary bank only for a measurable gap — a domain barely covered, or explanations that are not moving your accuracy — not for novelty.
- Stop buying more question access once your KBA exit criteria are met; redirect that time to the OSCE.
Bottom line
PA Practice is an inexpensive, broad KBA bank that fits the PARA knowledge paper well, provided you use it for retrieval rather than memorisation: set domain floors, code and act on misses once, use transfer questions instead of repeats, and measure readiness on unseen timed blocks — including a second bank such as iatroX. Just remember the boundary — the PARA OSCE is a separate 16-station clinical exam, and no MCQ bank, iatroX included, prepares you for it. Compare options on the iatroX comparison hub.
FAQ
Is PA Practice enough for Physician Associate Registration Assessment on its own? For the knowledge-based assessment it can be a sufficient primary bank, because it is inexpensive, broad and aligned to the PA clinical matrix — but it is not enough for the whole exam. PARA also has a 16-station OSCE that PA Practice does not address, so "enough" only ever applies to the KBA layer, and even there you should add the official RCP material for calibration and a small unseen set to confirm readiness.
Which Physician Associate Registration Assessment component does PA Practice not reproduce well? The OSCE. PA Practice is a single-best-answer knowledge bank with no clinical-station or communication component, so it does nothing for the 14 clinical scenarios (plus two rest stations) that make up the PARA OSCE. It also, being a fixed bank, reproduces sustained unseen exam conditions less well the more of it you have seen — which is why a protected unseen pool and a second bank matter near the exam.
How many PA Practice questions should I complete per day for Physician Associate Registration Assessment? There is no official figure, so treat this as guidance: most candidates working alongside clinical placements manage 30–50 fully reviewed items a day, where each miss is coded and given one corrective action rather than skimmed. Given the 200-item KBA and your weeks remaining, quality of review beats raw count, and the daily target should flex with the domains still below floor.
When should I stop using PA Practice and move to mixed mocks? Move to predominantly timed, mixed, full-length blocks once every domain is above your coverage floor, your unseen first-attempt accuracy is stable, and your pacing is comfortable across the four KBA papers. Keep a little topic-filtered practice only for domains still weak. The trigger is measurable coverage and stable performance, not finishing the bank — and in parallel you should already be doing OSCE practice.
How should I combine PA Practice with iatroX without duplicating practice? Give each bank one job. Use PA Practice for broad, low-cost KBA retrieval, and use the iatroX UK question bank, which lists PARA on quiz-landing, purely as the unseen transfer-and-measurement layer: when you miss a PA Practice item, answer a fresh iatroX item on the same principle instead of repeating the original, and keep a protected pool of unseen iatroX items for timed readiness checks. That prevents overlap and preserves honest calibration — and neither tool substitutes for OSCE preparation.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. Product figures (question counts, prices, features) are vendor-reported from the PA Practice product pages on that date and change without notice — verify the current count and price before purchase. Disclosure: iatroX operates a competing UK question bank; its role in this article is confined to unseen transfer practice and readiness measurement for the KBA — jobs PA Practice does not claim to do for you — and it explicitly does not reproduce the PARA OSCE; no proprietary adaptive algorithm is claimed for either product. Corrections are welcome via the feedback route on iatrox.com.
References: RCP PARA information for candidates and PARA KBA/OSCE blueprints (rcp.ac.uk); GMC physician associate content map and regulation pages (gmc-uk.org); PA Practice product page (papractice.com); iatroX, "Your Q-Bank Percentage Is Not Your Exam Score"; iatroX, "Calibrating AI-graded SAQs and OSCEs"; iatroX, "The two-Q-bank rule".