If you are a physician associate preparing for the Physician Associate Registration Assessment (PARA) and you are using RCP Revise, this workflow is for you. It addresses the written Knowledge-Based Assessment (KBA) only — not the OSCE — and its principal limitation is honest: RCP Revise is a revision product authored by the college that runs the exam, not the exam papers themselves. Treated carelessly it becomes recognition practice; treated as calibration material it tells you how the real paper is built.
What RCP Revise offers for PARA right now
The current-state summary below was last checked on 20 July 2026. Figures that a vendor publishes change frequently, so verify them on the product page before you buy.
| Item | What we could confirm (20 July 2026) |
|---|---|
| Producer | The RCP (Royal College of Physicians), the body that delivers PARA — sold through store.rcplondon.ac.uk |
| Format | Single-best-answer written questions mapped to the PARA KBA blueprint |
| PARA components covered | The written KBA (knowledge) layer only — not the OSCE |
| Question count | Not published in a form we could confirm at check — verify the current count on the RCP store |
| Price / access period | Not confirmed at check — verify on the RCP store on the day you buy |
| AI / adaptive features | None advertised — verify |
The single most useful thing to understand about RCP Revise is its provenance. Because the RCP writes the questions and also writes the exam, the item style tracks the official blueprint more closely than a generic UK bank adapted from PLAB or foundation material. That is a genuine strength for calibration. It is not, however, a licence to grind the bank until you recognise the answers — that destroys precisely the signal you are paying for.
The PARA exam anchor: what the blueprint actually specifies
PARA is delivered by the RCP and sits under GMC regulation; the GMC has regulated physician associates since December 2024, and PARA is the assessment newly qualified PAs pass to join the register. It has two components. The written KBA comprises 200 single-best-answer questions split into four papers of 50, one hour per paper (roughly four hours plus breaks). The OSCE, held at The Spine in Liverpool, runs 16 stations (14 scored clinical stations and two rest stations), eight minutes per station with two minutes of reading.
The KBA blueprint (June 2025) weights clinical areas as follows: Cardiovascular 8%; Acute and emergency care 7%; Gastrointestinal 7%; Surgery, Respiratory, Mental health, Neurosciences and Endocrine/metabolic 6% each; Child and adolescent health, Infection, Musculoskeletal, Obstetrics and gynaecology and Renal/urology 5% each; Dermatology 4%; Clinical haematology, ENT and Medicine of older people 3% each; and Healthcare service, Health promotion, Ophthalmology, Palliative care and Research/statistics 2% each. A second axis weights the cognitive task: Clinical management 40%, Diagnosis by choice of investigations 20%, Diagnosis by interpretation 20%, and Biomedical knowledge 20%. Confirm the current blueprint and marking scheme on rcp.ac.uk, because a college can revise weightings between diets.
What counts as genuinely official — and what does not
Candidates routinely conflate "produced by the RCP" with "the exam". They are not the same, and the distinction changes how you should use each item. The genuinely official, finite calibration material is: the KBA and OSCE blueprints, the released sample questions, the examination regulations, and the annual/examiner report. That material is the gold standard for format because it is written and sanctioned specifically to represent the live paper. RCP Revise is a commercial revision bank authored by the same college — closer to the blueprint than most third-party banks, but still a preparation product with a large, refreshable item pool. Third-party banks (PLABable for PAs, PassMAP, Matrix Education, iatroX and others) sit a further step out: useful for volume and unseen measurement, but not authored by the exam body.
Use that hierarchy deliberately. The scarce, official sample items are your truest calibration and should be spent once, unseen and timed. RCP Revise is your high-fidelity practice volume. iatroX and other banks supply the fresh, unseen items you need after RCP Revise questions become familiar.
The seven blueprint signals to extract
Most candidates read RCP Revise for content — the fact in the explanation — and miss the structural signals that tell them how the real paper behaves. Extract these seven deliberately, ideally by annotating a sample of 40–50 items rather than eyeballing them.
| Signal | What to look for | Why it matters for PARA |
|---|---|---|
| Stem length | Word count and how much clinical narrative precedes the question | Long stems reward reading discipline and time budgeting at roughly one minute per item |
| Option construction | Are distractors plausible near-misses or obvious fillers? | RCP-style options tend to be close; guessing on partial recognition fails |
| Cognitive level | Is the item recall, interpretation, or management? | The blueprint puts 40% on management and 40% across the two diagnosis tasks — pure recall is the minority |
| Image and data use | Frequency of ECGs, imaging, charts, lab panels | Signals how much data-interpretation practice you need |
| Timing | Time the item honestly under exam conditions | Calibrates whether you can hold ~60 seconds per question across 50-question papers |
| Marking behaviour | Whether every item is worth attempting | Treat all items as answerable and confirm the current marking scheme in the RCP regulations |
| Domain emphasis | Which clinical areas appear most | Should mirror the blueprint: cardiovascular, acute care and GI lead |
Build the side-by-side matrix
Now compare the official sample with whatever your main PARA bank is — RCP Revise if that is your primary, or a third-party bank if RCP Revise is your calibration reference. Do not copy item text; score each source on the signals above. A worked version might look like this.
| Signal | Official RCP sample | Your main Q-bank | Read the gap |
|---|---|---|---|
| Median stem length | Long, full vignette | Shorter, fact-led | Bank under-trains reading stamina |
| Distractor plausibility | High | Medium | Bank flatters your score |
| Management vs recall | Management-heavy | Recall-heavy | Bank tests the wrong cognitive level |
| Data interpretation | Regular | Occasional | Add ECG/imaging/lab practice |
| Pace under timing | ~60s/item | Untimed habit | Introduce timed blocks now |
The matrix is diagnostic, not decorative. Each row that diverges is an instruction for the weeks ahead.
Read the discrepancies diagnostically
A discrepancy is not automatically a fault in the bank; it tells you what the bank is for. If your bank is easier or more factual than the official sample, it is a foundations tool — good early, misleading as a readiness signal. If it is narrower, it needs supplementing across the low-weighted domains. If it is more sharply worded, it may over-penalise you now but will harden your discrimination. Decide, per source: harder, easier, narrower, more factual, or differently worded — and act accordingly, rather than assuming a high percentage anywhere means you are ready.
Preserve the calibration value
Calibration material only works while it is unseen. The moment you have sat an official sample twice, your score reflects memory of those items, not your ability to reason through new ones. So ring-fence the official sample and any RCP Revise questions you designate as calibration: sit them unseen, under timing, once. Everything you rehearse repeatedly should be transfer practice — new items testing the same principle — not the calibration set. This is the same logic behind treating your Q-bank percentage as a diagnostic rather than a predicted score.
From signals to quotas: translate findings into practice
Convert the matrix into numbers. If data interpretation is under-represented in your bank relative to the official sample, set a weekly quota — say 30 fresh data-interpretation items. If your pace signal shows you drifting past 75 seconds per question, every block from now runs on a clock. If a low-weighted domain such as ophthalmology or palliative care is absent from your recent history, schedule a small, deliberate top-up rather than trusting the algorithm or your instincts to surface it. Quotas turn a vague sense of weakness into a finite, checkable plan.
A seven-day worked example
Here is one week for a PA using RCP Revise as the high-fidelity practice bank and iatroX for fresh, unseen transfer measurement. No proprietary-algorithm claims are made — this is a manual schedule you control.
| Day | RCP Revise job | iatroX job (unseen measurement) |
|---|---|---|
| Mon | Sit one 50-item block under timing; annotate the seven signals | — |
| Tue | Review Monday's misses; write one transfer prompt per miss | Answer 15 fresh unseen items on the same principles |
| Wed | Topic top-up in your weakest blueprint area | — |
| Thu | Second timed 50-item block | 20 unseen mixed items, timed, to check transfer |
| Fri | Error-code review; update the matrix | — |
| Sat | Rest official sample (unseen, once, timed) as a calibration point | — |
| Sun | Plan next week's quotas from the calibration result | 25 unseen mixed items as a readiness read |
The official sample is spent once, on Saturday, and never re-sat. RCP Revise supplies volume; iatroX supplies the unseen check that RCP Revise, once familiar, can no longer give.
Three calibration mistakes this is designed to stop
First, re-sitting official sample items until the score climbs — that is measuring memory, not readiness. Second, reading RCP Revise only for content and ignoring the structural signals, so you arrive on the day surprised by long stems and management-weighted questions. Third, trusting a single high percentage as a pass signal; a percentage on a familiar bank is not a pass prediction, and the completion-is-not-coverage problem means a finished bank can still leave blueprint holes.
Decision checklist: continue, supplement, switch or stop
| Signal | Action |
|---|---|
| RCP Revise items still feel unseen and your matrix gaps are closing | Continue |
| Your bank is narrower or more factual than the official sample | Supplement with unseen items and data interpretation |
| You are recognising RCP Revise answers before reading the stem | Switch daily drilling to a fresh bank; keep official sample for calibration |
| Coverage floors met, pacing stable, and a recent unseen sample holds up | Stop adding material; consolidate and rest |
Base every move on a measurable gap — a blueprint hole, a pacing figure, an unseen-score trend — not on novelty or the sunk cost of a bank you have already paid for.
Bottom line
RCP Revise is a strong option for PARA precisely because the college that writes the exam writes the questions, so it carries more blueprint fidelity than an adapted generic bank. Its value is calibration, and calibration is fragile: the instant the material becomes familiar it stops measuring anything. Extract the seven signals, build the matrix, spend the official sample once, and route your daily repetition through fresh, unseen items so your practice keeps testing reasoning rather than recognition.
Frequently asked questions
Is RCP Revise enough for PARA on its own? For the written KBA it is a strong, blueprint-aligned foundation because it is authored by the RCP, but "enough" depends on whether it gives you sufficient unseen volume once you have worked through it, and it does not prepare you for the OSCE at all. Most candidates should pair it with a second bank for fresh items and with dedicated OSCE practice elsewhere. Verify the current question count on the RCP store, because a bank that is smaller than you assume will run out of unseen items before your exam.
Which PARA component does RCP Revise not reproduce well? The OSCE. RCP Revise is a written single-best-answer resource for the knowledge component; it does not train the 16-station OSCE, its communication and procedural domains, or the timed, examiner-observed performance those stations require. Use OSCE-specific preparation — Matrix Education runs OSCE courses, for example — and treat any written bank, including RCP Revise and iatroX, as covering only the knowledge layer.
How many RCP Revise questions should I complete per day for PARA? There is no official figure; a sustainable, evidence-friendly target for most working PAs is 30–50 questions a day with full review, rising to timed 50-item blocks that mirror a single KBA paper as the exam nears. Prioritise reviewing every miss over raw volume — 30 well-reviewed items beat 100 skimmed ones. Adjust to the weeks you have and protect a pool of unseen questions for timed assessment rather than spending them early.
When should I stop using RCP Revise and move to mixed mocks? Switch the emphasis when you start recognising items before reading the stem, when your blueprint coverage floors are met, and when your pace holds around 60 seconds per question. At that point RCP Revise has given you its calibration value; further repetition mostly measures memory. Move to timed, mixed, unseen blocks — using a fresh bank such as iatroX for the unseen items — and keep any remaining official sample for one final calibration sitting.
How should I combine RCP Revise with iatroX without duplicating practice? Give each a distinct job: RCP Revise is your blueprint-faithful practice volume and calibration reference; iatroX is your fresh, unseen measurement bank for the written knowledge layer. Never re-answer the same item across both to inflate a percentage — that duplicates practice and corrupts calibration. When an RCP Revise item exposes a weak principle, test it with a new iatroX item on the same concept, following the logic of 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. Vendor figures — question counts, prices and access periods — are vendor-reported, were not fully confirmable at check for RCP Revise, and change frequently; verify them on the RCP store before purchase. Disclosure: iatroX operates a competing UK question bank and covers the PARA written/SBA knowledge layer only, not the OSCE and not the RCP's own product; this article confines iatroX's role to the unseen-measurement job that a familiar bank can no longer do. Corrections are welcome via the feedback route on iatrox.com.
References: RCP — Physician Associate Registration Assessment (information for candidates, KBA blueprint and OSCE blueprint, June 2025), rcp.ac.uk; GMC — PA registration assessment content map, gmc-uk.org; RCP Revise Question Bank for Physician Associates, store.rcplondon.ac.uk. Internal: Your Q-Bank Percentage Is Not Your Exam Score; the two-Q-bank rule; the iatroX comparison hub.