If you are asking what to study for the Physician Associate Registration Assessment (PARA), the honest answer is that "covered" is not a number of questions and not a timetable — it is a body of evidence. Before you stop adding new questions, you should be able to show that every domain of the GMC content map has been sampled, that your interpretation and recency gaps are closed, and that your accuracy holds up on unseen, timed, mixed blocks. This article turns that into a checklist you can run and, where it flags a gap, points you to the right next activity.
This is the exam-level hub for a single question: how do I verify PARA coverage rather than assume it? The platform-specific audits in this cluster (Matrix Education, PLABable for PAs, PassMAP) link back here rather than repeat the framework.
The minimum evidence before you can say you have covered PARA
Stop treating a completion bar as a finish line. You have covered PARA when you can produce five things: (1) a blueprint coverage table showing every content-map area sampled to a target volume; (2) first-attempt accuracy on unseen items, not review accuracy on items you have already seen; (3) evidence you can interpret images, ECGs, radiographs, laboratory trends and calculations under time; (4) a recency log dating your sources against current UK guidance; and (5) a plan for the components a knowledge bank cannot touch — the OSCE and clinical skills. If any of those five is missing, you have not covered PARA; you have covered the part of it that is comfortable.
Everything below expands each of those into a concrete verification step.
Current exam snapshot: what you are actually being examined on
PARA is delivered by the Royal College of Physicians (RCP) on behalf of the General Medical Council. The GMC began regulating physician associates and anaesthesia associates on 13 December 2024, with a transition period toward mandatory registration; PARA replaced the older PANE from September 2025. To register, candidates must pass both PARA components within the preceding two years, with four attempts allowed at each part (verify current rules on gmc-uk.org and rcp.ac.uk).
The two components (verified against RCP and GMC pages, last checked 20 July 2026):
| Component | Format | Detail |
|---|---|---|
| Knowledge-Based Assessment (KBA) | 200 single-best-answer questions, online | Four one-hour papers of 50 questions each; tests diagnosis, investigation, management and the professional context of practice |
| OSCE | 16 stations (14 scored + 2 rest) | Each scored station 8 minutes (7-minute warning) plus 2 minutes reading; 35 marks/station, 490 total; pass the standard and a minimum of 9 stations |
The authoritative blueprint is the GMC PARA content map, supported by the RCP's own KBA and OSCE blueprints. The content map is organised into four domains — professional values and behaviours; clinical capabilities; areas of clinical practice; and professional knowledge (plus core procedures) — and lists 18 areas of clinical practice: acute and emergency care (including toxicology), cardiovascular, child and adolescent health, clinical haematology, dermatology, ENT, ophthalmology, obstetrics and gynaecology, endocrine and metabolic, gastrointestinal, infection (including sexually transmitted infections), mental health, musculoskeletal, neurosciences, palliative and end-of-life care, renal and urology, respiratory, and surgery. Treat the RCP and GMC pages as the primary sources; treat any third-party "blueprint" as a summary to check against them, not a substitute.
One important honesty point up front: iatroX — and every other single question bank — sits on the knowledge/SBA layer only. A bank helps you verify and close the KBA gap. It does not, and cannot, reproduce the OSCE. Keep that boundary in mind throughout this checklist.
Build a blueprint coverage table
The single most useful artefact in your revision is a table that forces you to look at every area, not just the ones you enjoy. Give it five columns: official area, questions attempted, first-attempt accuracy, last reviewed, and confidence. A worked fragment, using invented data for a fictional candidate, "Amara", eight weeks out:
| Content-map area | Official weight | Questions attempted | First-attempt accuracy | Last reviewed | Confidence (1–5) |
|---|---|---|---|---|---|
| Cardiovascular | Listed (no numeric weight published) | 210 | 74% | 12 days ago | 4 |
| Respiratory | Listed | 180 | 71% | 9 days ago | 4 |
| Acute & emergency care | Listed | 160 | 63% | 6 days ago | 3 |
| Mental health | Listed | 60 | 58% | 21 days ago | 2 |
| Clinical haematology | Listed | 35 | 52% | 26 days ago | 2 |
| Ophthalmology | Listed | 18 | 44% | 34 days ago | 1 |
| Palliative & end-of-life | Listed | 22 | 50% | 30 days ago | 2 |
| Professional values/ethics/law | Listed | 40 | 61% | 19 days ago | 2 |
Note the column that trips candidates up: official weight. The RCP and GMC publish the areas but do not publish a public numeric percentage weight for each. So do not invent one. Instead, treat every listed area as examinable, set your own minimum sampling target per area (for example, no area below 40–50 first-attempt items before you claim coverage), and use the KBA and OSCE blueprints to see which areas are mandatory OSCE specialties. Amara's table shows the pattern almost everyone produces on first inspection: heavy over-sampling of cardiology and respiratory, and thin, stale, low-accuracy rows for ophthalmology, haematology, palliative care and ethics. Those thin rows are the content gap. The finish line is when no row is both low-volume and low-accuracy and low-confidence at once.
The ten domain-level blind spots self-selected practice hides
Left to our own devices, we practise what we can already half-do. These ten areas are the ones most likely to stay hidden behind a flattering overall percentage, because they are low-volume on the day and easy to skip in revision. Each should be sampled deliberately and — before you rely on your own judgement — reviewed by a clinician who knows the PARA standard.
- Ophthalmology — the acute red eye, sudden painless visual loss, and when "refer same day" is the answer.
- ENT — vertigo differentiation, epistaxis management, the unilateral symptom that is a red flag.
- Dermatology — rash recognition from images, the pigmented lesion, and drug eruptions.
- Clinical haematology — anaemia work-up, interpreting a full blood count and film, anticoagulation and bleeding.
- Palliative and end-of-life care — symptom control, anticipatory prescribing, and the ethics of DNACPR and ceilings of care.
- Mental health — structured risk assessment, capacity, and the basics of the Mental Health Act pathway.
- Child and adolescent health — safeguarding thresholds, the unwell child, developmental red flags, weight-based prescribing.
- Obstetrics and gynaecology — early-pregnancy problems, contraception counselling, and gynaecological red flags.
- Endocrine and metabolic emergencies — diabetic emergencies, thyroid storm and myxoedema, adrenal crisis.
- Professional values, ethics, law and evidence — consent and capacity, confidentiality, duty of candour, safeguarding, and the statistics/critical-appraisal items candidates routinely under-practise.
If your coverage table has fewer than about 40 first-attempt items in any of these, you have a documented gap — not a hunch.
Format checklist: the parts a bank cannot verify
Tick these explicitly, because a knowledge bank will silently let you ignore them:
- OSCE and clinical-skills practice is deliberate, not incidental. You have rehearsed history-taking, examination, communication and core procedures against the RCP OSCE domains (Communication 30%, Diagnosis 30%, Clinical management 25%, Therapeutics and procedures 15%), with a live observer. Reading SBAs does not train these.
- You have covered the mandatory OSCE specialties. The OSCE draws at least one station from each of nine mandatory areas (acute care, cardiovascular, child health, gastrointestinal, mental health, musculoskeletal, neurosciences, obstetrics and gynaecology, respiratory). Confirm you have practised stations across all nine.
- Blueprint versioning is current. You have checked the dated version of the GMC content map and the RCP KBA/OSCE blueprints you are revising from, and confirmed it is the version in force for your diet. Blueprints are revised; an old PANE-era summary is a trap.
Interpretation checklist
The KBA is not pure recall; PARA items include images, ECGs and X-rays. Verify you can, under time:
- Images — recognise the common dermatological, ophthalmological and ENT presentations from a picture, not a description.
- ECGs — read rate, rhythm, axis, ischaemia and the handful of "act now" traces (complete heart block, STEMI patterns, hyperkalaemia).
- Radiographs — a systematic chest and abdominal film read, plus the classic fractures.
- Laboratory trends — interpret a changing set of results (rising creatinine, falling sodium, a coagulation screen), not a single snapshot.
- Calculations — drug doses, fluid and renal-function calculations done accurately at speed.
- Ethics and statistics — sensitivity/specificity, predictive values and number-needed-to-treat, plus applied consent, capacity and safeguarding judgements.
For each, the standard is not "I have seen these" but "I got them right, first pass, against the clock."
Recency checklist
PARA is a UK exam, and UK guidance moves. Build a small recency log for guidance-sensitive topics and, for every fact you commit, record the source and the date. High-yield areas to re-date: type 2 diabetes and hypertension management, asthma and COPD, atrial fibrillation and anticoagulation, sepsis recognition, antimicrobial choices, contraception, and safeguarding thresholds. Two rules keep you honest: use current UK sources (NICE, CKS, SIGN, the SmPC/eMC for medicines, and NHS clinical content), and note the jurisdiction — some public-health and prescribing specifics differ across England, Scotland, Wales and Northern Ireland. A correct answer from a 2019 note is a wrong answer if the guideline changed in 2025.
Performance checklist
Coverage without performance is a library, not a readiness signal. Verify:
- Unseen, timed, mixed blocks. Your evidence comes from items you have not seen before, sat at the real pace (200 questions across four one-hour papers is roughly one minute per item), in a mixed block that samples the whole blueprint — not a single-topic set.
- Speed. You finish papers with a margin, and your accuracy does not collapse in the last ten minutes.
- High-confidence errors. You track the items you were sure about and got wrong. These are the dangerous ones; they do not feel like gaps, so they never get revised unless you log them.
- Retention. A topic you "did" three weeks ago still scores when re-tested cold. If it does not, you learned it for a day, not for the exam.
- Official-material calibration. You have compared your bank performance against the RCP's own sample material where available, because that is the closest calibration to the real standard.
Your overall Q-bank percentage is the least useful of these numbers — see Your Q-Bank Percentage Is Not Your Exam Score for why review-inflated percentages mislead.
Stop / continue decision tree
Run the checklist, then act on the measured gap rather than on how tired you are of questions:
- Continue new questions if any content-map area is below your sampling target or below roughly 60% first-attempt accuracy on unseen items. You have a knowledge gap; new volume is the right tool.
- Consolidate (stop adding, start spacing) if coverage is complete and first-attempt accuracy is adequate but retention is patchy — you are forgetting faster than you are learning. Switch to spaced re-testing of misses.
- Simulate if knowledge is solid but you have never sat a full timed, mixed, four-paper block, or have not rehearsed OSCE stations under observation. The gap is performance, not content.
- Seek teaching if a specific area stays low despite volume — that is a comprehension gap a tutor or clinician fixes faster than more questions.
- Rest if your unseen scores are stable and your errors are now random rather than systematic. More questions past this point buys fatigue, not marks.
The one-page checklist (copy this)
- Every one of the 18 content-map areas sampled to my minimum first-attempt target
- No area simultaneously low-volume, low-accuracy and low-confidence
- Each of the ten blind-spot domains sampled and clinician-reviewed
- Images, ECGs, radiographs, lab trends and calculations verified under time
- Ethics, consent, capacity, safeguarding and statistics items practised
- Recency log completed with source + date + jurisdiction for guidance-sensitive topics
- At least one full timed, mixed, unseen four-paper block sat
- High-confidence errors logged and re-tested
- Retention confirmed on cold re-test of older topics
- OSCE stations rehearsed across all nine mandatory specialties with a live observer
- Blueprint version confirmed current for my diet
Worked example
Amara, eight weeks out, has a flattering 72% overall. The table exposes it: cardiology and respiratory are over-sampled at 200-plus items each; ophthalmology (18 items, 44%), haematology (35, 52%) and palliative care (22, 50%) are thin, stale and shaky; ethics sits at 61% but untested under time. Her decision is not "keep going" or "I'm done" — it is targeted. She schedules new-question blocks in the four weak areas, books two OSCE practice sessions covering the mandatory specialties she has not rehearsed, and logs a recency review of diabetes and anticoagulation. Two weeks later her weak rows have volume and her first-attempt accuracy on unseen mixed blocks has risen to the high sixties across the board. Now she consolidates and simulates. That is what "covered" looks like — a closed table, not a big number.
FAQ
How do I know whether I have covered the full Physician Associate Registration Assessment blueprint? You know when your blueprint coverage table shows every one of the 18 content-map areas sampled to a defined minimum of first-attempt, unseen items, with no area left simultaneously low-volume, low-accuracy and low-confidence, and when the ten blind-spot domains have each been reviewed by a clinician who knows the PARA standard. Coverage is a documented table, not a completion percentage; if you cannot produce the table, you have not verified coverage — you have assumed it. Cross-check the areas against the current GMC content map and RCP KBA blueprint, because the version in force can change between diets.
Can one question bank be enough for Physician Associate Registration Assessment? One bank can be enough for the knowledge layer if it covers the whole content map at adequate depth and currency, but no single bank is enough for the whole exam, because none reproduces the OSCE. Even for the KBA, a second, unseen bank is worth adding once your first is mostly seen, so you are measuring on fresh items rather than re-recognising familiar ones — the reasoning is set out in the two-Q-bank rule. Treat one bank as your workhorse and a second, plus dedicated OSCE practice, as your verification and performance layers.
What should I measure instead of my overall Q-bank percentage for Physician Associate Registration Assessment? Measure first-attempt accuracy on unseen items broken down by content-map area, your accuracy under real time on mixed blocks, your rate of high-confidence errors, and your retention on cold re-tests — not the single headline percentage, which is inflated by repeated exposure to items you have already reviewed. The overall figure tells you how comfortable your bank has become with you; the per-area unseen figures tell you where you would still fail. The reasoning behind this distinction is in Your Q-Bank Percentage Is Not Your Exam Score.
When should I stop doing new Physician Associate Registration Assessment questions? Stop adding new questions when your coverage table is closed, your first-attempt accuracy on unseen mixed blocks is stable and adequate across every area, and your errors have become random rather than clustered — at that point additional volume buys fatigue, not marks, and your time is better spent consolidating misses, simulating full timed papers and rehearsing OSCE stations. Until then, any area below your sampling target or below roughly 60% first-attempt accuracy is a live reason to keep going. The decision is driven by the measured gap, never by how many questions are left in the bank.
Which Physician Associate Registration Assessment resource should I use for my weakest component? Match the tool to the deficit: if the weak component is knowledge in a specific domain, use a question bank with strong coverage and explanations in that area; if it is interpretation, use image, ECG and data-heavy question sets; if it is the OSCE or clinical skills, use a structured OSCE course with live observation and feedback, because no MCQ bank trains a station; and if it is performance under time, use full mixed mocks and official sample material. iatroX is well suited to the knowledge and unseen-measurement layer via the PARA bank on the quiz landing page, but it is explicitly not an OSCE simulator, so pair it with dedicated clinical-skills practice for that component.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. Exam-format figures are taken from RCP and GMC pages on the date shown and can change between diets; verify the current KBA question count, OSCE station count and blueprint version on rcp.ac.uk and gmc-uk.org before you rely on them. Disclosure: iatroX operates a UK question bank that competes on the PARA knowledge layer; its role here is confined to the jobs a bank can do — unseen measurement and spaced retrieval of confirmed gaps — and it does not reproduce the OSCE. Corrections are welcome via the feedback route on iatrox.com.
References: RCP PARA information for candidates and KBA/OSCE blueprints (rcp.ac.uk); GMC information about the Physician Associate Registration Assessment and the PARA content map (gmc-uk.org); the modality-gap companion, what MCQ banks cannot prepare you for in PARA; Your Q-Bank Percentage Is Not Your Exam Score; the iatroX comparison hub.