Before you can honestly say you have covered the PSA, you need evidence — not a finished question bank. This is a checklist, not a timetable. It sets out the minimum you should be able to show across the eight question types and seven clinical domains: unseen timed accuracy, safe prescription construction, fast navigation of the exam's online formulary, sound calculation under pressure, and confirmation against current UK guidance. Work the checklist; stop when the boxes are evidenced, not when the bank runs out.
Current exam snapshot
The PSA is 60 items in two hours, computer-based, sat by final-year UK medical students and some foundation doctors, delivered by the British Pharmacological Society with the Medical Schools Council, with several sittings a year and a result valid for two years. Standard-setting is by modified Angoff. The eight question types are not equally weighted — this matters enormously for where your marks come from:
| Question type | Items | Marks each | Marks |
|---|---|---|---|
| Prescribing | 8 | 10 | 80 |
| Prescription Review | 8 | 4 | 32 |
| Planning Management | 8 | 2 | 16 |
| Calculation Skills | 8 | 2 | 16 |
| Adverse Drug Reactions | 8 | 2 | 16 |
| Drug Monitoring | 8 | 2 | 16 |
| Communicating Information | 6 | 2 | 12 |
| Data Interpretation | 6 | 2 | 12 |
| Total | 60 | — | 200 |
The seven clinical domains are Medicine, Surgery, Elderly Care, Paediatrics, Psychiatry, Obstetrics and Gynaecology, and General Practice. Note the arithmetic: the eight Prescribing items carry 80 of the 200 marks — 40% of the paper — so an hour spent shoring up prescribing is worth far more than the item count alone suggests. Candidates are given online access to the exam's built-in formulary before and during the assessment; navigating that reference at speed is itself an examined skill. For medicines facts in your revision, work from the Summary of Product Characteristics via the eMC, and from NICE and CKS for management — and confirm the current format on the official PSA site, because sitting arrangements are updated each cycle.
The blueprint coverage table (build this first)
Copy this and fill it in from your own data; a type is not covered until every column is evidenced.
| Question type | Official weight (marks) | Items attempted (unseen) | First-attempt accuracy | Last reviewed | Confidence (H/M/L) |
|---|---|---|---|---|---|
| Prescribing | 80 | ||||
| Prescription Review | 32 | ||||
| Planning Management | 16 | ||||
| Calculation Skills | 16 | ||||
| Adverse Drug Reactions | 16 | ||||
| Drug Monitoring | 16 | ||||
| Communicating Information | 12 | ||||
| Data Interpretation | 12 |
The point of the marks column is to stop you optimising the wrong thing: pushing Data Interpretation from 80% to 90% adds at most a mark or two; the same effort on Prescribing can move eight to ten marks.
Ten domain-level blind spots that self-selected practice tends to hide
Require exam-specific clinician review before you trust your coverage; these are the areas candidates most often leave weak because they rarely choose them:
- Anticoagulation initiation, bridging and reversal, and interacting drugs.
- Insulin regimens and variable-rate intravenous insulin.
- Paediatric weight-based dosing and safe volumes.
- Prescribing in renal and hepatic impairment (dose adjustment and avoid-lists).
- Prescribing in pregnancy and breastfeeding.
- Analgesia, including opioid conversion and step decisions.
- Antimicrobial choice, allergy de-labelling and duration.
- High-risk monitoring (lithium, methotrexate, aminoglycosides, digoxin) — confirm targets in the SmPC.
- Fluid and electrolyte prescribing, including replacement and maintenance.
- Deprescribing and the elderly-care polypharmacy review.
Each of these maps to real marks across the Prescribing, Prescription Review, Drug Monitoring and Adverse Drug Reaction types.
These blind spots share a mechanism: candidates avoid them precisely because they are uncomfortable, so self-selected practice systematically under-samples them and a bank percentage climbs while the real risk sits untouched. That is why exam-specific clinician review matters before you declare yourself ready — a reviewer who works in the relevant setting will spot the missing anticoagulation-reversal scenario or the absent variable-rate insulin item that your own selection quietly skipped. Build the list into your coverage table as named rows, not as a vague intention to get to them later.
The exam's reference tool: practise the lookup, not just the recall
Candidates are given online access to the exam's built-in formulary before and during the assessment, which changes how you should revise. The skill under test is not memorising every dose; it is knowing which facts you can safely look up and doing so fast enough that the lookup costs seconds, not minutes. Practise finding a licensed dose, an interaction, a monitoring parameter and a contraindication at speed, and build the reflex of checking rather than guessing on anything you are less than sure of. For your own study facts, work from the Summary of Product Characteristics via the eMC so that what you rehearse matches a primary source, and reserve memorisation for the high-frequency, time-critical facts you cannot afford to look up mid-question.
Format checklist
Tick these off as deliberate, timed practice — not incidental exposure:
- Fast, accurate navigation of the exam's online formulary: find a dose, an interaction or a monitoring parameter in seconds. The reference is provided in the exam, but only speed makes it useful.
- Complete, safe prescription construction: drug, dose, route, frequency, formulation, start and stop, signature and date — in the exam's writing format.
- Correct dose, route and frequency selection under time, including maximum doses and adjustments.
- Whole-paper time management: 60 items in two hours, with the heavily weighted Prescribing items given the time they deserve.
Interpretation checklist
- Calculations without a slip: infusion rates, unit conversions, paediatric mg/kg, and renal-function estimates.
- Data interpretation: laboratory trends (U&Es, LFTs, INR, drug levels) driving a prescribing decision.
- Adverse drug reaction recognition and identification of the causative agent.
- Drug monitoring: which test, when, and the action threshold — targets confirmed against the SmPC.
- Communicating information accurately to a patient or colleague where the item requires it.
Prescription construction: the marks candidates leave behind
Prescribing items carry the most marks and the most partial-credit traps, so construct every prescription in full even when the drug choice feels obvious. A complete, safe prescription specifies the drug (generic name, correctly spelled), the dose with the right units, the route, the frequency or timing, the formulation where it matters, the maximum dose or duration where relevant, and the date with a signature — all checked against the patient's allergy status, renal and hepatic function, and current medicines for interactions. The commonest avoidable losses are an omitted allergy check, a dose that is right for a fit adult but wrong in renal impairment, a missing frequency, and a prescription that is safe in isolation but interacts with something already on the chart. Practise writing the whole thing, not just choosing the agent.
Calculations: the slips that cost most
Calculation items are low-mark individually but cluster into an easy, avoidable loss under time. Drill the specific slips rather than arithmetic in the abstract: misplaced decimal points, unit conversions (micrograms to milligrams, millimoles), per-kilogram paediatric doses confused with total doses, infusion rate versus total volume, and rounding that turns a safe dose into an unsafe one. Do these under the clock, because the error rate that matters is the one you make when hurried, not the one you make when checking calmly. A reliable method — write the units at every step, sanity-check the order of magnitude, and confirm the final figure against a licensed maximum in the SmPC — turns calculations from a liability into free marks.
Recency checklist
Identify the guidance-sensitive topics (anticoagulation, antimicrobials, diabetes, venous-thromboembolism prophylaxis, contraception and common interactions) and, for each source you rely on, record the date and the jurisdiction. UK practice is the reference frame; confirm against NICE, CKS, SIGN and the SmPC via the eMC, and re-date anything you revised more than a few months ago.
Performance checklist
- Unseen, timed, mixed blocks — not just the topic sets you chose.
- Speed: are you finishing with time proportionate to the marks, Prescribing first?
- High-confidence errors: track items you were sure of and got wrong — the safety-critical category.
- Retention: spaced re-tests holding.
- Official-material calibration: performance on the official practice papers, treated as the gold-standard for style and standard.
Reading your coverage table: marks, not averages
Read the completed table by marks, not by a flat average across the eight types. Because Prescribing and Prescription Review together carry 112 of the 200 marks, a candidate at 60% on those two and 90% on everything else is in more danger than the headline average suggests, while the reverse profile is comfortable. Weight your remaining revision by the marks column: an hour that lifts Prescribing accuracy is worth several hours spent perfecting Data Interpretation. Read the confidence column too — a type marked high-confidence but only medium-accuracy is a calibration problem, and high-confidence errors on prescribing are exactly the safety-critical failures the exam is built to catch.
Three mistakes this checklist is designed to stop
The first is equating a finished bank with a covered blueprint; completion is exposure, not evidence of unseen, timed accuracy across every type. The second is optimising the wrong types — pushing an already-strong low-mark type higher while the heavily weighted Prescribing items stay soft, because a single percentage hides where the marks actually sit. The third is revising to a stale guidance picture; anticoagulation, antimicrobial and diabetes recommendations change, so date every guidance-sensitive source against current UK guidance before you rely on it. The checklist exists to convert a vague feeling of readiness into a marks-weighted, dated, unseen-tested record you can actually defend.
Stop / continue decision tree
- Continue new questions if: any type is below floor on unseen blocks, or blind spots are unaddressed.
- Consolidate (stop new, review) if: coverage is broad but retention is slipping or high-confidence errors persist.
- Simulate if: coverage and retention are solid — move to full timed papers and the official practice material.
- Seek teaching if: a specific domain (insulin, anticoagulation) resists self-study.
- Rest if: all boxes are evidenced and performance is stable — more questions now buy anxiety relief, not marks.
The one-page checklist (copy this)
- Blueprint coverage table complete, weighted by marks.
- Ten common blind spots reviewed with a clinician.
- Format skills (formulary navigation, prescription construction, timing) drilled under the clock.
- Interpretation and calculation error rate low and stable.
- Guidance-sensitive topics dated against UK sources.
- Performance measured on unseen, timed mixed blocks.
- At least one official practice paper sat as final calibration.
Run the standard-mode syllabus audit first to expose the gaps, then switch to adaptive practice for the confirmed weak types.
Worked example (invented data)
A final-year student's coverage table shows Prescribing at 62% first-attempt on unseen items (medium confidence), Calculation Skills at 55% (low), and everything else at or above 75%. Weighted by marks, that is a problem: Prescribing and Calculation are where the paper is won or lost by the unprepared. The plan writes itself — two weeks of timed Prescribing and Calculation blocks with error-coding, blind spots 1–5 reviewed with a clinician, then unseen re-measurement and one official practice paper. Exit criterion: Prescribing and Calculation at or above overall accuracy on unseen timed blocks — not "bank finished".
A second, subtler profile: a student at 85% overall but with three high-confidence prescribing errors on unseen items. The average looks safe; the pattern is not, because confident prescribing errors are the safety-critical category and standard-setting weights that failure heavily. The action is not more questions but targeted review of those three items with a clinician, a check of the underlying rule in the SmPC, and re-measurement on fresh prescribing items until the high-confidence error rate falls to near zero.
FAQ
How do I know whether I have covered the full Prescribing Safety Assessment blueprint? You have covered it when your blueprint table is evidenced in every column — each of the eight question types attempted on unseen items, first-attempt accuracy at or above target, recency dated and confidence justified — with extra weight on Prescribing, because it carries 40% of the marks. Completion of a bank is not coverage; a filled, marks-weighted table is. The completion-is-not-coverage pillar sets out how to build the matrix for any exam.
Can one question bank be enough for Prescribing Safety Assessment? It can be enough for content breadth, but not for the two things that decide the PSA: safe prescription construction in the exam's format with fast use of the in-exam reference, and an honest readiness signal from unseen items. Pair one bank with the official practice papers (for standard and style) and an unseen measurement layer, and it can suffice; used alone and re-scored, it will overstate your readiness.
What should I measure instead of my overall Q-bank percentage for Prescribing Safety Assessment? Measure marks-weighted, unseen performance by question type — especially Prescribing and Prescription Review, which together carry more than half the paper — plus your calculation error rate, your high-confidence error rate, and your speed on the heavily weighted items. An overall percentage treats a Prescribing item and a Data Interpretation item as equal; the exam does not. See the linked article on why the percentage misleads.
When should I stop doing new Prescribing Safety Assessment questions? Stop when every question type is at floor on unseen timed blocks, your ten common blind spots are reviewed, calculations are reliable under time, and your high-confidence error rate is low. Past that point, sit the official practice papers as final calibration and rest. New questions beyond this buy reassurance, not marks.
Which Prescribing Safety Assessment resource should I use for my weakest component? Match the tool to the failure. If prescription construction is weak, drill full prescriptions in a bank that marks them, and practise the in-exam reference against the clock. If calculations are weak, use timed calculation sets until the error rate falls. If a clinical domain (insulin, anticoagulation, paediatric dosing) resists, use focused teaching or a reference source, confirming specifics in the SmPC — then re-measure that component on unseen items.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Exam structure is taken from the official PSA blueprint (July 2024) and the delivering bodies' current guidance; the paper is 60 items marked out of 200, weighted heavily toward Prescribing. Confirm the current format and sitting arrangements on the official PSA site, as they are updated each cycle. Any product figures are vendor-reported and should be verified on the product page. Medicines facts should be confirmed against the Summary of Product Characteristics via the eMC and current NICE, CKS and SIGN guidance. Disclosure: iatroX operates a PSA question bank; this checklist is exam-agnostic in method and confines iatroX's role to unseen measurement and gap confirmation. Corrections via the feedback route on iatrox.com.
References: Prescribing Safety Assessment official site and blueprint (prescribingsafetyassessment.ac.uk) and the British Pharmacological Society (bps.ac.uk); official PSA practice papers via the BPS Assessment learner portal; NICE, CKS and SIGN, and the SmPC via the eMC; iatroX PSA bank, comparison hub, the completion-is-not-coverage blueprint-matrix pillar and "Your Q-Bank Percentage Is Not Your Exam Score".