This audit is for UK medical students and foundation doctors weighing the Geeky Medics PSA offering as exam-condition practice. First, an honest reframing: the PSA is a written prescribing exam, not a spoken, branching consultation, so there is no voice, no station circuit and no examiner to simulate — a PSA "simulator" can only mean timed, exam-interface mock papers plus item practice. Geeky Medics markets exactly that (a PSA question bank and three timed mock papers) and does not itself call it a simulator. Its principal limitation is the one every finite mock library shares: repeat it and you measure recognition, not readiness.
Current state (last checked 19 July 2026)
| Attribute | Reported position |
|---|---|
| Product | PSA question bank + three timed mock papers (not marketed as a "simulator") |
| Question count | 600+ questions across all eight PSA question types (vendor-reported) |
| Mock exams | Three timed papers, aligned to the official PSA blueprint, timed under exam conditions with an option to add time (vendor-reported) |
| Task realism | Users can write prescriptions; performance broken down by specialty |
| Extras | "Notebook" of 350+ topic summaries; webinar recordings; advanced analytics; filter by specialty and PSA category |
| AI / adaptive | None advertised on the pages checked |
| Price / access | Not shown on the pages checked — verify on geekymedics.com |
All figures are vendor-reported and subject to change; confirm on the product page.
Exam anchor
The PSA is 60 items in two hours, computer-based, sat by final-year students and some foundation doctors. Its eight question types are unequally weighted — the eight Prescribing items alone carry roughly 40% of the marks (80 of 200) — across seven clinical domains (Medicine, Surgery, Elderly Care, Paediatrics, Psychiatry, Obstetrics and Gynaecology, General Practice), with modified Angoff standard-setting. Candidates are given online access to the exam's built-in formulary before and during the assessment. The official practice papers from the delivering bodies are the calibration gold-standard; a third-party product's "blueprint-aligned" claim is a claim to verify, not an official guarantee. Confirm medicines facts against the SmPC via the eMC and management against NICE and CKS.
Format map: what the product actually reproduces
Because the PSA has no consultation to branch and no station to walk, the honest format map is short and specific:
| PSA task | Reproduced? |
|---|---|
| Prescribing (write a full, safe prescription) | Yes — the bank and mocks let you write prescriptions |
| Prescription Review | Yes |
| Planning Management | Yes |
| Communicating Information | Yes (as written items) |
| Calculation Skills | Yes |
| Adverse Drug Reactions | Yes |
| Drug Monitoring | Yes |
| Data Interpretation | Yes |
| Whole-paper timing (two hours) | Yes — mock papers are timed |
| Branching decisions / time advancement / voice / stations / examiner domains | Not applicable — the PSA has none |
So the fidelity question is not "does it branch?" (there is nothing to branch) but "does the timed paper reproduce the interface, the weighting and the time pressure of the real thing?"
Fidelity test
Compare against current official candidate guidance on four axes: timing (two hours, weighted toward Prescribing), the writing interface (can you construct a full prescription, and is it marked on the components that matter?), the scoring categories (are the eight types scored separately so you can see where marks are lost?), and permitted actions. The one thing no third party can identically reproduce is the exam's own built-in online formulary and its exact interface; you can practise the skill of fast reference-navigation, but the live environment differs, so treat the official practice papers as the definitive interface calibration.
Case-mix audit
Check whether the seven clinical domains and the common-versus-rare, acute-versus-chronic and safety-critical scenarios appear in realistic proportions, and whether the heavily weighted Prescribing items span those domains rather than clustering in general medicine. A finite library of a few hundred items plus three papers will not cover every permutation; use the specialty analytics to find the domains you are quietly avoiding.
Feedback audit
Distinguish three things the feedback may conflate: observable correctness (did you pick the right drug and dose — objective), inferred competence (does one right answer mean you can prescribe it safely in a different patient — an inference), and any model-generated commentary (useful, but to be checked against the SmPC and NICE/CKS, not trusted because it reads fluently). Where scoring is automated, calibrate it against a known-good source before you trust a borderline judgement; the pillar on calibrating automated feedback sets out how.
Repetition risk
Three mock papers and a few hundred items are a finite library. Worked twice, they teach the answers rather than the reasoning — false fluency. Sit each mock once under strict timing, and preserve at least one paper (and a pool of unseen items elsewhere) for genuine final calibration. Rotate in unseen scenarios so your readiness signal reflects transfer, not memory.
What it cannot test
The exact in-exam reference and interface; the real standard-setting; novel drug combinations outside its library; and the local prescribing workflow and formulary constraints you will meet as a foundation doctor. None of this is a criticism of the product — it is the boundary of any off-exam mock. Close it with unseen practice and the official papers.
A practical way to hold that boundary in mind is to list, before you start, the three things you will not learn here — the live reference interface, the official standard, and the comorbidity combinations outside the library — and to plan explicitly where each is covered instead: the official practice papers for the first two, and unseen mixed blocks for the third. Naming the gap in advance is what stops a comfortable mock from being mistaken for a finished preparation, and it keeps your final week focused on the layer the simulator cannot supply rather than on re-sitting familiar papers.
Worked example: a seven-day plan for students and foundation doctors
Days 1–4: one defined job — work the Geeky Medics question bank by weak PSA type (say Prescribing and Calculations), writing full prescriptions and error-coding your misses. Day 5: sit one Geeky Medics timed mock under strict two-hour conditions and score it by type. Day 6: on iatroX, take an unseen, timed mixed PSA block to measure transfer, and open a missed concept in the Socratic Tutor to test the principle on a fresh item (no proprietary-algorithm claims). Day 7: review high-confidence and calculation errors, and preserve the remaining mock unseen. The point is the division of labour: the mock gives interface and timing rehearsal; the unseen block gives the honest readiness signal.
Reading your mock results: three signals
A timed mock is only as useful as the way you read it. Read the by-type breakdown first, weighted by marks: a dip in the heavily weighted Prescribing type matters far more than the same dip in Data Interpretation, and the specialty analytics will show whether the weakness is a drug-knowledge gap or a domain you have avoided. Read your timing profile next: finishing with time to spare but making errors is a knowledge problem, while running out of time on the prescribing items is a pace problem with a different fix. Read your high-confidence errors last and hardest — the items you were sure of and still got wrong are the ones that matter most, because confident prescribing mistakes are exactly what the assessment is built to detect.
Case mix in practice: find the domain you are avoiding
Use the specialty breakdown deliberately rather than passively. Most candidates over-practise general medicine and under-practise psychiatry, obstetrics and paediatric prescribing, because those feel less familiar; the analytics exist to surface exactly that avoidance. After each mock, note which of the seven domains you scored lowest in and whether the heavily weighted prescribing items in that domain were the culprit, then spend your next session there rather than on the domain you already enjoy. A realistic case mix is not the one that feels comfortable — it is the one that matches the proportions of the paper, including the specialties you would rather skip.
Three mistakes this audit is designed to stop
The first is treating "simulator" as if the PSA were an OSCE — there is no branching consultation or examiner to rehearse, so the fidelity that matters is timing, interface and marks-weighting, not voice or stations. The second is exhausting a finite library early: three mock papers re-sat become a memory test, so preserve at least one for clean final calibration. The third is trusting automated feedback because it reads well; scoring and commentary should be calibrated against the SmPC and current NICE and CKS guidance before you act on a borderline judgement. Each mistake is avoided by keeping unseen items in reserve and reading results by marks, not by a single percentage.
Decision checklist
- Continue if: you have unused mocks and your weak types are still improving.
- Supplement if: a domain or calculation skill stays weak — add targeted practice, not a second whole product.
- Switch if: you have exhausted the library and are re-seeing items (recognition, not learning).
- Stop if: coverage is evidenced, unseen timed performance is solid and high-confidence errors are low — sit the official practice paper and rest.
Decide on measured gaps, not on novelty or on money already spent.
FAQ
Is Geeky Medics PSA Simulator enough for Prescribing Safety Assessment on its own? It can carry the interface-and-timing rehearsal and a good share of content, but "enough" needs two additions it does not itself provide: the official PSA practice papers as the standard-setting gold-standard, and unseen measurement so your score reflects transfer rather than recognition of a finite library. As a sole resource that you re-sit, its three mocks and item pool will flatter you near the exam.
Which Prescribing Safety Assessment component does Geeky Medics PSA Simulator not reproduce well? The live exam environment itself — the built-in online formulary and its exact interface, and the real standard-setting — which no third party can replicate identically. It reproduces the item types and the timed-paper pressure well; it cannot reproduce the official reference tool or guarantee the official standard, which is why the official practice papers remain the definitive calibration.
How many unseen Geeky Medics PSA Simulator cases or stations should I preserve for final Prescribing Safety Assessment calibration? Preserve at least one full timed mock paper and, ideally, a further unseen block from a different source, untouched until the final week. With only three mock papers available, using them all up early leaves you nothing clean to calibrate on; one unseen paper sat under strict timing in the last few days is worth more than re-sitting papers you half-remember.
When should I stop using Geeky Medics PSA Simulator and move to mixed mocks? Move to full timed mixed papers once your weak question types are at floor on unseen items — typically the final two weeks. Do not wait to finish the bank; the mocks and the official practice papers are where you rehearse the whole-paper skill. If you are re-seeing items, you have already passed the point of switching.
How should I combine Geeky Medics PSA Simulator with iatroX without duplicating practice? Give each a distinct job: Geeky Medics for prescription-writing rehearsal, timed mock papers and specialty analytics; iatroX for unseen, timed measurement and for opening a missed item in the Socratic Tutor to test the principle on a fresh question. Never re-answer the same items in both — that duplicates exposure and corrupts the signal. The two-Q-bank rule formalises the split.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Geeky Medics figures (600+ questions, three mock papers, Notebook summaries) are vendor-reported as at that date and change; verify on geekymedics.com, and note the product is marketed as a question bank and mock papers rather than a "simulator". Exam facts are from the official PSA blueprint and the delivering bodies; confirm the current format on the official PSA site. Medicines facts should be confirmed against the SmPC via the eMC and current NICE and CKS guidance. Disclosure: iatroX operates a competing PSA bank and Socratic Tutor; this audit confines iatroX's role to unseen measurement and transfer practice — jobs the mock papers do not claim — and is positive about Geeky Medics where its coverage and timed papers warrant. Corrections via the feedback route on iatrox.com.
References: Prescribing Safety Assessment official site and blueprint (prescribingsafetyassessment.ac.uk); Geeky Medics PSA question bank and mock exams (geekymedics.com); the calibrating-automated-feedback and two-Q-bank-rule pillars; iatroX PSA bank, comparison hub and "Your Q-Bank Percentage Is Not Your Exam Score".