SCA Prep is aimed at ST3 GP trainees who want a purpose-built MRCGP SCA simulator: voice AI-patient roleplays, an unlimited case generator with mark schemes, and a guidelines-trained AI tutor. Of the common SCA tools it is among the closest to the true AI-patient-simulator archetype, and it tests consultation skills MCQ banks cannot — data gathering, safety-netting, spoken management. Its principal limitation is the flip side of "unlimited": generated cases vary in quality, and automated marking still cannot judge the non-verbal rapport a human examiner sees.
What SCA Prep offers for MRCGP SCA right now
Checked against the vendor's public pages on 19 July 2026, all figures vendor-reported. SCA Prep combines three things: voice AI-patient roleplays ("100+ AI Patient Roleplays" with life-like voices drawn from 40+ GP-written cases); an AI case generator that produces effectively unlimited SCA cases with patient notes, a doctor's briefing and a mark scheme (the vendor reports "8,000+ SCA cases generated" across its user base — a usage statistic, not a fixed library); and a guidelines-trained AI tutor and revision bot, text-based, accessible via web and WhatsApp, described as trained on UK clinical guidelines. Marking is returned instantly and split into Data Gathering, Management and Interpersonal Skills, with worked exemplars of "what good looks like." Pricing, at last check, started around £14.95/month for a basic plan (AI tutor plus case generation) and around £34.95/month for the "AI Actors" plan that adds voice roleplay, on a monthly subscription you can cancel anytime; a promotional discount code was live. Verify the current plan structure, roleplay count and prices on the product page.
The exam your practice must answer to
The SCA is twelve simulated twelve-minute consultations, 144 minutes total, sat remotely from ST3, nine diets a year, fee around £1,207. Role-players consult by video or telephone through the RCGP platform; you gather a history, request examination findings verbally, and state management aloud. Crucially, the exam is marked across three examiner-judged domains — Data Gathering and Diagnosis, Clinical Management and Medical Complexity, and Relating to Others. This matters for reading SCA Prep honestly: the platform advertises coverage of "12 RCGP domains," but that is its clinical-content taxonomy (paediatrics, sexual health, mental health, prescribing and so on), not the exam's marking domains, of which there are three. Its own marking split — Data Gathering, Management, Interpersonal — maps neatly onto those three, and that is the distinction to keep straight: three domains judge you; the twelve are content areas.
Format map: which SCA tasks SCA Prep reproduces
SCA Prep is built to reproduce the consultation itself, and its format map is the fullest of the tools in this series.
| SCA task | Reproduced by SCA Prep? | Where, and the caveat |
|---|---|---|
| Spoken consultation with a patient voice | Yes | Voice "AI Actors" roleplay — the feature MCQ banks entirely lack |
| 12-minute timed consultation | Yes | Time the roleplay yourself to the exam clock |
| History and data gathering | Yes | Marked as a distinct category |
| Verbal management planning | Yes | Marked, with exemplars of good practice |
| Examination findings on request | Verify | Confirm how the case briefing surfaces requested findings |
| The three examiner domains | Mapped | Marking split (Data Gathering / Management / Interpersonal) aligns with the three |
| Unseen case exposure at volume | Yes, unlimited | The generator's strength — and its quality-variance risk |
| A full 12-case circuit | Assemble it | Run twelve generated cases back to back to mimic the diet |
| Interpersonal rapport and non-verbal signals | No | Voice can carry tone; a model cannot judge true rapport |
The middle of this table is where SCA Prep earns its place: an MCQ bank can test whether you know the management of a condition, but it cannot make you say it out loud to a patient who interrupts, worries and needs safety-netting. The voice roleplay rehearses the spoken skills the SCA actually assesses, and the mark scheme names them.
Fidelity test: timing, interface, scoring and permitted actions
On timing, the roleplays can be run to the twelve-minute clock, though you must impose it. On interface, SCA Prep does not replicate the RCGP exam platform, so treat it as a rehearsal environment. On scoring, the three-way split is well aligned to the exam's domains and the exemplars are genuinely useful for calibrating what "good" looks like on the observable behaviours — but instant automated marks on management quality and, especially, interpersonal skill are inferences, not examiner verdicts. On permitted actions, the platform keeps management verbal, matching the exam. Two fidelity cautions specific to a generator: first, verify how requested examination findings are handled, since a generated case that invents or withholds findings trains a bad habit; second, because cases are machine-generated at scale, the occasional case will be atypical or subtly off-guideline, so read each case critically rather than assuming it is exam-calibrated.
Case-mix audit: is the spread realistic?
An unlimited generator solves the volume problem and creates a distribution problem. You can produce endless cases, but you choose the topics — and left to preference you will over-generate the areas you enjoy. The "12 RCGP domains" taxonomy is a useful scaffold here: use it as a checklist and generate deliberately across paediatrics, women's and sexual health, mental health, multi-morbidity, undifferentiated presentations, and the ethically and safety-loaded cases that carry equal marks. Because generation is cheap, there is no excuse for a thin domain — but only a conscious audit against the RCGP's published case content prevents an unbalanced set. The blueprint discipline in completion is not coverage is the right habit: quantity is not coverage.
Feedback audit: observable, inferred and generated
SCA Prep's instant marking blends the same three layers, and the split makes them easy to separate. Data Gathering is the most observable category — whether you screened red flags, took a structured history and explored ideas, concerns and expectations is checkable against your transcript, so trust it most. Management is partly observable (did you state a plan?) and partly inferred (was it right?), so treat a weak Management mark as a knowledge prompt rather than a verdict — this is exactly where a thin-looking plan usually means outdated medicine, not poor communication. Interpersonal Skills is the least reliable category for any automarker, because rapport is substantially non-verbal and even a voice model cannot see your manner; use it as a rough signal and rely on human observation for the real judgement. The full method is in our pillar on calibrating automated feedback before you trust the score.
Repetition risk and preserving unseen cases
Paradoxically, an unlimited generator can lower and raise repetition risk at once. It lowers it because you need never run out of fresh cases. It raises it because generated cases share templates and rhythms, so you can develop a false fluency with the generator's style — anticipating its structure rather than a real patient's unpredictability — that will not transfer to the RCGP's human-authored, role-player-delivered cases. Guard against it two ways: vary your generation prompts widely so cases do not converge on a house style, and preserve a set of cases you have not run, ideally GP-written rather than freshly generated, for the final fortnight so your last rehearsals measure cold performance against exam-like material.
What it cannot test
SCA Prep cannot reproduce examiner variability across a real diet, nor the RCGP's exact platform and remote logistics. Its voice roleplay can carry tone but cannot judge genuine rapport, so the Relating to Others dimension is only partly rehearsed and not reliably scored. It cannot replace the real clinical exposure that populates your differential instincts, and a generated case is only as good as the model that made it, so it cannot guarantee every case is guideline-perfect. These are the honest limits of even a well-built simulator; naming them is not a criticism of SCA Prep but a boundary every candidate should hold in mind.
Where iatroX honestly sits
iatroX is not a consultation simulator and does not replace SCA Prep's roleplays — state that plainly. iatroX is the knowledge layer beneath the consultation: unseen multiple-choice measurement and current clinical knowledge. When SCA Prep marks your Management thin, or when you suspect a generated case was off-guideline, iatroX is where you check what the management should have been against current UK practice and then test, on unseen questions, whether that knowledge holds cold. It measures the medicine; SCA Prep rehearses the consultation. Keep the caveat from why your Q-bank percentage is not your exam score in view: neither a roleplay mark nor a question percentage is a pass prediction.
A seven-day pattern for busy ST3 trainees
This plan uses SCA Prep for one defined job — rehearsing and scoring the spoken consultation — and iatroX for the separate job of measuring clinical knowledge on unseen items, with no claim to any proprietary algorithm. Monday: two voice roleplays, timed to twelve minutes, feedback reviewed against Data Gathering behaviours only. Tuesday: generate two cases in a domain your audit flagged as thin; run and mark them. Wednesday: a knowledge session in iatroX's MRCGP SCA area on the Management points marked weak. Thursday: two roleplays plus the AI tutor to close specific guideline gaps. Friday: a peer-observed role-play so a human judges the interpersonal dimension the model cannot. Saturday: a timed mini-circuit of three cases, reviewed by category. Sunday: rest, preserving unseen GP-written cases. SCA Prep rehearses and scores the consultation; iatroX keeps the medicine inside it current.
Continue, supplement, switch or stop
Continue while your spoken structure, timing and management currency improve and you are generating across the full domain spread. Supplement with human-observed practice for the interpersonal and management judgement the automarker cannot fully assess. Switch only for a measurable gap — if you need the human examiner realism a taught course and mocks provide, add those rather than abandoning the simulator. Stop leaning on freshly generated cases in the final fortnight and move to preserved, GP-written cases under strict time. Decide by measured gap, not by the novelty of unlimited generation or by sunk cost.
Three mistakes this audit is designed to stop
First, confusing the "12 domains" with the marking scheme: three domains judge you, and mistaking the content taxonomy for the marking model leads trainees to chase the wrong feedback. Second, treating unlimited generation as unlimited coverage — quantity in your favourite topics is not breadth, and only a domain audit fixes it. Third, over-trusting the instant Interpersonal Skills mark; it is the least observable category, so weight corrections toward the reliably-observed Data Gathering feedback and confirm rapport with a human before you act on it.
Frequently asked questions
Is SCA Prep enough for MRCGP SCA on its own? Its voice roleplays, unlimited generator and guidelines-trained tutor can carry much of the consultation rehearsal, because they test spoken skills MCQ banks cannot — but automated marking cannot reliably judge rapport, generated cases vary in quality, and it does not measure your clinical knowledge on unseen questions, so add human-observed practice and an unseen knowledge check.
Which MRCGP SCA component does SCA Prep not reproduce well? The Relating to Others domain and examiner variability: a voice model can carry tone but cannot judge genuine non-verbal rapport, and no simulator reproduces the calibrated human panel of a real diet, so treat the Interpersonal mark as a rough signal only.
How many unseen SCA Prep cases or stations should I preserve for final MRCGP SCA calibration? Preserve at least twelve unseen cases — ideally GP-written rather than freshly generated — for the final fortnight, so your last timed circuit measures cold performance against exam-like material rather than the generator's familiar house style.
When should I stop using SCA Prep and move to mixed mocks? When your spoken structure, timing and management currency are stable and your domain audit is even — usually the final two to three weeks — shift to full timed twelve-case circuits, ideally with a human observer for the interpersonal judgement the automarker cannot supply.
How should I combine SCA Prep with iatroX without duplicating practice? Use SCA Prep to rehearse and score the spoken consultation and iatroX to keep the clinical management inside it current on unseen questions — the roleplay flags a thin plan, iatroX tells you what the plan should have been. This is the two-source design in the two-Q-bank rule, applied across a simulator and a knowledge bank; our comparison hub helps you place both.
The bottom line for ST3 trainees
The honest verdict on SCA Prep for the SCA: among the most purpose-built of the common simulators, with voice roleplays, an unlimited generator and a guidelines-trained tutor that between them rehearse the spoken consultation skills MCQ banks cannot reach. Its strengths are real; so are its limits. Keep the three marking domains distinct from the twelve content areas, audit your generation for breadth rather than trusting volume, read each generated case critically because quality varies, and treat the instant Interpersonal mark as the least reliable of the three. Calibrate the automated feedback by observability, preserve GP-written cases for the run-in, and let an unseen knowledge source measure whether your management is current. Used that way, SCA Prep rehearses and scores the consultation, and iatroX keeps the medicine inside it honest.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; SCA Prep figures (100+ voice roleplays, 40+ GP-written cases, an unlimited generator with "8,000+ cases generated" as a usage statistic, and pricing from roughly £14.95 to £34.95 per month) are vendor-reported and were checked on that date — confirm current plans, counts and prices on the product page. MRCGP SCA format and its three marking domains are per the RCGP; the "12 domains" SCA Prep advertises is a clinical-content taxonomy, not the exam's marking scheme. Disclosure: iatroX operates a competing clinical-knowledge and question-bank platform and is not a consultation simulator, so this audit names SCA Prep's strengths plainly and confines iatroX's role to the unseen knowledge measurement the product does not claim. Corrections via the feedback route on iatrox.com. References: RCGP Simulated Consultation Assessment pages (rcgp.org.uk); SCA Prep product pages (scaprep.co.uk); related reading: calibrating automated feedback before you trust the score and why your Q-bank percentage is not your exam score.