This workflow is for RACGP registrars and internationally trained GPs preparing for the Applied Knowledge Test (AKT) and Key Feature Problem (KFP) who already use, or are weighing up, PassGP and want a repeatable way to convert every miss into retained, verified knowledge. It addresses the two written components only. Its principal limitation is easy to state and easy to forget: an AI tutor can explain a question fluently and still be wrong, out of date, or written for the wrong health system — so the value lives in the verification step, not the chat.
What PassGP offers for the RACGP Fellowship right now
The table below is vendor-reported and last checked on 19 July 2026. Treat every figure as a marketing claim until you confirm it on the current product page.
| Attribute | Vendor-reported detail (verify on the product page) |
|---|---|
| Written coverage | AKT and KFP question banks; marketed as one package spanning AKT, KFP and CCE |
| Question volume | ~2,595 AKT items and ~1,395 KFP-style cases; a combined "4,900+" AKT, KFP and CCE questions |
| Practice assets | 14 mock exams and 500+ flashcards |
| AI / adaptive | "PassGP AI Virtual Tutor" and "adaptive" question banks (vendor terms; no independent adaptivity audit exists) |
| Access period | 6- or 12-month unlimited app and web access |
| Price | From ~A$595 for a crash course; full-package pricing varies — verify current pricing on the product page |
One caveat matters more than the numbers. PassGP markets "CCE questions", but the Clinical Competency Exam is not a multiple-choice component. Per the RACGP's own exam pages, the CCE is a live, online consultation exam delivered over video, built from case discussions and role-player clinical encounters and judged by trained GP examiners. A question-and-explanation bank — PassGP's or anyone else's — is knowledge preparation adjacent to the CCE, not a simulation of it. iatroX is in exactly the same position: it is a written-knowledge and unseen-measurement layer, and it is not a CCE simulator. Keep that boundary in view whenever a product advertises "CCE coverage".
The exam you are actually preparing for
Anchor your study to the official blueprint, not to a vendor's question count. Per racgp.org.au (checked 19 July 2026): the AKT comprises 150 single-best-answer questions set in an Australian general-practice context; the KFP comprises 70 independent multiple-selection questions, each case unrelated to the others, rewarding specific, safe decisions and penalising scattergun answers. Both are currently paper-based with computer-readable answer sheets and run for around four hours each — confirm the delivery mode and timing for your diet, because these details have changed before. The CCE is the separate clinical/consultation stage described above.
The official material is the anchor: the RACGP curriculum and syllabus, the published sample AKT and KFP questions, the candidate guidelines and the examiners' reports. PassGP's volume and "adaptive" claims are third-party. When the two disagree, the RACGP wins. Because the exam is explicitly Australian, every rule you learn should be traceable to Australian guidance — Therapeutic Guidelines, RACGP guidance, or an Australian disease-specific source — not to a generic or overseas explanation.
A prompt taxonomy for missed questions
Six families of prompt cover almost every reason a question is missed. Learn the families, not the individual scripts.
- Mechanism prompts — why the correct answer is correct at a physiological or pharmacological level.
- Discriminating-feature prompts — the single feature that separates the key from its closest distractor.
- Option-elimination prompts — why each wrong option is wrong, not merely why the key is right.
- Guideline-verification prompts — which Australian guideline, which edition, which date.
- Counterfactual prompts — change one stem variable and see whether the answer changes.
- Retrieval-testing prompts — a fresh, unseen question on the same concept, stripped of the original wording.
Twenty-five reusable prompts, grouped by the six errors that cost most marks
These are generic scaffolds. Never paste proprietary question text into a tutor; describe the concept instead.
Error 1 — Pure knowledge gap (you simply did not know the fact).
- "State the mechanism behind the correct answer in three sentences, then name the single highest-yield fact I should have known."
- "Give me the one-line rule this question tests, phrased so I could apply it to any similar case."
- "What first-principles reasoning makes this answer more likely than the others?"
- "List the three most examinable facts about this condition for an Australian GP exam."
Error 2 — Guideline or currency error (outdated or non-Australian rule). 5. "Which Australian guideline governs this, and what is its current recommendation and publication date?" 6. "Has first-line management changed in Australia in the last three years? If so, state old versus new." 7. "Is your answer based on Australian practice or another country's? Name the source." 8. "What would Therapeutic Guidelines say here, and does it differ from your explanation?"
Error 3 — Discriminating-feature error (you could not separate look-alikes). 9. "What single clinical feature discriminates the correct answer from the closest wrong option?" 10. "Give me a two-column table of features favouring diagnosis A versus diagnosis B." 11. "If the stem had said [alternative symptom], which answer would win, and why?" 12. "What is the classic trap feature examiners use to make the wrong option look right here?"
Error 4 — Option-elimination or misread error (you fell for a distractor or misread the task). 13. "Explain in one line each why every option is right or wrong." 14. "Restate exactly what the question is asking — the task, not the diagnosis." 15. "Which option is a true statement that nonetheless does not answer this specific question?" 16. "Which word in the stem should have changed my answer, and did I miss it?"
Error 5 — Management-threshold error (right diagnosis, wrong next step or threshold). 17. "What is the correct next step, and what threshold justifies it over watchful waiting?" 18. "At what point does this move from GP management to referral, and what defines that point?" 19. "What is the single safest first action here, and why is it safer than the alternatives?" 20. "Which investigation changes management, and which are unnecessary?"
Error 6 — KFP specificity or data-gathering error (vague or scattergun answers). 21. "Rewrite my vague answer as the specific, guideline-named action a KFP marker would credit." 22. "For this KFP-style case, list only the responses that would score and explain why the rest would not." 23. "Where would over-ordering or a shotgun list lose marks in a KFP?" 24. "Give the minimum sufficient set of actions that fully answers this key-feature question." 25. "Turn this into a fresh KFP-style prompt on the same key feature so I can retest without the original wording."
The anti-answer-leak rule
Commit before you open the tutor. Write your answer letter and a one-line rationale before any AI explanation is visible; if PassGP shows its explanation beside the answer, cover it. An explanation you read before committing trains recognition, not recall — and both the AKT and KFP test recall under time pressure. Reading first feels productive and quietly inflates your sense of readiness.
The jurisdiction check
Paste this after every tutor answer: "Before I accept this — does this reflect current RACGP and Australian general-practice guidance, or another health system? Name the specific guideline and its date. If you are not certain it is Australian and current, say so." A tutor trained on mixed international data will sometimes hand you a UK, US or generic answer, and on an explicitly Australian exam that is a silent failure mode you will not otherwise catch.
Create a misconception record
For every genuine miss, write four lines: the wrong rule you were using; the corrected rule with a named Australian source and date; a one-sentence transfer question that tests the same concept in a different scenario; and a review date to re-test it cold. If you cannot write the corrected rule with a named source, you have not verified it yet — you have only been told an answer.
Weekly verification sample
Once a week, take a small random sample — say five — of the tutor explanations you relied on and check them independently against a primary Australian source. Log every discrepancy. Two or more material discrepancies in five is a signal to verify harder and trust less. This is the standard AI-tutor audit — grounding, answer leakage, hallucination, retention — run on a schedule rather than by vibe.
A seven-day plan: PassGP for one job, iatroX for transfer
Give each tool exactly one job. PassGP does high-volume, tutor-assisted AKT and KFP practice; iatroX does unseen, timed transfer measurement. This is a division of labour, not a claim about anyone's proprietary algorithm.
- Monday: 40 PassGP AKT items in your weakest domain; commit-then-verify; build misconception records.
- Tuesday: convert Monday's misses into transfer questions; 30 KFP-style items; jurisdiction-check each explanation.
- Wednesday: a fresh, timed, mixed unseen block in iatroX; no tutor until after scoring; record first-attempt accuracy by domain.
- Thursday: target Wednesday's two weakest domains in PassGP using mechanism and discriminating-feature prompts only.
- Friday: KFP specificity day — 25 items marked against minimum-sufficient answers.
- Saturday: weekly verification sample of five explanations versus primary sources; update records.
- Sunday: a short iatroX unseen block to see whether Thursday's targeting moved the number; then rest.
Wednesday and Sunday exist to give you a clean read. Unseen items you have never met in PassGP are the only fair test of whether learning transferred. Your PassGP percentage is not your exam score; treat it as evidence of effort, not attainment.
Decision checklist: continue, supplement, switch or stop
- Continue if first-attempt accuracy on unseen items is rising and your weak domains are narrowing.
- Supplement with iatroX or official material if your PassGP percentage is high but unseen transfer is flat — that gap is recognition masquerading as recall.
- Switch emphasis if verification samples keep surfacing non-Australian or dated tutor answers faster than you can correct them.
- Stop adding new content and move to mixed timed mocks once coverage is broad, unseen accuracy is stable at target, and your remaining errors are pacing rather than knowledge.
Decide on measured gaps, never on novelty or sunk cost.
FAQ
Is PassGP enough for RACGP Fellowship on its own? For the AKT and KFP, PassGP can supply enough question volume and explanation to be a candidate's main written bank, and its vendor-reported figures (checked 19 July 2026) are large. But "enough" also requires broad blueprint coverage, unseen measurement you have not already seen inside PassGP, and separate preparation for the CCE, which is a consultation exam PassGP does not reproduce. Used alone, a single bank tends to over-train recognition of its own item style, so most candidates are better served pairing it with an independent unseen bank and the official RACGP materials.
Which RACGP Fellowship component does PassGP not reproduce well? The CCE. It is a live, examiner-judged consultation exam built from case discussions and role-player clinical encounters, and no question-and-answer product — including PassGP's advertised "CCE questions" and including iatroX — simulates the interactive, timed, communication-assessed reality of it. PassGP's CCE material is best understood as knowledge that supports the CCE, not practice of the CCE itself.
How should I verify PassGP AI answers for RACGP Fellowship? Commit to your own answer and rationale first, then read the tutor. Run the jurisdiction check on every explanation to force it to name an Australian guideline and a date. Weekly, sample five explanations you relied on and check them against a primary Australian source such as Therapeutic Guidelines or the relevant RACGP guidance, and log discrepancies. If material discrepancies recur, downgrade how much you trust the tutor and lean harder on primary sources.
When should I stop using PassGP and move to mixed mocks? When three things are true at once: your blueprint coverage is broad rather than concentrated in comfortable domains, your first-attempt accuracy on unseen items has been stable at your target for a couple of weeks, and your remaining errors are about timing and stamina rather than missing knowledge. At that point additional single-topic practice yields little, and full-length mixed timed mocks — under the real four-hour, penalty-free-marking conditions — become the higher-value activity.
How should I combine PassGP with iatroX without duplicating practice? Give them non-overlapping jobs. PassGP is your learning-and-explanation bank; iatroX is your unseen, timed measurement layer, run on fresh items you have not met elsewhere. Never let an iatroX item become "the answer" to a PassGP item or vice versa — the whole value of the second bank is that it is clean. This is the two-bank rule: one bank teaches, the other measures, and you protect the measurement bank from contamination.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; platform question counts, features, access periods and prices are vendor-reported and change frequently — verify each on the current PassGP product page before relying on it. Disclosure: iatroX operates a competing RACGP AKT and KFP question bank, so this article deliberately confines iatroX to the jobs PassGP does not claim to own — unseen, timed measurement and underlying-knowledge practice — and states plainly that iatroX is not a CCE simulator. Corrections are welcome via the feedback route on iatrox.com.
References: RACGP — Applied Knowledge Test (AKT) and Key Feature Problem (KFP) exams, and Clinical Competency Exam (CCE) pages (racgp.org.au); PassGP RACGP product page (passgp.au); iatroX — Your Q-Bank Percentage Is Not Your Exam Score; iatroX — how to audit an AI medical exam tutor; iatroX — RACGP exams explained: AKT, KFP and CCE; iatroX comparison hub.
Open a missed RACGP item and run the verification prompt in iatroX →