There is no single best RACGP Fellowship resource, and any article that names one is selling something. The right stack depends on which component you are sitting, how many weeks you have, what you can spend, and, above all, what kind of mistakes you make. This article is a decision tree: take a measured baseline, answer four questions, and follow the branch that fits your profile. iatroX is used here as the baseline that tells you which branch you are on, not as the answer to every branch.
Start with the exam, not the product
The RACGP Fellowship pathway has three assessment components, and they reward different resources. The AKT is 150 single-best-answer questions on applied knowledge in Australian general practice. The KFP is 70 independent key-feature cases with multiple-selection responses, testing whether you can name the decisions that change outcomes. The CCE is the clinical and consultation assessment. Because these are genuinely different tasks, the correct move is to choose per component rather than to buy one all-in-one product and hope it covers everything. Confirm current formats, durations and dates on racgp.org.au before you commit.
Segment yourself honestly
Most candidates fit one of six profiles, and the profile, not the marketing, should drive the choice.
- First attempt, adequate runway: you need breadth and calibration more than intensive coaching.
- Retaker: you need a precise diagnosis of the domain or component that failed, then targeted work, not a fresh full course.
- International medical graduate new to Australian practice: your gap is usually jurisdiction, PBS realities and Australian guidance, not raw medicine.
- Weak foundations: you need teaching and structure before volume, or the volume simply confirms the gap.
- Strong knowledge, weak pacing: you need timed full-length practice and pace measurement, not more content.
- Strong on MCQ, weak on the practical or KFP reasoning: you need key-feature and consultation practice, which an SBA bank will not give you.
Define the minimum stack
Before adding anything, define the smallest stack that can pass. For most candidates it is four things: one primary question bank for volume and analytics; the official calibration material, meaning the RACGP sample questions and curriculum, which is finite but authoritative; one teaching or reference source, added only where a domain is genuinely weak; and one modality tool where the component demands it, such as a dedicated KFP bank or supervised CCE practice. Anything beyond this minimum should earn its place by closing a measured gap, not by being new.
Budget bands
Verify every price on the day you buy, because vendors change them and this article does not restate them.
- Free or low cost: the RACGP sample questions and curriculum, plus a free or low-cost bank for volume. iatroX's Australian bank sits here as an affordable measurement layer (price is vendor-reported; check the current figure on iatrox.com).
- One premium resource: add a single paid RACGP bank or a structured course if you need teaching and accountability. Do not buy two premium banks at once; you rarely have time to use both.
- Comprehensive stack: a full course provider for teaching and mocks, one bank for volume, and iatroX as the independent unseen-measurement bank. Justified mainly for retakers or candidates with weak foundations and time to use it all.
Time bands: what to omit, not what to add
The hard skill in exam planning is subtraction. State plainly what you will not do.
- Under four weeks: stop learning new topics. Do timed mixed blocks, review errors the same day, and calibrate against official material. Omit new courses and second banks entirely.
- Four to twelve weeks: alternate topic blocks with weekly timed mocks. Add teaching only for your two weakest domains. Omit anything that does not map to a measured gap.
- More than twelve weeks: build breadth systematically, one or two domains a week, with monthly full-length mocks. This is the only band where a full course is comfortably worth the hours.
Decision matrix: map platforms to jobs, not to rankings
Rather than ranking providers, map each to the single job it does best, then pick the job you need. Summarised deliberately here; the detailed evidence lives in the narrow child audits linked below, so this hub does not duplicate them.
| Job you need done | Resource type that does it | Note |
|---|---|---|
| Structured teaching and accountability | A RACGP course provider (for example GP Institute) | Best for weak foundations and retakers; see the course audit |
| Large-volume AKT and KFP practice | A dedicated RACGP bank or course bank | Volume is not fidelity; check explanation quality |
| Realistic timed mocks | A provider with full-length AKT and KFP mocks | Essential in the final month |
| Independent unseen measurement | iatroX Australian AKT bank | Keep it separate from your primary bank |
| Official calibration | RACGP sample questions and curriculum | The gold standard; finite, so supplement for volume |
| KFP key-feature reasoning | A dedicated KFP resource | An SBA bank will not train this |
| CCE consultation skill | Supervised role-play, a course, real feedback | No MCQ product substitutes for this |
Note the honest gap that recurs across every branch: a platform whose only listed exams are, for example, ACEM, GSSE, RACP and AMC does not cover RACGP at all, so it never enters this matrix. Check that a resource actually holds RACGP content before you compare it on anything else.
Cannibalisation guardrail
This article is the exam-level hub for the single question of which resource to use. It intentionally keeps platform descriptions short and sends you to the child audits for the evidence: the AceTheExam coverage audit and the GP Institute course audit. If you find yourself re-reading long platform descriptions here, follow the link instead; duplicating them would only make both pages worse.
Three worked examples
Priya, first attempt, ten weeks, moderate budget. Baseline shows 74 per cent overall but 55 per cent in cardiology and 58 per cent in women's health. Branch: four to twelve weeks, adequate foundations. Stack: one bank for volume, official samples for calibration, teaching only for the two weak domains, and iatroX weekly for unseen measurement. Exit criterion: first-attempt unseen accuracy above 70 per cent across all domains and stable pacing before she books.
Daniel, retaker, six weeks, failed KFP. Baseline confirms his AKT knowledge is fine but his KFP key-feature discrimination is weak. Branch: retaker with a known component gap. Stack: a dedicated KFP resource, deliberate practice at naming key features, and no new AKT course, which would waste his time. iatroX is used only to confirm his AKT has not slipped. Exit criterion: consistent identification of key features under time before he re-sits.
Mei, international medical graduate, sixteen weeks, comprehensive budget. Baseline shows strong medicine but weak Australian jurisdiction and primary-care context. Branch: IMG with more than twelve weeks. Stack: a structured course for Australian practice and accountability, one bank for volume, and iatroX as the independent measurement bank. Exit criterion: unseen accuracy on jurisdiction-sensitive items matches her accuracy on pure medicine.
Evidence hierarchy
When two sources disagree, rank them. Official RACGP material wins for format and blueprint. Primary Australian guidance, the RACGP guidelines and Therapeutic Guidelines, wins for clinical content. Vendor pages are acceptable only for product facts such as question counts and price, and should be labelled as vendor-reported. Independent testing and your own measured results settle questions of user experience and, crucially, of readiness. Never let a vendor claim outrank the College on format, or your overall percentage outrank your unseen, timed, domain-level accuracy on readiness.
Reading your baseline into a branch
A baseline is only useful if you convert it into a branch, and the conversion is mechanical rather than intuitive. Look first at the shape of the profile, not the average. A flat profile that is uniformly low means weak foundations, so teach before you drill, because more volume will only confirm the gap. A spiky profile, strong in some domains and weak in others, means targeted work: teach the two lowest domains and leave the rest alone. A high but flat profile paired with slow pace puts you in the pacing branch, where you should stop learning content and start timing. And a profile that is strong on pure knowledge but weak on key-feature or consultation-style items points to the practical branch, where an SBA bank will not help you and you should move to KFP and CCE practice. Then overlay weeks-to-exam and budget, which do not change the direction but set how much of the recommended branch you can realistically complete. The baseline chooses the direction; the calendar and the wallet set the scope.
Two traps this tree is built to avoid
The first trap is sunk cost: continuing with a resource because you have already paid for it rather than because it closes a measured gap. If your data says you need key-feature practice, an unfinished AKT course is not the answer, however much it cost. The second trap is novelty: adding a shiny new bank in the final fortnight because your progress feels slow. New volume introduced late rarely helps and usually displaces the timed simulation you actually need at that stage. Let the branch decide the next purchase, not guilt about money spent or restlessness about pace.
Bottom line
There is no universal best RACGP resource, only the resource that fits your component, your weeks, your budget and your errors. Take a measured baseline, read its shape into a branch, define the smallest stack that can close the gap, and add resources only when a measured weakness demands one. Keep your primary bank and your measurement bank separate, so a good score reflects transfer rather than familiarity, and reserve the CCE for real, supervised consultation practice. Follow the branch, not the marketing.
Frequently asked questions
How do I know whether I have covered the full RACGP Fellowship blueprint? You know it not from finishing a bank but from a blueprint-coverage matrix: list every RACGP curriculum domain, and for each record questions attempted, first-attempt accuracy, the date last reviewed and your confidence. Coverage means every domain has enough recent, unseen, correctly answered items, not that a progress bar reads one hundred per cent. Completion is not coverage, as our blueprint-matrix pillar sets out.
Can one question bank be enough for RACGP Fellowship? For the AKT, one good bank plus the official samples can be enough if it genuinely covers the curriculum and you measure by domain. For the Fellowship as a whole it cannot, because the KFP is a different item type and the CCE is a consultation exam; a single SBA bank addresses neither. One bank for volume plus a second, unseen bank for measurement is the safer structure.
What should I measure instead of my overall Q-bank percentage for RACGP Fellowship? Measure first-attempt accuracy on unseen, timed, mixed blocks, broken down by domain; your pace against the exam's pace; your rate of high-confidence errors, which are the dangerous ones; and your retention of previously missed items after a spacing gap. Your headline percentage rises simply from seeing questions again and tells you little about exam-day performance.
When should I stop doing new RACGP Fellowship questions? Stop adding new questions when your first-attempt accuracy on unseen items has plateaued across domains and your remaining errors are careless rather than knowledge gaps. At that point new volume yields little, and your time is better spent on timed full-length mocks, error review and, for the KFP and CCE, the specific reasoning and consultation practice those components demand.
Which RACGP Fellowship resource should I use for my weakest component? Match the resource to the component. For a weak AKT, use a curriculum-mapped SBA bank and teach the weak domains. For a weak KFP, use a dedicated key-feature resource and practise naming decisions under time; SBA volume will not fix it. For the CCE, use supervised consultation practice and real feedback, because no question bank, iatroX included, reproduces a live consultation.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026; any vendor figures cited elsewhere in this series are vendor-reported and change without notice, so verify current prices and counts on the day. Disclosure: iatroX operates a RACGP AKT-style bank, so in this decision tree it appears only as the baseline and unseen-measurement input, never as the universal answer; the decision is yours and your data's. Corrections are welcome via the feedback route on iatrox.com. References: RACGP Applied Knowledge Test and Key Feature Problem exams (racgp.org.au); RACGP Curriculum and Syllabus for Australian General Practice; iatroX, "Your Q-Bank Percentage Is Not Your Exam Score"; iatroX, "Question Bank Completion Is Not Coverage"; and the AceTheExam and GP Institute child audits in this series.