Ordinary multiple-choice practice cannot assess most of what the CCFP examination is built to test. A question bank can measure whether you recognise the right answer among options; it cannot measure how you construct management under uncertainty in a Short Answer Management Problem (SAMP), how you communicate in a Simulated Office Oral (SOO), or how you actually run a consultation — increasingly a virtual one. This article names those skills precisely and gives a modality ladder to train each, because the CCFP is moving toward short-menu formats but remains, at heart, a test of family-medicine performance rather than recall.
The CFPC examination format map
The College of Family Physicians of Canada (CFPC) Certification Examination in Family Medicine has two components (verify the current structure on cfpc.ca):
- SAMPs (Short Answer Management Problems): a computer-based paper, four hours, of case-based problems delivered through Prometric. Historically these required written-in answers; from April 2026 the CFPC is transitioning SAMP cases away from write-in responses toward multiple-choice and short-menu formats — vendor/College-reported as up to 25% of cases in MCQ or short-menu in Fall 2026, and all cases in that format by 2027. Verify the timeline and any per-component counts on cfpc.ca; exact case and question counts are not published.
- SOOs (Simulated Office Orals): five 15-minute simulated office interviews (75 minutes total), delivered virtually, assessing clinical communication and the physician–patient relationship against defined competencies.
Two features matter for how you prepare. First, even as SAMP moves to short-menu, it is testing clinical management under uncertainty, not trivia — the format is changing, the construct is not. Second, the SOO and the virtual delivery are performance assessments; no MCQ bank touches them.
Knowledge versus performance: what a correct selected answer proves — and does not
A correct answer on a selected-response item proves one narrow thing: that, presented with the right option, you could recognise it. That is necessary but not sufficient for family practice. It does not prove that you would have generated that option unprompted; that you would have gathered the information that made it correct; that you could sequence management over time; that you could explain it to a frightened patient; or that you could do all of this inside a short consultation, on camera, with an agenda you did not set. The CCFP is designed to test the generation and delivery of care, not just its recognition. When your revision is entirely selected-response, you are training the one sub-skill the exam has always tried to look past.
The skills MCQ practice under-tests, in detail
The SAMP short-menu transition
A short-menu item still asks you to prioritise: to pick the two investigations that change management now, not the six that could be relevant; to name the safety-netting step that matters; to sequence. The risk in the 2026–27 transition is subtle. Candidates who over-drill generic MCQs can learn to recognise single best answers while losing the family-medicine reasoning the SAMP rewards — the breadth of an appropriate differential, prioritisation under uncertainty and management across time rather than a single decisive test. Train the reasoning first, then practise expressing it in the short-menu format; do not let the format train the reasoning out of you. The observable skill is disciplined prioritisation with a defensible rationale, whatever the answer mechanism.
SOO communication
The SOO assesses how you build a relationship and manage a consultation: identifying the patient's agenda and context, responding to cues, sharing decisions, and handling the complexity a real family-practice encounter carries. These are behaviours, scored by an examiner against competencies — not knowledge you can select from a list. You cannot revise for them by reading; you rehearse them, out loud, with someone giving structured feedback against the actual domains.
Virtual consultation performance
Increasingly the SOO is delivered virtually, and virtual consulting is itself a distinct skill. Managing rapport through a screen, handling silences and technology, structuring a remote encounter, safety-netting without a physical examination — these behave differently from an in-person consultation and reward specific practice. Rehearsing only in person leaves a gap exactly where the exam now sits.
One observable behaviour, task, feedback source and exit standard per skill
For each under-tested skill, define four things before you practise: the observable behaviour, the deliberate-practice task, the feedback source and the exit standard. Vague intentions ("get better at SOOs") do not improve performance; a specific behaviour with a real feedback source does.
| Skill | Observable behaviour | Deliberate-practice task | Feedback source | Exit standard |
|---|---|---|---|---|
| SAMP short-menu reasoning | Prioritises the few management steps that change the outcome, with a rationale | Timed short-menu cases; write the rationale, then select | Model answers/marking guides; a peer or tutor | Consistent, defensible prioritisation under time across unfamiliar cases |
| SOO communication | Elicits agenda and context, responds to cues, shares decisions | Role-played 15-minute SOOs against the CFPC domains | An examiner, GP tutor or trained peer | Meets each SOO competency on cases you have not seen |
| Virtual consultation | Builds rapport and structures a safe encounter on screen | Recorded video consultations, reviewed | Self-review of the recording plus a tutor | A safe, structured remote consultation without prompting |
A four-week modality ladder
Train performance the way skills are built — isolated first, then integrated, then under exam conditions. Four weeks is illustrative; scale it to your timeline.
- Week 1 — isolated skill: drill one component at a time. Short-menu prioritisation drills; a single SOO sub-skill (agenda-setting, or shared decisions) rehearsed repeatedly; a solo recorded virtual consultation reviewed for structure.
- Week 2 — coached case: run whole cases with a coach present who stops you, gives feedback against the rubric, and has you re-run the segment. This is where most improvement happens.
- Week 3 — timed integrated case: full-length, timed SOOs and SAMP cases with no coaching mid-case; feedback afterwards only. You are now rehearsing under exam constraints.
- Week 4 — unseen simulation: fresh, unseen cases, timed, scored against the official competencies and treated as a mock. Protect these; do not rehearse them into recognition.
Notice that the MCQ and short-menu knowledge layer sits underneath the whole ladder — it is the entry ticket, not the ladder itself.
When AI feedback helps, when it misleads, and when you need a human
AI feedback is useful for the parts of preparation that are structured and knowledge-adjacent: checking whether your differential is complete, whether your management aligns with a guideline, whether your written rationale names the key step. It is genuinely helpful there, and it scales. It is unreliable, however, for the graded human behaviours at the centre of the SOO: warmth, timing, reading a patient's unspoken agenda, and the judgement of whether a shared decision truly happened. An automated score on a communication performance can look precise and mean little — calibrate any such score against a human examiner's judgement before you trust it. And some things require a human, full stop: a trained examiner or GP tutor judging a SOO against the competencies, and a clinician confirming that your management is safe. Use AI for breadth and structure; use a human for the performance and the safety call.
A worked example: auditing one candidate's CCFP preparation
Consider a final-year family-medicine resident, twelve weeks out, whose question-bank percentage has plateaued in the low 80s and who feels "nearly ready". A modality audit reframes the picture. Her knowledge coverage, mapped against the CFPC priority topics, is genuinely strong — three domains aside, she can generate management, not merely recognise it. But she has done two role-played SOOs in six months, both with the same colleague, both cases she had seen before, and she has never had a consultation scored against the CFPC domains. She has never rehearsed a virtual encounter under time. The audit's conclusion is not "study more medicine"; it is "stop adding MCQs and move the marginal hour to performance". Her next four weeks follow the ladder: week one, isolated agenda-setting and safety-netting drills plus one recorded virtual consultation; week two, three coached SOOs on unfamiliar cases with a GP tutor scoring against the domains; week three, timed integrated SOO and short-menu SAMP cases; week four, unseen simulated cases treated as a mock. The knowledge layer ticks over underneath on a small maintenance quota. The measurable signal she now tracks is not her Q-bank percentage but her SOO competency ratings on unseen cases — which is the number the exam actually cares about.
A balanced case and task matrix
Candidates naturally practise the cases they enjoy and are already good at. That is how people arrive expert in chest pain and untrained in the awkward consultation. Build a matrix so your practice samples the breadth of family practice deliberately: across the life cycle (child, adult, older adult, end-of-life), across presentation type (undifferentiated symptom, chronic-disease review, mental health, prevention, the psychosocial or "difficult" consultation), and across setting (in-person and virtual). Tick cases off the matrix, not off your comfort. The exam samples breadth on purpose; your practice should too.
Red flags that your CCFP preparation is off track
- Memorised scripts: reciting a fixed SOO opening rather than responding to the patient in front of you. Examiners see through it, and real patients do not follow the script.
- Repeated cases: re-running the same practice cases until you recognise them, which trains recognition, not consultation skill.
- Generic feedback: "that was good" tells you nothing; feedback must map to a specific competency and give one thing to change.
- Uncalibrated scoring: trusting a number — from a peer or from an AI — that has never been checked against the official rubric.
- No official-rubric check: practising without ever comparing your performance to the CFPC's own competencies and sample material. The official material is the calibration gold-standard; a bank supplies volume, not the standard.
Three mistakes this modality audit is designed to stop
The first is mistaking a high Q-bank percentage for readiness. A percentage measures recognition on items you have often already seen; it is silent on whether you can consult, communicate or prioritise under uncertainty — and it is silent by design, because those are not what an MCQ tests. The second is treating the SAMP short-menu transition as a reason to do more generic MCQs. The format is moving to short-menu, but the construct is still family-medicine reasoning; drilling recognition of single best answers can actively erode the differential breadth and prioritisation the SAMP rewards, so the transition is a reason to train reasoning harder, not to grind more recall. The third is rehearsing performance without a rubric or a human. Running the same practice SOO with the same friend until it feels smooth trains familiarity, not competence; without scoring against the CFPC domains and without a clinician's safety check, you cannot tell whether a consultation that felt good would actually pass. Each mistake has the same root: substituting the modality that is easy to practise and easy to measure — MCQs — for the modalities the exam is built around. The audit exists to redirect your marginal hour to where the real gap is.
Bottom line
A question bank is the foundation of CCFP knowledge, and as the SAMP moves to short-menu it will look more like exam practice than ever — but it still cannot assess the things the exam is built around: reasoning under uncertainty, communication, and consultation performance, increasingly on screen. Cover the knowledge layer efficiently, then spend your marginal hour where MCQs cannot reach: coached, rubric-referenced, unseen practice of the SOO and the virtual consultation. iatroX can give you the unseen knowledge baseline and measure it honestly; it is not, and does not claim to be, a SOO simulator. Train the performance where performance is trained.
Frequently asked questions
How do I know whether I have covered the full CCFP blueprint? Map your practice against the CFPC's published priority topics and assessment objectives rather than your Q-bank's internal categories, and mark each area as covered only when you can perform in it — generate the management, and for consultation skills demonstrate the behaviour — not merely recognise an answer. A blueprint-coverage matrix that tracks both knowledge and performance, ticked against the official objectives, is the most reliable way to see the gaps; a completion percentage on one bank is not.
Can one question bank be enough for CCFP? No. One bank can be enough for part of the knowledge layer, but no bank assesses the SOO, the virtual consultation or the full reasoning a SAMP case demands, so a bank alone cannot prepare you for the CCFP. Treat the bank as necessary groundwork and add coached, rubric-referenced performance practice for the components MCQs cannot touch.
What should I measure instead of my overall Q-bank percentage for CCFP? Measure blueprint coverage (are all priority areas addressed?), first-attempt performance on unseen questions under time, and — separately and most importantly — performance on the SOO and virtual consultation, scored against the CFPC competencies by a human. Your overall percentage conflates seen and unseen items and says nothing about consultation skill; it is the least informative number you have. Your Q-bank percentage is not your exam score.
When should I stop doing new CCFP questions? Stop adding new questions once your knowledge coverage is adequate and your first-attempt performance on unseen items is stable, and redirect that time to the performance components, which is where most candidates' marginal gains actually are. Continuing to grind fresh MCQs past the point of stable knowledge is comfortable and low-yield; unseen SOO practice is uncomfortable and high-yield.
Which CCFP resource should I use for my weakest component? Match the resource to the modality: for knowledge gaps, a question bank and current Canadian guidance; for SAMP reasoning, timed short-menu cases with model-answer feedback; for SOO communication, role-play with a GP tutor or trained peer scoring against the CFPC domains; for virtual consultation, recorded video consultations reviewed by a clinician. The weakest component is almost never fixed by more of the resource you already use most — usually it is the performance layer, and that needs a human and a rubric, not another bank.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Exam-format details, including the 2026–27 SAMP transition, are vendor/College-reported and subject to change — verify current components, timeline and any case counts on cfpc.ca; exact per-component counts are not published, so none are asserted here. Disclosure: iatroX operates a competing question bank; in this article iatroX is confined to the unseen knowledge-measurement job and is explicitly not positioned as a SOO or consultation simulator. Corrections via the feedback route on iatrox.com.
References: College of Family Physicians of Canada — Certification Examination in Family Medicine format and objectives, and the SAMP transition notice (cfpc.ca); the iatroX CCFP hub (https://www.iatrox.com/canada/exam/ca-ccfp); "Your Q-Bank Percentage Is Not Your Exam Score" (https://www.iatrox.com/blog/qbank-percentage-not-your-exam-score); the completion-is-not-coverage blueprint-matrix pillar (https://www.iatrox.com/blog/question-bank-completion-is-not-coverage-how-to-build-a-blueprint-coverage-matrix-for-any-medical-exam); calibrating AI-graded feedback (https://www.iatrox.com/blog/ai-graded-saqs-and-osces-how-to-calibrate-automated-feedback-before-you-trust-the-score); the iatroX comparison hub (https://www.iatrox.com/compare).