This workflow is for international medical graduates and Canadian final-year students who have bought, or are weighing, AIMedQs for the MCCQE Part I — the multiple-choice paper that gates the Licentiate of the Medical Council of Canada. AIMedQs markets an adaptive engine that pushes you towards weak areas, which is genuinely useful on a single-best-answer exam. Its principal limitation is equally predictable: an algorithm optimised for your rising overall score can quietly starve small blueprint domains you rarely see. This guide shows you when to follow it, when to override it, and when to stop tuning and simply sit mixed timed blocks.
What AIMedQs offers for the MCCQE Part I right now
The figures below are vendor-reported and last checked on 19 July 2026. Treat them as a starting point and confirm the current numbers at checkout, because they move.
| Item | What AIMedQs states (vendor-reported, 19 July 2026) |
|---|---|
| MCCQE Part I question volume | The MCCQE1 page headlines "over 1,300 practice MCQs aligned to MCC objectives," while the same page's FAQ cites "over 500 MCCQE1-specific" items. Treat the true MCCQE1-specific pool as uncertain and verify it before you pay. |
| Adaptive/AI features | "Adaptive question selection" that identifies weak areas and prioritises them, spaced repetition of missed items, and performance analytics by discipline, plus written explanations. There is no conversational chatbot tutor. |
| Access and pricing | Tiered monthly plans listed at roughly $29, $49 and $79 per month (Starter/Pro/Premium), with a small free trial pool. Confirm the currency and current tiers at checkout. |
| Components supported | The written MCCQE Part I MCQ paper only. AIMedQs does not simulate the NAC OSCE or any clinical station, and it is not the exam itself. |
The important honesty point is the one word doing a lot of work in searches: "tutor." AIMedQs sells an adaptive selection engine plus explanations, not a Socratic chat partner you can interrogate. That distinction changes the workflow. You are not outsourcing reasoning to a tutor; you are deciding, block by block, whether the machine's routing serves your blueprint or your ego.
The exam you are actually training for
The MCCQE Part I is now, since the 2025 change, a multiple-choice-only computer-based examination. The clinical decision-making (CDM) component that older candidates remember has been removed. The current paper is 230 MCQs delivered in two sections of 115 questions, each section timed at two hours and forty minutes, so budget for roughly five and a half hours in the seat plus tutorials and breaks. Items carry one best answer among three to five options; confirm the live count and timing on mcc.ca before your diet, as the MCC updates this page.
The blueprint matters more than the raw count. MCC objectives sit under the CanMEDS roles and are sampled across two axes: Dimensions of Care (health promotion and illness prevention, acute, chronic, and psychosocial aspects) and Physician Activities (assessment and diagnosis, management, communication, and professional behaviours). The official preparatory materials — the MCC's own practice test and self-assessment products — are the calibration gold standard because they are written to that blueprint; verify which products are currently sold on mcc.ca. Everything AIMedQs offers is third-party interpretation of those objectives, useful but not authoritative. Keep that hierarchy clear: official material calibrates, the bank supplies volume.
Baseline week: measure before you personalise
Before you let any algorithm shape your feed, take a reading it cannot influence. In week one, complete a small, blueprint-stratified unseen sample — around 120 to 150 questions spread deliberately across all four Dimensions of Care and all four Physician Activities, in timed mode, with explanations hidden until the block ends. The point is a clean baseline: your first-attempt accuracy per domain when nothing has been pre-selected for you.
Record that baseline in a simple table before AIMedQs starts adapting. Once the engine begins routing, your overall percentage and your domain mix are both moving, and you lose the ability to say which improvement is learning and which is the machine feeding you easier terrain. A fixed baseline is the anchor you will compare every later block against.
First pass: set domain floors the algorithm cannot hide
Adaptive selection has a structural blind spot. If it concentrates on your weak areas, it can under-sample areas you are merely mediocre in, and it can leave genuinely low-volume blueprint domains — psychosocial aspects, health promotion, professional behaviours — almost untouched while your headline score climbs on internal medicine and paediatrics. A rising overall percentage is not coverage.
So impose floors the algorithm has to respect. Set a minimum number of first-attempt questions per Dimension of Care and per Physician Activity — for example, at least 60 unique items in each of the eight cells before you trust any readiness signal. Track attempts per cell manually if AIMedQs's analytics do not expose it. When a cell is under floor, override the feed: use topic filters to force questions there, even if the engine would rather keep you on your weakest discipline. Following the algorithm is efficient for depth; overriding it is how you buy breadth. You need both.
An error taxonomy that tells you what to do next
"Got it wrong" is not a diagnosis. Every miss should be coded, because the code determines the corrective action. Use six categories:
- Knowledge gap — you did not know the fact or mechanism.
- Misread stem — you knew the content but misread the vignette, the negative, or the "most likely" qualifier.
- Premature closure — you anchored on the first plausible option and stopped comparing.
- Guideline/jurisdiction error — you applied a rule that is wrong, out of date, or not the Canadian convention.
- Calculation error — arithmetic, units, or a dosing/interval slip.
- Time-pressure error — you would have got it with thirty more seconds.
Tally these weekly. The distribution is diagnostic. A pile of knowledge gaps means more content review and more questions; a pile of misread stems means slow down and annotate the stem before looking at options; premature closure means force yourself to eliminate all five options in writing; time-pressure errors mean your pacing, not your knowledge, is the constraint.
Review intervals: not everything deserves an immediate repeat
The reflex to re-answer a missed AIMedQs item straight away mostly trains recognition of that item, which the exam will never show you. Sort misses by code. Knowledge gaps deserve a short source read (the relevant MCC objective, then a current Canadian guideline or a reference chapter) and then a new transfer question — a different item testing the same principle — rather than a repeat. Misread-stem and premature-closure errors deserve a spaced review of the reasoning habit, not the fact. Only genuinely high-yield facts you keep forgetting belong in spaced repetition. Let AIMedQs's spaced repetition handle those, but do not let it convince you that re-clearing old items is the same as demonstrating transfer on unseen ones.
When to move to mixed blocks
Topic-filtered practice cues you: when you know the block is all cardiology, you reason differently than you will on exam day. The switch to mixed, randomised, timed blocks should be criteria-driven, not calendar-driven. Move to predominantly mixed blocks when, and only when, all eight blueprint cells are above floor, your first-attempt accuracy on unseen items is stable across at least three sessions, and your pacing sits comfortably under the roughly 80 seconds per item the paper allows. Until then, keep filtering to fill floors. After then, filtered practice should fall to a small remedial minority and random mixed blocks should dominate, because that is the format you are actually sitting.
Exit criteria: what "ready" actually looks like
You are ready to reduce bank work when five things are simultaneously true: every blueprint cell is above your coverage floor; first-attempt accuracy on unseen mixed blocks is stable rather than still climbing steeply; your pacing is under the per-item budget with time to review flags; your error log shows retention (yesterday's fixed misconception is still fixed a week later); and your performance on official MCC practice material agrees with your bank performance. Bank completion is not on that list. Finishing every AIMedQs question proves you have seen them, not that you can transfer. Your Q-bank percentage is not your exam score — the calibration you trust is unseen, timed and blueprint-stratified.
A worked seven-day plan for an international graduate
Consider Priya, an IMG four weeks out, strong in medicine, shaky on psychosocial and health-promotion items and on Canadian conventions (consent thresholds, reportable conditions, screening intervals). She uses AIMedQs for one defined job — driving weak-domain depth — and iatroX for unseen transfer measurement, without either of us claiming to know AIMedQs's internal routing.
- Day 1: 40 mixed timed AIMedQs items; code every miss; tally by domain.
- Day 2: Override the feed — topic-filtered psychosocial and health-promotion blocks to lift both cells towards floor; one short source read per knowledge gap.
- Day 3: 40 AIMedQs items on management and communication; log Canadian-convention errors specifically.
- Day 4: Transfer day — a fresh, timed, mixed MCCQE Part I block in iatroX, no topic filter, to measure whether Day 2's work transferred to unseen items. Compare against the week-one baseline, not against AIMedQs's own dashboard.
- Day 5: Remediate the iatroX misses by principle; new transfer questions, not repeats.
- Day 6: One official MCC practice component under timed conditions to calibrate.
- Day 7: Rest or light spaced-repetition review only.
The division of labour is deliberate: AIMedQs personalises the grind, iatroX supplies the unseen ruler, and the official material is the tie-breaker. Nobody has to reverse-engineer a proprietary algorithm for this loop to work.
Decision checklist: continue, supplement, switch or stop
- Continue with AIMedQs if your floors are filling, coded errors are shifting from knowledge gaps to reasoning slips, and unseen iatroX blocks are rising in step with the AIMedQs dashboard.
- Supplement (add a second, unseen bank) if your AIMedQs percentage climbs but unseen transfer stalls — that gap means recognition, not learning. The two-Q-bank rule tells you how to add breadth without duplicating items.
- Switch if the MCCQE1-specific pool turns out to be small enough that you exhaust unseen items well before exam day, or if the Canadian-convention accuracy is poor.
- Stop adding bank time when all exit criteria are met; more questions past that point is reassurance-seeking, not preparation.
Make these calls on measurable gaps — coverage cells, transfer trend, pacing — never on novelty or on the sunk cost of a subscription you have already paid for.
Frequently asked questions
Is AIMedQs enough for MCCQE Part I on its own? For most candidates, no single third-party bank is enough on its own, and AIMedQs is no exception. Its adaptive routing is a reasonable depth tool, but the MCCQE1-specific pool is vendor-reported and inconsistently stated (the page cites both "1,300+" and "500+"), and a modest unseen pool is exhausted quickly. Pair it with the official MCC practice material for calibration and at least one unseen bank for transfer measurement before you treat any score as meaningful.
Which MCCQE Part I component does AIMedQs not reproduce well? AIMedQs reproduces the written MCQ paper's format acceptably but does not give you a conversational tutor to interrogate your reasoning, and it does not touch the clinical/OSCE side of Canadian licensing at all (the NAC OSCE is a separate assessment). Its weakest reproduction is exam-day randomness: adaptive, topic-aware routing is the opposite of the blind mixed sampling you face on the day, so you must deliberately switch to mixed blocks it has not curated.
How many AIMedQs questions should I complete per day for MCCQE Part I? There is no magic number, but a sustainable target for most full-time candidates is 40 to 60 first-attempt items per day, always reviewed by error code, plus spaced repetition of prior misses. Quality of review beats raw volume: 40 items you code and remediate properly will move you further than 100 items you skim. Scale down if your review backlog is growing faster than you clear it.
When should I stop using AIMedQs and move to mixed mocks? Move when all eight blueprint cells are above your coverage floor, unseen first-attempt accuracy is stable across at least three sessions, and pacing is under roughly 80 seconds per item. At that point, filtered adaptive practice has given you most of what it can, and the marginal gain comes from rehearsing the real format — full-length, mixed, timed, unassisted — and from calibrating against official material.
How should I combine AIMedQs with iatroX without duplicating practice? Assign each tool one job. Use AIMedQs for adaptive weak-area depth and its own spaced repetition; use iatroX only for fresh, timed, mixed blocks that measure transfer to unseen items, and never re-import an AIMedQs item into your iatroX sessions. The moment the same question appears in both, your iatroX score stops being an unseen measurement and becomes a memory test. Keep the measurement bank clean.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. All AIMedQs figures (question counts, tiers and prices) are vendor-reported as displayed on the AIMedQs site on that date and can change without notice; confirm them at checkout and confirm the exam format on mcc.ca before relying on any of it. Disclosure: iatroX operates a competing MCCQE Part I question bank; to keep this useful rather than self-serving, iatroX is confined here to the one job AIMedQs does not claim — clean, unseen, timed transfer measurement — and is not presented as a replacement for the official MCC materials or for a purpose-built Canadian bank. No claim is made about AIMedQs's internal adaptive algorithm, which is not disclosed. Corrections are welcome via the feedback route on iatrox.com. References: Medical Council of Canada, MCCQE Part I exam-day and objectives pages (mcc.ca); AIMedQs MCCQE1 product and pricing pages; iatroX MCCQE Part I bank, the blueprint-coverage matrix method, and the comparison hub.