Lead with the finding, because it changes everything that follows: Rosh Review's emergency-medicine content — now on the Blueprint Prep platform — is built and blueprinted to the American ABEM (the In-Training and Qualifying exams), and its own pages make no mention of the RCPSC, the Royal College, or any Canadian emergency-medicine certification. So for an RCPSC Emergency Medicine candidate, Rosh's much-praised analytics and "predict your likelihood of passing" feature are optimising readiness for a different exam. This audit explains what its engine actually does, where its signals mislead a Royal College candidate, and how to use it honestly anyway.
Current-state box (checked 19 July 2026)
Figures are vendor-reported by Rosh Review / Blueprint Prep and were last checked on 19 July 2026.
| Item | What Rosh/Blueprint reports (vendor-reported) |
|---|---|
| EM exams covered | ABEM In-Training Exam, ABEM Qualifying Exam, MyEMCert modules, EM CME/lifelong learning |
| RCPSC coverage | None stated — no Canadian or Royal College emergency-medicine product |
| Question volume | EM Resident QBank ~2,000 ABEM-formatted questions; Qualifying QBank ~3,000 or ~5,000; plus CME, critical-care, ECG and ultrasound add-ons |
| Analytics | Personal analytics dashboard with completion and knowledge-gap views; "predict your likelihood of passing"; tutor and test modes |
| Adaptive/AI | Marketed with adaptive study features; no detailed adaptive-difficulty algorithm published |
| Price | Not displayed on the reviewed page; confirm current pricing on roshreview.com / blueprintprep.com |
| Blueprint | ABEM EM Model, not the RCPSC objectives of training |
Exam anchor: what the RCPSC actually assesses — and how ABEM differs
RCPSC Emergency Medicine certification has two parts: a computer-based written multiple-choice examination and a separate Applied examination (structured oral and OSCE-style stations), both mapped to the Royal College objectives of training and CanMEDS roles. The Royal College does not publish an accessible question count or session breakdown, so verify the current structure on royalcollege.ca. Rosh, by contrast, is engineered to the ABEM EM Model — twenty clinical domains with an acuity distribution (roughly critical 30%, emergent 40%, lower acuity 21%) and minimum paediatric and geriatric proportions. The clinical knowledge overlaps heavily, because emergency medicine is emergency medicine. The blueprint weighting, the jurisdiction, and the entire Applied component do not. That mismatch is the theme of this audit.
Define every metric the platform shows you
You cannot audit a dashboard you have not defined. Here is what each Rosh metric means, and what it does not.
| Metric | What it measures | What it does not tell an RCPSC candidate |
|---|---|---|
| First-attempt accuracy | % correct the first time you saw an item | Readiness for the RCPSC blueprint or standard |
| Repeat accuracy | % correct on re-attempted items | Anything reliable — it is recognition, not knowledge |
| Percentile | Your rank against other Rosh users | Your standing against ABEM candidates, not RCPSC ones |
| Predicted pass likelihood | A model of your ABEM pass probability | Nothing about the Royal College exam it is not calibrated to |
| Coverage | % of the Rosh bank completed | % of the RCPSC blueprint covered |
| Difficulty | Item difficulty within the Rosh pool | Difficulty relative to the RCPSC standard |
| Time per item | Your pace on Rosh items | Your pace under RCPSC written conditions |
The single most important line in that table is the predicted-pass one: it is a genuine, useful model — for ABEM. Read as an RCPSC prediction, it is a category error.
Selection bias: why an adaptive feed distorts your percentage
Adaptive and "smart" feeds tend to over-sample your weak areas — that is their point — which means the mix of questions you see is deliberately harder and narrower than a representative paper. Two distortions follow. First, your running percentage is depressed relative to a balanced blueprint, so a "low" number may reflect the feed's difficulty rather than your standing. Second, and more dangerously, once you have driven your weak areas up, the feed may switch to easier familiar territory and your percentage climbs for the wrong reason. Either way, a self-selected or adaptive percentage is not comparable with a mixed, unseen, representative block — which is the only fair readiness signal.
Blueprint audit: map your attempts, do not trust the average
Do not trust the home-screen average; audit your attempted-question distribution against the RCPSC domain weighting yourself. Export or tally what you have actually attempted, bin it into the Royal College's emergency-medicine domains (not Rosh's ABEM bins), and look for thin cells. Because Rosh is weighted to the ABEM EM Model, expect the mapping to leave under-sampled areas wherever the two blueprints diverge — and expect Canadian-context topics to be thin or absent. This re-mapping is the completion-is-not-coverage exercise applied to a bank built for another country's exam.
The readiness test: what a credible signal requires
A number only counts as a readiness signal if it meets five conditions at once:
- Unseen — items you have never attempted, so it measures knowledge, not memory.
- Timed — under the pace the RCPSC written paper demands.
- Mixed — representative of the whole blueprint, not an adaptive feed of your weak spots.
- No assistance — no tutor mode, no hints, no lookups.
- Adequate sample — enough items that the estimate is stable, not a lucky ten.
Rosh's tutor mode and adaptive feed are excellent for learning; they violate at least three of these conditions, so they cannot double as your readiness measurement. Reserve a separate, unseen, representative block for that job.
Algorithm override rules
An adaptive engine optimises for its own objective, which may not be your blueprint. Override it deliberately:
- Force low-volume domains. If the feed rarely surfaces environmental emergencies, toxicology sub-topics or specific paediatric presentations, schedule them manually.
- Force image and ECG density. Ensure you meet a self-imposed quota of ECGs, radiographs and clinical images, which an accuracy-driven feed may under-serve.
- Force ethics, communication and medico-legal items — the material the Applied component probes and a knowledge feed under-weights.
- Force calculations — doses, rates, infusion maths — under time pressure.
- Force Canadian-context revision — the jurisdiction the bank does not cover.
Worked dashboard example (no pass prediction)
A candidate's Rosh dashboard reads: overall accuracy 74%, cardiology 82%, toxicology 61%, paediatrics 58%, environmental 55%, and a cheerful "likely to pass" flag. Ignore the flag — it is an ABEM model. Convert the rest into next week's quotas. The three domains below your 70% threshold (toxicology, paediatrics, environmental) become the week's learning targets: three tutor-mode blocks of 15, each fully reviewed. Then, and separately, book one unseen, mixed, timed block of 40 items — no tutor, representative of the RCPSC blueprint — as the measurement, not on Rosh's home average. The dashboard set your learning quotas; it did not, and cannot, predict your Royal College result.
Worked example: a seven-day plan around clinical work
Give Rosh one job — teaching and drilling defined weak EM domains at ABEM-level depth — and give an unseen bank the separate job of RCPSC-blueprint measurement.
- Monday (1h): Rosh tutor-mode block, 15 items, weakest domain (toxicology), every miss annotated.
- Tuesday (1h): Read the Rosh explanations for misses; flag any US-jurisdiction or availability assumptions.
- Wednesday (1h): Fresh Rosh block, same domain, timed, to confirm retention.
- Thursday (1h): Manual override — a block of images/ECGs and a set of calculation items the feed under-served.
- Friday (1.5h): Switch tools. Unseen, mixed, timed iatroX block mapped to the RCPSC EM blueprint, reviewed with the Socratic tutor that names the misconception behind each miss.
- Saturday (1h): Canadian-context revision on the week's topics; date every guidance-sensitive point.
- Sunday (1.5h): Oral/OSCE-style case practice with a colleague for the Applied component.
No item is tested twice, and no proprietary algorithm is assumed: Rosh teaches EM knowledge at depth, iatroX measures unseen transfer against the RCPSC blueprint, a human rehearses the oral.
Decision checklist: continue, supplement, switch or stop
- Continue Rosh for its EM teaching depth while its items are unseen and its explanations are moving your understanding.
- Supplement it — necessarily — with RCPSC-blueprint measurement, Canadian-context revision, and Applied-component practice, none of which it provides.
- Switch your measurement off Rosh entirely: never treat its ABEM percentile or pass prediction as an RCPSC readiness signal.
- Stop adding Rosh volume when unseen mixed RCPSC-mapped blocks are stable and your remaining gaps are Applied-component skills.
Frequently asked questions
Is Rosh Review and Blueprint Prep enough for RCPSC Emergency Medicine on its own? No — and this is the central caution. Rosh is a strong, deep emergency-medicine knowledge bank, but it is built and blueprinted to the American ABEM, it carries US-jurisdiction assumptions, its analytics and pass prediction are calibrated to ABEM rather than the Royal College, and it does nothing for the separate RCPSC Applied (oral/OSCE) component. It can supply shared EM clinical knowledge, but on its own it prepares you for a different exam. Use it for depth, and add RCPSC-blueprint measurement, Canadian-context revision and oral practice around it.
Which RCPSC Emergency Medicine component does Rosh Review and Blueprint Prep not reproduce well? The Applied examination, entirely — there is no oral, no OSCE station and no way to score live resuscitation discussion or communication. It also does not reproduce the RCPSC written blueprint weighting or the Canadian practice context, because both follow the ABEM EM Model and US norms. Its genuine strength is deep, well-explained EM knowledge at the item level; the blueprint fit, jurisdiction and the entire performative component sit outside its scope.
How many Rosh Review and Blueprint Prep questions should I complete per day for RCPSC Emergency Medicine? Volume matters less than review quality and blueprint targeting. For most trainees revising around shifts, one or two fully reviewed blocks of 15–20 items — chosen to hit your measured weak domains rather than whatever the feed serves — is more valuable than 60 rushed items. Because the banks are large, you can afford to be selective; spend the saved time writing a corrected rule for every miss and manually forcing the low-volume domains and Canadian-context topics the feed under-serves. Depth and targeting beat raw count.
When should I stop using Rosh Review and Blueprint Prep and move to mixed mocks? Move to mixed, unseen, timed blocks mapped to the RCPSC blueprint once your targeted Rosh domains are consistently strong and your errors are transfer, timing or jurisdiction issues rather than knowledge gaps. Continuing to grind an adaptive ABEM feed past that point cannot measure your Royal College readiness, because its percentile and pass prediction were never calibrated to the RCPSC. Keep Rosh for re-teaching any domain a mock exposes, but make unseen, RCPSC-mapped simulation the readiness signal.
How should I combine Rosh Review and Blueprint Prep with iatroX without duplicating practice? Give them separate jobs. Use Rosh for deep, ABEM-level teaching and drilling of named EM domains; use iatroX for the different job of unseen, mixed, timed measurement mapped to the RCPSC blueprint, reviewed with a Socratic tutor that names the misconception behind each miss — and never treat Rosh's ABEM analytics as an RCPSC signal. The rule is that you never re-test yourself on items you have seen in either bank; our two-Q-bank rule explains how to keep two banks from contaminating each other's calibration, and the comparison hub maps the division of labour.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; Rosh Review / Blueprint Prep's exams covered, question counts, analytics and prices are vendor-reported and its emergency-medicine content is blueprinted to the American ABEM, not the RCPSC — confirm the current details on roshreview.com or blueprintprep.com, and the current RCPSC structure on royalcollege.ca, before you rely on them. Disclosure: iatroX operates a competing Canadian question bank mapped to the RCPSC blueprint; this article confines iatroX's role to unseen RCPSC-blueprint measurement and misconception-level feedback that Rosh's ABEM analytics do not provide, and it does not position iatroX as a replacement for the Applied (oral/OSCE) examination. Corrections are welcome via the feedback route on iatrox.com.
References: Royal College of Physicians and Surgeons of Canada — Emergency Medicine examination format and applied exams (royalcollege.ca/en/eligibility-and-exams); Rosh Review / Blueprint Prep — emergency-medicine question banks (roshreview.com/em, blueprintprep.com); iatroX Canada exam hub (iatrox.com/canada); iatroX — "Your Q-Bank Percentage Is Not Your Exam Score"; iatroX framework pillars on completion-is-not-coverage and the two-Q-bank rule.
Run a fresh timed RCPSC Emergency Medicine block in iatroX →