If you came here to build an adaptive NEJM Knowledge+ study loop for the American Board of Emergency Medicine (ABEM) Qualifying Examination, begin with the finding that changes the plan: as of 19 July 2026, NEJM Knowledge+ does not publish an Emergency Medicine product. Its adaptive engine is real and well regarded, but it powers Internal Medicine, Family Medicine and Paediatrics — not EM. There is no NEJM Knowledge+ ABEM bank to follow, override or graduate away from. This article says so plainly, then shows how to run the same workflow on a bank that does cover ABEM.
What NEJM Knowledge+ actually covers right now
NEJM Knowledge+ is an adaptive board-review platform now delivered through AMBOSS, which acquired it. Its strength is a personalised learning engine built around spaced repetition and confidence-based questioning — the "Adaptive Personalized Learning" approach NEJM described in its own literature, with published associations between product use and internal-medicine maintenance-of-certification performance. That engine is genuinely good at what it does. The problem for an emergency physician is scope, not quality.
The table below reflects the product line as advertised on the publication date. Treat every count and price as vendor-reported and confirm the live figures before purchase.
| NEJM Knowledge+ product | Target certification | Use for ABEM |
|---|---|---|
| Internal Medicine Board Review | ABIM IM (initial + MOC) | Not applicable |
| Family Medicine Board Review | ABFM | Not applicable |
| Paediatrics Board Review | ABP | Not applicable |
| Pain Management & Opioids module | CME / risk education | Not applicable |
| Emergency Medicine | — | Does not exist |
A necessary caveat, because it prevents a wrong turn: AMBOSS — the parent platform — does market a separate Emergency Medicine (ABEM) board review. That is an AMBOSS-branded product, not a "NEJM Knowledge+ Emergency Medicine" bank, and it should be evaluated on its own current question count, blueprint mapping and price. If a colleague tells you they "used NEJM Knowledge+ for EM boards", they almost certainly mean the AMBOSS EM bank or one of the dedicated EM vendors. The NEJM Knowledge+ brand itself has no EM offering to audit.
The ABEM Qualifying Examination blueprint
Because the workflow has to be re-homed on an EM bank, it helps to fix the target first. The ABEM Qualifying Examination is a computer-based, single-best-answer multiple-choice test delivered at Pearson VUE, with roughly 305 scored items. It is built on the Model of the Clinical Practice of Emergency Medicine — the EM Model — which organises content into around twenty condition and organ-system categories layered over a set of physician tasks. ABEM publishes an acuity frame for the blueprint: approximately 30% critical, 40% emergent and 21% lower-acuity presentations, with minimum representation for paediatric content (at least 8%) and geriatric content (at least 6%), plus image and pictorial items. Passing the Qualifying Examination is the gateway to the Oral Certifying Examination; the written paper does not itself confer certification. No multiple-choice bank — NEJM Knowledge+, AMBOSS, or anyone else — reproduces that oral component.
Keep the distinction between official requirements and third-party claims. The EM Model and the acuity weighting are ABEM's; a vendor's "aligned to the blueprint" statement is a mapping claim you should spot-check, not an official guarantee.
Why the adaptive workflow still applies — just not on this product
The adaptive loop that NEJM Knowledge+ made popular is sound: baseline honestly, let an algorithm surface your weak areas, correct the underlying error, then prove the correction on fresh material. The loop is portable. What is not portable is running it on a product that has no EM questions. So the remaining sections describe the loop applied to whichever EM bank you choose — Rosh Review, ACEP PEER, BoardVitals, the AMBOSS EM bank — with iatroX as the unseen-measurement layer that keeps the adaptive feed honest.
Baseline week: a blueprint-stratified unseen sample
Before you let any engine personalise your feed, take a small, deliberately mixed sample so you know where you actually stand. Draw fifteen to twenty items from each of the large EM categories — cardiovascular, respiratory, neurological, abdominal and gastrointestinal, trauma, toxicology, environmental, paediatric and geriatric presentations — under timed, no-notes conditions. That is your reference distribution. Do not "study" from it; use it to set floors. A rising overall percentage on a personalised feed means little if you never sampled the domains the algorithm quietly de-prioritised.
First pass: set domain floors so the algorithm cannot hide gaps
Adaptive feeds optimise for your aggregate accuracy, which is not the same as blueprint coverage. Set a minimum number of unique attempted items per EM Model category and per acuity band before you trust any headline score. Practically: define floors such as "at least 60 unique items and 70% first-attempt accuracy in toxicology and environmental" or "at least 8% of total attempts in paediatric presentations, matching the blueprint minimum." Track attempted-question distribution, not just accuracy. This is the completion-is-not-coverage problem, and it is exactly what a blueprint-coverage matrix is for.
An error taxonomy for EM board questions
An immediate re-attempt of a missed item teaches recognition, not reasoning. Before you decide what to do with a mistake, label it:
- Knowledge gap — you did not know the fact or the guideline.
- Misread stem — you missed a qualifier, an age, a vital sign or the actual question being asked.
- Premature closure — you anchored on the first plausible diagnosis and stopped.
- Guideline error — you applied an outdated or wrong threshold, dose or disposition rule.
- Calculation error — anion gap, corrected sodium, paediatric weight-based dosing, drip rates.
- Time-pressure error — you knew it but rushed under the clock.
The taxonomy matters because each error type has a different fix, and lumping them together is how candidates waste the final month.
Review interval: repeat, transfer, or read the source
Map each error type to a response. A knowledge gap earns a short source read (a guideline, a society statement, a focused reference) and a new transfer question on the same concept a few days later — not a re-attempt of the same stem. Premature closure and misread-stem errors earn a process fix (a fixed reading routine) rather than more content. Guideline and calculation errors earn a written rule you can rehearse and then test on unseen items. Time-pressure errors earn timed mixed blocks, not untimed topic drilling. The unifying principle: prove the correction on a question you have never seen, so you are measuring transfer rather than memory of the explanation.
The mixed-block switch
Topic-filtered practice is a scaffold for the early and middle weeks. You should reduce it once two objective conditions hold: every EM Model category has cleared its coverage floor, and your first-attempt accuracy on topic blocks has been stable for two to three weeks. At that point, shift the balance toward timed, random, full-length-style blocks that mirror the roughly one-minute-per-item pace of the real paper. The switch is a data decision, not a calendar decision.
Exit criteria
You are ready to taper active drilling when four things are true, and completing a bank is not one of them: coverage floors are met across all categories and acuity bands; first-attempt accuracy on unseen mixed blocks is stable, not merely high on reviewed items; your pace is comfortable at exam speed; and you have calibrated against genuinely official material — ABEM's own content specifications and any released sample items — rather than a single vendor's estimate. Retention over two to three weeks matters more than a peak score on a Tuesday.
A seven-day worked example
Here is one week for an emergency physician who has chosen an EM bank for adaptive drilling and uses iatroX purely to measure transfer on unseen items. No proprietary-algorithm claims are made about either tool; this is a scheduling pattern, not a black box.
- Monday — 40 timed items on your EM bank in your two weakest categories from the baseline. Tag every miss with the taxonomy.
- Tuesday — Source reads for Monday's knowledge gaps only; write three one-line rules (a threshold, a dose, a disposition).
- Wednesday — 30 unseen mixed items in iatroX, timed. Compare the error types against Monday, not the wording of individual questions.
- Thursday — 40 items on the EM bank in the next weakest category; refresh paediatric and geriatric floors.
- Friday — Calculation and image day: ECGs, radiographs, toxidromes, weight-based dosing.
- Saturday — One timed random block at exam length and pace; log pacing and first-attempt accuracy.
- Sunday — Review the week's taxonomy tally; set next week's floors from the gaps, not from novelty.
The division of labour is deliberate: your main EM bank drives the adaptive learning; iatroX supplies fresh, unseen items so your percentage reflects reasoning you can transfer, not questions you have already seen. That is the two-Q-bank rule in practice.
Decision checklist: continue, supplement, switch or stop
- Continue the adaptive loop while unseen first-attempt accuracy is still improving and coverage floors are not yet met.
- Supplement with a second, unseen bank when your main feed keeps recycling familiar items and your reviewed-item score has decoupled from your unseen score.
- Switch banks only for a measurable reason — a category the current bank barely covers, or a persistent blueprint gap — not for novelty or because a discount landed.
- Stop active drilling when the exit criteria are met; extra volume past that point buys fatigue, not marks.
Three mistakes this workflow is designed to stop
First, mistaking a rising personalised percentage for readiness. Adaptive feeds are engineered to make you feel better over time; only unseen, mixed, timed blocks tell you whether that feeling is earned. Second, treating every miss as a knowledge gap. Most late-stage errors are process errors — misreads, premature closure, pacing — and no amount of extra content fixes them. Third, shopping for the "perfect" bank. The NEJM Knowledge+ episode is instructive: candidates lose weeks searching for a named product to run a loop, when the loop matters far more than the brand, and the brand in question does not even cover EM.
Bottom line
NEJM Knowledge+ is a strong adaptive platform for internal medicine, family medicine and paediatrics, and nothing here disputes that. For ABEM it is simply the wrong door: there is no NEJM Knowledge+ Emergency Medicine bank to follow or override. Choose an EM bank that maps to the EM Model, run the baseline-floors-taxonomy-mixed-block loop, and use iatroX to keep your percentage honest on unseen items. That combination — not the search for a specific logo — is what moves the needle.
Frequently asked questions
Is NEJM Knowledge+ enough for ABEM on its own? No, and the reason is more fundamental than "not enough": NEJM Knowledge+ does not currently publish an Emergency Medicine product, so there is no ABEM bank to be enough or otherwise. Its Internal Medicine, Family Medicine and Paediatrics banks are aimed at different certifications and do not map to the EM Model. For ABEM you need a dedicated EM bank; NEJM Knowledge+ is not one.
Which ABEM component does NEJM Knowledge+ not reproduce well? It reproduces none of them, because it has no EM content. The more useful version of the question applies to every multiple-choice bank: none reproduces the ABEM Oral Certifying Examination, which tests structured verbal case management rather than single-best-answer recognition. Plan the oral separately with case-based verbal practice, not a Qbank.
How many NEJM Knowledge+ questions should I complete per day for ABEM? Zero, because there are no NEJM Knowledge+ EM questions to complete. On whichever EM bank you do choose, a sustainable target is roughly 40 to 60 timed items per day in the core months, front-loaded into your weakest EM Model categories, with pace and error-type tracking mattering more than raw volume.
When should I stop using NEJM Knowledge+ and move to mixed mocks? For ABEM the question does not arise — you were never on a NEJM Knowledge+ EM feed to leave. On your chosen EM bank, move from topic-filtered drilling to timed mixed blocks once every EM Model category has cleared its coverage floor and your first-attempt accuracy has been stable for two to three weeks.
How should I combine NEJM Knowledge+ with iatroX without duplicating practice? For ABEM you would not combine them, because NEJM Knowledge+ has no EM bank to combine. If you were sitting an internal-medicine, family-medicine or paediatrics board, you would keep NEJM Knowledge+ as the adaptive learning tool and use a second, unseen bank only for measurement. For ABEM, pair a dedicated EM bank with iatroX for unseen transfer testing.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Question counts, product scope and prices are vendor-reported and change without notice; verify the current details on each product's own page before relying on them. Disclosure: iatroX operates a competing question bank, so this audit confines iatroX's role to a job NEJM Knowledge+ does not claim to do for EM — supplying unseen, blueprint-stratified ABEM items for transfer measurement — and takes no position on NEJM Knowledge+'s internal-medicine, family-medicine or paediatrics products beyond noting they are not EM. Corrections are welcome via the feedback route on iatrox.com. References: ABEM Qualifying Examination content specifications and the Model of the Clinical Practice of Emergency Medicine (abem.org); NEJM Knowledge+ and AMBOSS product pages (knowledgeplus.nejm.org; amboss.com); iatroX ABEM bank (iatrox.com/abem-emergency-medicine); "Your Q-Bank Percentage Is Not Your Exam Score" (iatrox.com/blog/qbank-percentage-not-your-exam-score); and the blueprint-coverage matrix guide (iatrox.com/blog/question-bank-completion-is-not-coverage-how-to-build-a-blueprint-coverage-matrix-for-any-medical-exam).