This workflow is for emergency physicians using, or considering, BoardVitals for the ABEM Qualifying Examination who want to exploit its adaptive feed without letting it quietly starve small blueprint domains. It addresses the written Qualifying Examination only. The principal limitation to design around: an adaptive feed that prioritises your weak areas by overall performance can hide unattempted, low-frequency domains behind a rising headline percentage — and those domains still appear on the exam.
What BoardVitals offers for ABEM right now
Current state, last checked 19 July 2026; the figures below are vendor-reported (BoardVitals) and change — verify on the product page before you buy.
- Content: more than 900 emergency medicine board-review questions, described as following the American Board of Emergency Medicine content outline and targeting the ABEM Qualifying and In-Training examinations.
- Access and price: subscription tiers of roughly Cram (1 month) $169, Prepare (3 months) $299 and Master (6 months) $419, plus a free trial and a "Buy Now, Start Later" option of up to six months.
- Adaptive/AI: the vendor advertises "Adaptive Learning Technology" and AI-powered risk-assessment analytics that flag at-risk subject areas — treat both as vendor-reported and do not assume a specific algorithm behind them.
- Modes and extras: a timed testing mode to replicate exam conditions, detailed explanations for correct and incorrect answers, around 35 AMA PRA Category 1 Credits as a paid add-on, and a "100% Pass Guarantee."
At roughly 900 questions, this is a moderate bank rather than an exhaustive one, which sharpens the low-volume-domain risk: there is less slack for a personalising feed to cover everything, so the discipline below matters more, not less.
The ABEM Qualifying Examination, briefly
The Qualifying Examination is approximately 305 single-best-answer MCQs, computer-based at Pearson VUE, built on the EM Model — 20 clinical domains crossed with physician-task and patient-acuity axes. Acuity is weighted towards the sick: roughly Critical 30%, Emergent 40% and Lower acuity 21%, with paediatric content at least 8% and geriatric at least 6%, and image or pictorial items throughout. The blueprint concentrates marks in high-acuity presentations, but it also samples the long tail of smaller domains — haematologic and oncologic emergencies, immune and endocrine disorders, psychobehavioural presentations, renal and genitourinary problems, ENT and ophthalmology, environmental and procedural content. Individually small, collectively decisive at the borderline. A vendor's "adaptive" label is a third-party feature, not part of this official specification; keep the two separate in your head.
Baseline week: a blueprint-stratified unseen sample before you personalise
Before you let the feed adapt to anything, sit a small, deliberately stratified baseline — for example 60 to 100 questions drawn to touch every major domain and both high-acuity bands, timed and unassisted. The point is not the score; it is a map. Personalising a feed before you know your real distribution of gaps means the algorithm optimises against a picture that is mostly noise, and it tends to double down on the domains that happen to be over-represented in your first random draw.
First pass: set domain floors so the feed cannot hide a gap
An adaptive feed chasing your overall percentage will happily raise it by feeding you more of what you are already reasonably good at. Counter this by setting a minimum number of attempted questions per blueprint domain — a domain floor — and tracking attempts, not just accuracy. Concretely: commit to a floor for every domain (including the small ones) and refuse to call the first pass complete until each floor is met, whatever the headline number does.
| First-pass control | What it prevents |
|---|---|
| Attempts counted per domain, not just overall accuracy | A rising percentage masking whole domains you have barely touched |
| An explicit floor for each low-volume domain | The feed starving haematology, endocrine, psych, ENT and the like |
| Acuity tracked alongside topic | Strength on stable items hiding weakness in critical and emergent ones |
| A reserved unseen subset, untouched | Having no clean material left for a genuine readiness check |
Error taxonomy: name the failure, not just the topic
When you review, sort every miss into one of six buckets, because each demands a different response and topic alone hides the real fault.
| Error type | Signal | First response |
|---|---|---|
| Knowledge gap | You did not know the fact or step | Short source read, then fresh questions on the topic |
| Misread stem | You knew it but answered the wrong question | Deliberate stem-reading practice on unseen items |
| Premature closure | You anchored before reading the whole vignette | Slow the final read; practise differentials |
| Guideline error | Outdated or wrong management standard | Read the current guideline, not a bank rationale alone |
| Calculation error | Dose, rate or score wrong under pressure | Drill the specific calculation to fluency |
| Time-pressure error | Right method, ran out of time | Pacing drills in timed blocks |
Review interval: not every miss deserves an immediate repeat
Match the response to the error type rather than re-attempting everything at once. A knowledge gap deserves a short source read and then a fresh transfer question a few days later, not an immediate re-answer of the same item (which mainly trains recognition of that item). A misread stem or premature-closure error deserves spaced, deliberate practice on the technique. A guideline error deserves a read of the primary source. Reserve immediate repeats for genuine calculation fluency. Spacing the return of a miss, on a new question, is what tests whether the fix held.
Mixed-block switch: when to reduce topic filtering
Topic-filtered practice is right early, when you are building and repairing domains. Move deliberately towards timed, random, mixed blocks once three objective conditions hold: every domain floor is met, your first-attempt accuracy on filtered blocks has stabilised rather than still climbing steeply, and your errors have shifted from raw knowledge towards pacing and discrimination. Increasing the share of mixed blocks in the final weeks rehearses the real cognitive task — switching domains and acuity cold, under the clock — that topic-filtered practice never trains.
Exit criteria: readiness, not bank completion
You are ready to taper when a set of measurable conditions is met — not when the bank hits 100%.
- Coverage floor: every blueprint domain attempted to its floor, with the low-volume ones explicitly checked.
- Stable first-attempt performance: first-pass accuracy holding across mixed, unseen blocks, not just on reviewed items.
- Pacing: you finish timed blocks with margin at the real per-item rate.
- Retention: spaced re-tests of earlier weak topics still pass on fresh questions.
- An external calibration: a strong result on unseen material outside BoardVitals, so the number is not just familiarity with one bank's style.
A seven-day worked example
BoardVitals does one job here — adaptive, blueprint-stratified building and review — while a second, unseen bank measures transfer. No proprietary-algorithm claim is made for either tool.
- Day 1 — baseline: a 60- to 80-item stratified, timed BoardVitals block; log attempts and accuracy by domain and acuity.
- Day 2 — floors: set domain floors; note which low-volume domains the baseline barely touched (say immune, endocrine, psychobehavioural).
- Day 3 — build: targeted BoardVitals blocks on those under-sampled domains; code every error into the six-bucket taxonomy.
- Day 4 — transfer: a fresh, timed iatroX ABEM block filtered to the same low-volume domains — unseen items, no assistance — to check the learning generalises beyond BoardVitals wording.
- Day 5 — technique: a short mixed block aimed at stem-reading and premature closure; review only those error types.
- Day 6 — space: re-test Day-3 topics on new unseen items; a five-item image and ECG set.
- Day 7 — mix and decide: a 40-item timed random block spanning all domains and acuity bands; apply the decision checklist.
Decision checklist: continue, supplement, switch or stop
Base each call on a measurable gap, not novelty or sunk cost.
- Continue with BoardVitals as your adaptive driver if domain floors are being met, low-volume domains are improving, and unseen transfer results track your BoardVitals results.
- Supplement with an unseen bank if your BoardVitals percentage is high but you cannot yet prove transfer, or if the moderate bank size is leaving specific domains thinly sampled.
- Switch your primary driver if you have effectively seen the bank and are re-answering familiar items, so the adaptive feed has nothing fresh to work with.
- Stop and change method if a domain keeps failing on unseen questions despite meeting its floor and targeted review — change the source, not the question count.
Bottom line
BoardVitals is a usable adaptive bank for ABEM, and its analytics can genuinely surface weak areas — provided you do not let a rising overall percentage stand in for coverage. Set domain floors, track attempts and acuity alongside accuracy, code your errors properly, and prove readiness on unseen material before you taper. The adaptive feed is a tool for sequencing your practice, not a substitute for the blueprint.
Frequently asked questions
Is BoardVitals enough for ABEM on its own? It can carry a well-prepared candidate a long way, but at a vendor-reported bank size of around 900 questions it is a moderate resource, and "enough" depends on your baseline and on whether you can demonstrate transfer to unseen items. Many candidates use BoardVitals as their adaptive core and add a second, unseen bank for breadth and a final readiness check, precisely so a strong headline percentage is corroborated on material the bank did not train them on.
Which ABEM component does BoardVitals not reproduce well? BoardVitals reproduces the written Qualifying Examination's single-best-answer format and offers a timed mode, but it does not reproduce the Oral Certifying Examination, the examiner-led case component that follows a Qualifying pass. It also cannot, by itself, reproduce the sustained multi-hour, two-section endurance of the real sitting from short filtered blocks — you have to build that with deliberate full-length, mixed, timed practice.
How many BoardVitals questions should I complete per day for ABEM? There is no official figure, and consistency plus error analysis beat raw volume. For most working emergency physicians, roughly 30 to 50 questions on a study day, fully coded and reviewed, is sustainable and productive; a much larger count skimmed without triage tends to inflate exposure while teaching little. Distribute the count to honour your domain floors rather than letting the feed concentrate it where you are already strong. (Counts, prices and the adaptive feature are vendor-reported; verify on boardvitals.com.)
When should I stop using BoardVitals and move to mixed mocks? Shift towards timed, random, mixed blocks once every domain floor is met, your first-attempt accuracy on filtered blocks has stabilised, and your errors are mostly pacing and discrimination rather than knowledge — usually the final three to four weeks. Mixed blocks rehearse the real task of switching domain and acuity cold under time, which filtered topic practice never does, so make them the majority of your work as the exam approaches.
How should I combine BoardVitals with iatroX without duplicating practice? Assign each bank a distinct job: BoardVitals for adaptive, blueprint-stratified building and review, and iatroX as the unseen-question layer for timed transfer measurement, so you are never re-answering BoardVitals items to generate a "score." Because the pools are separate, you avoid duplication while gaining a genuine external calibration — the two-Q-bank rule in practice. Point iatroX at the low-volume domains your BoardVitals floors exposed, keep it timed and unassisted, and read its percentage as a study metric.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; product figures — question counts, prices, CME and the adaptive/AI features — are vendor-reported and change, so verify the current details on the BoardVitals product pages before relying on them. Disclosure: iatroX operates an ABEM question bank that competes with BoardVitals; this article confines iatroX to the unseen-measurement and transfer role that a single adaptive bank cannot fill for itself, and makes no proprietary-algorithm claim about either product. Corrections are welcome via the feedback route on iatrox.com.
References: American Board of Emergency Medicine — EM Model and Qualifying Examination content specifications (theabem.org); BoardVitals Emergency Medicine board review product page (boardvitals.com); the iatroX ABEM bank; the iatroX comparison hub; "Your Q-Bank Percentage Is Not Your Exam Score"; and the blueprint-coverage-matrix method.