This workflow is for GP trainees who have bought, or are considering, eMedica for the MRCGP AKT and want a disciplined sequence rather than a pile of features used at random. eMedica's domain-weighted bank and full 160-item mocks cover the applied-knowledge core well; its principal limitation is that it will not organise your week for you. Left to default settings, most candidates run four versions of the same passive activity — reading, re-reading, re-attempting and re-marking — and mistake motion for progress.
What eMedica offers for the AKT right now
Vendor-reported, last checked 19 July 2026 — verify on the eMedica product pages:
- Bank: over 2,250 questions across the three AKT domains, with concise explanations.
- Mocks: four mini-mocks plus a full 160-question Mock A included; Mocks B and C sold separately (vendor-reported around £30 each, or about £69 bundled).
- Access and price: vendor-reported from about £49 (one month) to about £129 (twelve months), with tiers in between.
- Feedback and analytics: per-item explanations and basic topic feedback; no adaptive sequencing, no AI tutor, no spaced-repetition scheduler.
That feature set is enough to build a complete workflow — provided you supply the sequencing the platform does not.
The exam anchor
The AKT is 160 single-best-answer questions in two hours forty minutes, four sittings a year at Pearson VUE, weighted roughly 80% clinical / 10% evidence-based practice / 10% organisational (RCGP, 19 July 2026), at about a minute per item. Everything below is organised to hit that target: broad coverage, current UK guidance, and recognition speed under time.
Give every feature exactly one job
The core move is to stop four features from doing the same thing. Assign one job each:
| Feature | Its one job | Not its job |
|---|---|---|
| Explanations/notes | Teach a specific gap just in time | Being your primary reading source |
| Topic question sets | Test and drill recognition | Passive re-reading |
| Full mocks | Simulate pace across 160 items | Everyday practice |
| Topic feedback | Diagnose where to point the above | A score to admire |
| A second, unseen bank | Measure transfer | Duplicating eMedica's items |
If a feature is doing two jobs, one of them is being done badly. Notes are for teaching a gap you just exposed, not for reading cover to cover; mocks are for simulation, not casual daily use.
Start with a blueprint-stratified baseline
Do not follow the platform's default order blindly. Begin with a baseline that samples all three domains proportionally — roughly 80/10/10 — so your first data reflect the exam's shape, not whichever topic eMedica lists first. Use that baseline to choose your first modules. A blueprint-coverage matrix turns the baseline into a map of where to spend hours, and stops you over-revising comfortable topics.
Build the weekly sequence: learn, test, retest, integrate
Run a simple four-stage loop each week:
- Learn only identified gaps — use eMedica's notes on the specific topics your baseline or last week's errors flagged, not everything.
- Test them with eMedica topic sets, committing answers before reading explanations.
- Retest later — the same ideas on fresh, unseen questions after two to five days, so you measure recall, not recognition.
- Integrate the corrected topics into a mixed, timed block that ignores topic order.
This loop, repeated, is the whole workflow. Its discipline is that nothing gets ticked off after a single correct attempt; a topic is "done" only when it survives a delayed, unseen retest.
Protect your assessment assets
Unseen questions and full mocks are finite, and their value is that you have not seen them. Do not consume Mock A, B or C during casual study, and do not burn your second-bank items by using them for learning. Ration full mocks to a handful across the cycle, taken cold and timed. Once you have seen a question, it can only ever test recognition again — so spend your unseen stock deliberately.
Set switch criteria based on observed errors
Decide in advance what moves you between resources, based on the error you actually see:
- Wrong because you did not know the rule → switch to notes (teach), then test.
- Wrong because you could not retrieve it under time → switch to timed mixed blocks (drill pace).
- Wrong because the guidance has moved → switch to the primary source (NICE, CKS, SmPC/eMC) and update your note.
- Wrong because of misreading → switch to a slower, deliberate re-read protocol, not more volume.
- A whole domain stays low on unseen items → switch to a second bank for that domain's breadth.
Matching the fix to the error type is what stops you defaulting to "just do more questions," which only helps one of these five failure modes.
Exit criteria
Stop the acquisition phase and move to consolidation when four things are true: coverage is broad across all three domains on a real matrix; first-attempt accuracy on unseen items is stable at or above the exam standard; your pacing is correct (you finish full mocks with time in hand); and your errors are knowledge gaps rather than timing failures. Completion percentage is not on that list.
Three mistakes this workflow is designed to stop
The one-job-per-feature discipline exists to prevent three specific failures:
- Running four passive activities at once. Reading notes, re-reading notes, re-attempting seen questions and re-marking them can fill a whole evening while testing nothing new. Each feature must do a different job — teach, test, simulate or diagnose — or three of the four are wasted motion dressed up as revision.
- Admiring the score instead of using it. A rising percentage on questions you have seen before is mostly recognition, and it flatters you into stopping too early. The number that matters is unseen first-attempt accuracy; the eMedica dashboard is a diagnostic pointer to where you should spend hours, not a trophy to polish.
- Spending assessment assets as study material. Sitting a full mock in week one, or grinding your unseen second-bank items for learning, destroys the very thing that made them useful — their unseen quality. Ration mocks and protect unseen stock, because a question you have already seen can only ever test recognition again.
The thread connecting all three is calibration. Every time you learn and test on the same seen items, you inflate your sense of readiness; every time you learn on one bank and prove it on unseen questions from another, you deflate that inflation back to something honest. The workflow is really just a set of guards against fooling yourself — which, three or four weeks out from a 160-item paper at roughly a minute each, is the failure mode most worth engineering against.
A worked seven-day plan for a busy trainee
Four clinical days, short evenings, one weekend session. Give eMedica one job: learn-and-drill your two weakest domains.
- Monday: 30 eMedica items in your weakest domain; commit answers first, then read only the explanations for errors.
- Tuesday: re-write the corrected rules in your own words; queue them for a delayed retest.
- Wednesday: a fresh, unseen iatroX block on the same domain — never a replay — to measure transfer on questions you have not seen.
- Thursday: a mixed, interleaved eMedica block across all domains at exam pace.
- Friday: rest, or a ten-minute misconception-log review.
- Saturday: one full eMedica mock, cold and timed, if you are within four weeks of the exam; otherwise more mixed blocks.
- Sunday: plan next week from the delayed retests that failed.
No proprietary-algorithm claim is needed: you learn and drill on eMedica and prove it on a separate, unseen bank. That is the two-Q-bank rule, and it keeps your readiness signal uncontaminated by familiarity.
Decision checklist: continue, supplement, switch or stop
- Continue while your unseen first-attempt scores rise and coverage broadens.
- Supplement a single stubborn low domain with unseen items, not more re-reading.
- Switch the primary bank only for a measured coverage gap, never for novelty.
- Stop buying more mocks once pacing is solid and errors are knowledge gaps — more mocks will not teach content.
Frequently asked questions
Is eMedica enough for MRCGP AKT on its own? For a well-prepared UK trainee, eMedica's domain-weighted bank plus the official RCGP materials can be a sufficient primary resource, and its full mocks handle the pacing rehearsal well. But "enough" is defined by stable, unseen first-attempt performance across all three domains, not by finishing the bank. If your transfer score on questions you have never seen sits comfortably above the standard, one bank is enough; if a domain lags, add unseen volume there rather than re-reading.
Which MRCGP AKT component does eMedica not reproduce well? The platform reproduces clinical recognition strongly but offers less scaffolding for the interpretive edge of the two 10% domains — applied statistics and current practice administration — because concise explanations state the rule without always training the calculation or the reasoning. It also does not automate the diagnose-to-retest loop, so the workflow around the content is left to you. Verify those slices on a coverage matrix and top them up with unseen items.
How many eMedica questions should I complete per day for MRCGP AKT? A sustainable target for a working trainee is roughly 30–50 timed items on study days, with the emphasis on properly reviewing and retesting errors rather than maximising raw volume. Thirty questions you convert into corrected, retested rules beat sixty you never revisit. Scale up only in the final few weeks, and only if your review discipline holds under the extra load.
When should I stop using eMedica and move to mixed mocks? Transition to predominantly mixed, full-length, timed mocks in the final three to four weeks, once coverage is broad and your errors have shifted from "did not know" to "misread under time." Ration the full mocks so you enter that window with unused papers to sit cold. If whole topics are still untouched, keep building coverage before going mock-only.
How should I combine eMedica with iatroX without duplicating practice? Give each a distinct job: eMedica to learn and drill a topic to standard, and a free UK-core iatroX block to measure transfer on unseen items you have never met. Never mirror the same questions across both, because the whole point of the second bank is an honest readiness signal — and a Q-bank percentage only means something when it is unseen. One bank teaches; the other tests.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; eMedica's question counts, prices and access periods are vendor-reported and change frequently, so verify them on the product page before purchase. Disclosure: iatroX operates a competing MRCGP AKT question bank; this workflow confines iatroX to the unseen-measurement job that eMedica does not claim, and does not present it as a replacement for eMedica's bank or mocks. Corrections are welcome via the feedback route on iatrox.com.
References: RCGP, Applied Knowledge Test, rcgp.org.uk/mrcgp-exams/applied-knowledge-test; eMedica MRCGP AKT product pages, emedica.co.uk/aktmock.html and courses.emedica.co.uk; iatroX, "The Two-Q-Bank Rule"; iatroX, "Your Q-Bank Percentage Is Not Your Exam Score"; iatroX comparison hub.