Anki is the highest-return revision tool available to medical students — and the one most likely to be abandoned within three months. Not because it doesn't work (the evidence for spaced repetition is overwhelming), but because most people use it in a way that's unsustainable: too many new cards, badly designed prompts, and a growing review pile that becomes a source of guilt rather than learning.
Here's how to use it properly.
The settings that matter
New cards per day: 5–15. Not 30. Not 50. The most common Anki failure mode is setting new cards too high in the first week of enthusiasm, then drowning in reviews three weeks later. Every new card generates approximately 7–10 reviews over the following month. Fifteen new cards per day means 100–150 daily reviews within a month. That's sustainable. Fifty new cards per day means 350–500 daily reviews — that's a second job.
Maximum reviews per day: no limit (but manage via new card cap). Don't cap your reviews — that creates a hidden backlog. Instead, control the input (new cards) and the output (reviews) follows naturally.
Learning steps: 1m 10m 1d. The default learning steps are too aggressive for medical content. A card you get right after 1 minute might not survive 10 minutes. Adding a 1-day step before a card graduates to the review queue reduces early lapses.
Graduating interval: 1 day. Easy interval: 4 days. These are the defaults and they work. Don't increase them thinking you'll save time — you'll just see more lapsed cards later.
Lapse steps: 10m. When you get a mature card wrong, it re-enters learning with a 10-minute review step. This is appropriate for medical content where a wrong answer needs immediate reinforcement.
Card design: the make-or-break
Bad cards teach you to recognise answers. Good cards teach you to recall decisions.
Bad card: "What are the side effects of ramipril?" → [long list]. This tests your ability to recite a list, which isn't what exams test and isn't what clinical practice requires.
Good card: "A 62-year-old man on ramipril develops a dry persistent cough. What is the mechanism and what do you switch to?" → ACE-inhibitor induced bradykinin accumulation. Switch to ARB (e.g., losartan/candesartan). This tests clinical reasoning — the same skill the exam tests.
The rules:
One fact per card. If your card tests two things, split it into two cards.
Test the decision, not the list. "What is the first-line treatment for X?" is better than "List all the treatments for X."
Use cloze deletions for factual recall. "The CHA₂DS₂-VASc score awards {{c1::1 point}} for diabetes" is efficient and specific.
Include context. "In a patient with {{c1::eGFR <30}}, metformin should be stopped" — the clinical context makes the card useful beyond the exam.
Delete cards that don't serve you. If a card is consistently easy (you've never got it wrong in 6 months), suspend it. If a card is poorly written and confusing, delete it and write a better one. Anki is a tool, not a collection to preserve.
The daily workflow
Reviews first, always. Before you add new cards, before you study anything else, clear your review queue. This is non-negotiable. The entire value of spaced repetition collapses if you skip reviews.
New cards after reviews. Add new cards from your current study topic. Keep to your daily cap. If you're tempted to add more, don't — your future self will pay for today's enthusiasm.
Total daily time: 15–30 minutes. If your Anki session regularly exceeds 30 minutes, your new card rate is too high or your cards are too complex. Fix the cause rather than spending more time.
Never miss two days in a row. Missing one day doubles the next day's reviews. Missing two days makes the backlog feel unmanageable. If you genuinely can't do a full session, do 10 minutes — maintain the habit even if you can't maintain the volume.
When to use Anki vs a qbank
Anki and question banks serve different purposes:
Anki is for long-term retention of specific facts: drug doses, diagnostic criteria, management thresholds, classification systems. It's efficient for the "you either know it or you don't" component of exams.
Qbanks (like iatroX, PassMedicine, AMBOSS) are for clinical reasoning: applying knowledge to scenarios, interpreting investigations, choosing between management options. They test the "what would you do?" component.
The optimal revision strategy uses both: Anki for factual scaffolding, qbanks for clinical application. Most students should spend 60–70% of revision time on qbanks and 30–40% on Anki — not the other way around.
The burnout trap
If Anki feels like a burden rather than a tool, something is wrong. The most common causes: too many new cards (reduce immediately), too many decks (consolidate into one or two), cards that are too complex (rewrite or delete), or perfectionism about streak maintenance (a missed day is not a failure).
Anki should take 15–30 minutes per day and feel like maintenance, not a workout. If it doesn't, fix the system before blaming your motivation.
iatroX offers an adaptive qbank that complements Anki-based revision — AI-generated explanations and spaced repetition built in. See our spaced repetition guide for the evidence behind the method.