# Spaced Repetition for USMLE Flashcards > How do you use spaced repetition for USMLE flashcards? A working method for spaced-repetition flashcards: what belongs on a card, how question practice feeds the deck, and how to clear a backlog before test day. Spaced repetition schedules each flashcard for review just before you would forget it, so every repetition is an act of recall rather than recognition. Make one card per fact the moment a question exposes a gap, answer it from memory, grade yourself honestly, and let the FSRS algorithm set the next interval. Review daily, without exception. ## How do you use spaced repetition for USMLE flashcards? Spaced repetition is a review schedule built from your recall history, not a calendar: cards you answer confidently return in weeks, cards you fumble return tomorrow. Passive re-reading fails at USMLE volume because rereading produces fluency, not memory, and fluency collapses under a vignette. Testing yourself on studied material yields markedly better delayed recall than restudying it ([Karpicke & Roediger, 2008](https://pubmed.ncbi.nlm.nih.gov/18276894/)), and spacing those tests out beats bunching them ([Cepeda et al., 2006](https://pubmed.ncbi.nlm.nih.gov/16719566/)). **FSRS**, the Free Spaced Repetition Scheduler in current flashcard software, models each card's stability and difficulty from your grading history and schedules its next review for the day your predicted recall falls to your retention target. > **If you only remember one thing:** a card you recognize is not a card you know. Seeing the answer before producing it wastes the repetition. ## The forgetting curve is what the intervals are fitted to Memory for a new fact decays fastest in the hours and days after learning, then more gradually. Each successful retrieval flattens that curve, so intervals expand: a card recalled today returns in days, then weeks, then months. A fixed daily loop wastes the flat part; cramming spends every repetition inside the steep part and leaves nothing scheduled after it. Active recall is the ingredient that makes the schedule pay off, because an expanding interval helps only when each repetition is a retrieval attempt with the answer hidden. Flipping through cards you read rather than answer buys spaced *exposure*, leaving the scheduler to plan around a memory strength you do not have. ## What belongs on a card, and what never does One card tests one fact, phrased so exactly one answer is correct. "Describe the nephrotic syndromes" is a study session; "which nephrotic syndrome shows spike-and-dome subepithelial deposits?" is a card. Put **cloze deletions on the tested term**, never on a random word: hide the drug, the enzyme, the deposit pattern, the cutoff. Mechanism cards work best when they link cause to consequence in one line, so the card rehearses the reasoning a vignette demands. Never paste a question explanation or textbook paragraph onto a card, because a long card cannot be graded honestly. Split a whole table into one card per row or contrast. ## Make a card the moment a question exposes a gap Question practice is the best card-writing prompt available, because it exposes gaps you did not know you had. Write the card during review, in your own words, before the explanation fades. That habit is the retrieval half of the workflow in [how to review UWorld incorrect questions](/blog/how-to-review-uworld-incorrect-questions), and it keeps [pharmacology](/topics/pharmacology) cards attached to real stems rather than drug lists. Deleting is the other half: a card you can no longer justify was written badly. Suspend it, since review time is the scarce resource. ## Worked example: one missed question, two good cards A stem describes a patient three months into tuberculosis treatment with new numbness and tingling in both feet. You answered ethambutol; the answer is isoniazid, which depletes pyridoxine. That miss yields two cards. - **Card 1, the gap:** "Isoniazid causes peripheral neuropathy by depleting {{pyridoxine, vitamin B6}}; give B6 with it." - **Card 2, the discriminator:** "Which anti-tuberculous drug causes optic neuritis with red-green color desaturation?" -> "Ethambutol." Most students skip the second card, yet it fixes the real error: ethambutol filed under the wrong toxicity. ## A scheduler that reads your question data finds weak spots for you Card selection is what manual deck-building does worst: the gaps you notice are not the gaps costing you points. StepGenie's **20,000+ flashcards** are mapped to the same concepts as its question bank, so a missed question surfaces that concept's cards automatically, on the FSRS schedule. ## Passive re-reading vs active recall vs spaced repetition The three study modes differ on the dimensions that decide a board score. | Dimension | Passive re-reading | Active recall | Spaced repetition | | --- | --- | --- | --- | | **What you do** | Reread and highlight text | Answer from memory, source hidden | Answer from memory the day you are about to forget | | **Time investment** | Low per pass, passes never stop | Higher per item, effortful by design | Lowest to sustain, since intervals lengthen | | **Retention** | Fades quickly; builds familiarity | Substantially more durable at delayed tests | Most durable of the three | | **Where it fails** | Feels productive while retention decays | Over-tests what you already know | A missed week leaves a backlog | | **USMLE applicability** | Weak; the exam tests retrieval, not recognition | Strong within one subject | Strongest across months and thousands of facts | ## What a deck costs per day, and how to stop a backlog Daily review load is driven by how many new cards you add, not by deck size, so the number to control is your new-card cap. At roughly ten seconds per card, realistic only for one-fact cards, two hundred due cards is about half an hour, which fits the flashcard slots in a [Step 1 study schedule](/blog/usmle-step-1-study-schedule). Work down a backlog in this order: 1. **Set new cards to zero** until the queue clears. 2. **Clear the oldest-due cards first**, since those have decayed most. 3. **Never bulk-reset the queue**, which destroys the scheduling history. 4. **Suspend low-yield cards** rather than promising to reach them. ## Common flashcard mistakes and their corrections Three habits waste more review time than any scheduler setting. - **Grading generously.** Marking a card "good" when you needed the first letter teaches a stability you lack. Correction: no unprompted answer, no pass. - **Building cards from lectures, not errors.** Correction: write from missed questions and concepts you cannot explain aloud. - **Skipping reviews to fit more questions.** Correction: reviews are the interest on hours already spent. ## Sources - [Cepeda et al., distributed practice meta-analysis](https://pubmed.ncbi.nlm.nih.gov/16719566/) - [Karpicke & Roediger, retrieval practice and long-term retention](https://pubmed.ncbi.nlm.nih.gov/18276894/) Stop building a deck by hand as your exam date closes in. [Try StepGenie spaced repetition flashcards -- 20,000+ cards synced with your QBank](https://dashboard.stepgenie.app/sign-up), and let your missed questions decide tomorrow's review. ## Frequently asked questions ### How many new flashcards should you add per day? Set a new-card cap you can sustain on your worst day, because every card you add generates reviews for weeks afterwards. The right number is whatever keeps your total daily reviews inside the time you have blocked for them. If reviews are routinely spilling past that block, lower the cap rather than skipping days, since skipped days compound into a backlog far faster than a low cap costs you coverage. ### When should you start spaced repetition before Step 1? Start as early in the pre-clinical years as you can, because the intervals need months to stretch out and deliver their advantage. Starting a large deck four weeks before the exam mostly buys you a backlog. If you are already inside your dedicated period, do not import thousands of cards; make cards only from questions you miss, and let the question bank define your deck. ### Are premade decks or your own flashcards better for the USMLE? Both have a job. A premade deck gives you coverage of high-yield material you have not met yet, which is hard to write for yourself. Cards you write from missed questions fix the specific gaps costing you points, and you remember them better for having written them. The practical combination is a curated premade deck for breadth plus your own error-derived cards for depth. ### What is FSRS and how does it differ from older schedulers? FSRS, the Free Spaced Repetition Scheduler, is the algorithm behind current flashcard software. Older schedulers such as SM-2 multiplied each interval by a per-card ease factor nudged up or down by your grades. FSRS instead fits a model of memory stability and difficulty to your own review history for each card, then schedules the review for the day your predicted recall falls to the retention target you choose. ### How do you clear a flashcard backlog before the exam? Set new cards to zero until the queue is empty, then work the oldest-due cards first, since those have decayed the most and teach you the most. Suspend low-yield cards permanently instead of promising to reach them later. Never bulk-reset or mass-fail the queue to make the number disappear, because that erases the review history the scheduler depends on to time your next repetition. --- Canonical page: [Spaced Repetition for USMLE Flashcards](https://www.stepgenie.app/blog/spaced-repetition-usmle-flashcards) Topic hub: [Study Strategy](https://www.stepgenie.app/topics/pharmacology)