The Science of Spaced Repetition in Medical School (What the Evidence Says)

What is spaced repetition?

Spaced repetition is reviewing material at increasing intervals — for example after 1 day, 3 days, a week, a month — instead of massing reviews into one session. Each review is a retrieval attempt (you answer before seeing the solution), which is itself a powerful learning event.

What does the evidence actually show?

• The spacing effect is among the oldest findings in experimental psychology, first documented by Ebbinghaus in 1885 with his forgetting-curve experiments. A meta-analysis of 254 studies (over 14,000 participants) confirmed distributed practice reliably outperforms massed practice for retention. [Cepeda 2006, Psychol Bull]
• The testing effect: retrieving information from memory produces better long-term retention than re-studying it, even when re-study time is equal. [Roediger & Karpicke 2006, Psychol Sci]
• Ranking the techniques: a review of ten common study techniques rated practice testing and distributed practice as the only two with "high utility" [Dunlosky 2013, Psychol Sci Public Interest] — above highlighting, re-reading and summarizing.
• In medical education: randomized trials of "spaced education" with physicians-in-training showed durable knowledge gains from spaced, test-based review [Kerfoot 2007], and medical students' use of spaced-repetition flashcards predicted USMLE Step 1 performance. [Deng 2015, Perspect Med Educ]

How does the SM-2 algorithm work?

SM-2 — the algorithm family behind Anki and Finito Medicine — tracks an ease factor per card. Answer correctly and the interval grows (roughly ×2.5 at default ease); struggle and the interval shrinks and the card repeats sooner. The goal is reviewing each card near the edge of forgetting, which maximizes retention per minute of study. Newer schedulers (like FSRS) refine the same principle with memory models.

How to apply this in medical school

1. Convert lectures into questions the same week. The retrieval benefit needs question-formatted material — flashcards, not highlighted slides. (This is the step AI generation automates: Finito builds cards from your PDFs and slides so the conversion cost doesn't stop you.)
2. Do reviews daily, even 15 minutes. Skipped days pile up reviews and break the spacing schedule.
3. Keep cards atomic. One fact per card; long cards become re-reading in disguise.
4. Don't suspend the hard cards. The cards you keep failing are exactly where the algorithm is buying you the most.
5. Trust intervals that feel "too long." Reviewing just before forgetting feels uncomfortable — that difficulty is the mechanism, not a bug.

Common mistakes that cancel the effect

• Re-reading instead of answering. If you flip the card before attempting an answer, you removed the retrieval event.
• Making cards you never review. Coverage without review is just slower note-taking.
• Cramming "to be safe" before the exam anyway. A final consolidation pass is fine; abandoning the schedule for a week of cramming throws away the spacing benefit for next semester's cumulative exams.

Frequently asked questions

How many flashcards per day should a medical student review?

Most students sustain 100–300 reviews per day (20–45 minutes). The honest limit is whatever you can do every single day — consistency beats volume.

Is SM-2 outdated compared to FSRS?

FSRS models memory more precisely and can reduce review counts at equal retention, but the dominant factor is doing spaced retrieval at all. Both schedulers capture the large effect; the difference between them is second-order.

Does spaced repetition work for concepts, or only facts?

It works for both, but concepts need cards that force explanation or application ("why does X cause Y?"), not just recognition. Pair flashcards with practice questions for reasoning-heavy material.