Evidence

How many cases until you're actually good at it?

The orthobro · 2026-08-31 · 2 min read

Someone asks this in every residency, usually somewhere in the second year. How many do I need to do before I am actually good at this?

The literature will hand you a number. It should not.

A study that went looking for the plateau

Six orthopaedic surgeons, 90 robot-assisted total knee replacements, one institution. Operating time fell steadily across the run: 97.6 minutes over the first thirty cases, 86.6 over the second thirty, 76.7 over the third.

No plateau appeared within those 90 cases. The curve was still going down when the data ran out.

The number moves depending on how long anyone watches

Most "cases to proficiency" figures come from CUSUM analysis. A 2023 study in Surgical Endoscopy ran that analysis on the same surgeons twice — once across their first 50 cases, once across their first 100.

The case number where proficiency appeared went up when they added the extra data. Same surgeons, same operations, different answer.

The reason is mathematical rather than surgical. CUSUM scores you against the average of whatever sample you happen to sit in, so the peak is, in the authors' words, "nothing more than the case number when the console time reaches the average console time." Watch someone for longer and their average shifts, and the milestone slides out ahead of them.

The authors are also blunt about what the method cannot see: patient complexity, the team in the room, and whatever the surgeon walked in with.

The finding worth stealing

That same study turned up something that was not an artifact. Surgeons who got through their early cases faster in calendar time went on to show less variability in their operating times.

Not more cases. The same cases, closer together. Twelve knees over three months is not the education that twelve knees over two years is, and both of them log as "12".

The uncomfortable one

Outside surgery, a 2014 meta-analysis pooled the research on deliberate practice and performance. Practice explained 26% of the variance in games, 21% in music, 18% in sports, 4% in education — and under 1% in professions.

That number is easy to over-read, so be careful with it. It describes how much of the difference between people is explained by logged practice. It does not say practice fails to make you better. And none of the pooled domains were surgery.

Still — if you are keeping score by counting cases, that is roughly the bet you are placing.

So count something else

The case number is a proxy nobody has managed to make behave. What it stands in for is specific and knowable: which steps of this operation you can do unassisted, today.

Write that down after the case, not at the end of the rotation. Two lines is enough — the step you led, and the step you handed back.

OrthoVaultOS keeps notes against each procedure for this, so the honest answer from last time is sitting there the next time you open it before a case. Six weeks of that tells you more than any case count, because it is a list of what you cannot do yet, and that is the only list that gets shorter.