In my experience, women are much less likely to be "cowboys". Alternatively, I've seen few women surgeons routinely attempt riskier corrective surgeries. It could be more about patient selection than skill.
In the study, 1,000,000 patients were treated by a male surgeon, and 100,000 treated by a female surgeon.
That's quite the discrepancy. Doesn't explain the results, but it does show that there were far more male surgeons than female. Which might mean that there is a selection bias somewhere in the process.
Having on one side the top 1,000,000 male surgeons and on the other side the top 100,000 female surgeons makes a difference, which is really hard to measure. Of course both are not the top surgeons, but it is just harder to find more of a kind. Imagine looking for 100,000,000 male surgeons, which is probably impossible given the education demographic in the US.