I according to my Stava/Trailforks log, I have done about 1000 rides since I started riding again in 2012 after a hiatus. In that time, I have been to the ER twice, both times for stitches. Once was practicing for an enduro on jagged terrain in the Okanagan, and once was while trying to go fast on Asian Adonis when it was brand new and 6" wide. Over those ~1000 rides, I have not taken especially great measures to mitigate my risks. I have ridden within my limits, but have definitely pushed myself to progress. And I have crashed many, many times.
I would say 1000 rides is a decent sample size. Using my own records, I run a 0.2% chance of ending up in the ER on any given ride.
To me, that risk seems manageable, even more so if I avoid really pushing it over the next bit. Both my ER trips were crashes at speed, so I will especially be avoiding trails that allow speeds to pick up.
Of course, 0.2% chance of ending up in the ER doesn't mean it it happens on schedule 2 out of 1000 rides. It can happen at any time. But based on probability, continuing to ride right now while targeting a few aspects of riding that increase my risk seems like a reasonable thing to do.
Also, for comparison, I ended up in the ER once over the same period cutting my finger while cooking.