In my (limited) experience, programming classes, especially intro level, often end up with a binomial-ish distribution anyway. I was casually assisting some research on why this is when I was helping teach labs and such so was interested. I'm sure more research happened after I wasn't doing that any more, but I remember the best way of removing this at the time was catchup classes.

A lot of intro programming builds directly on previous lessons, much more so than, e.g., maths. If you missed how variables work (off sick, just didn't get it, whatever), you're still stuck when it comes to functions and anything else following and then you're going to fail - it was quite predictable. We studied other university courses and nothing came close to the pattern we were seeing, except "computing for chemistry" or something, which was basically the same sort of course just in a different department.

So we added explicit catch-up classes a few days after a topic was covered so if you missed it, you could get quite personal help on getting back up to speed. This really shifted the distribution to the right, then the people who failed were either those who just didn't care, or those under more extreme circumstances where this couldn't help (or those who just could not learn programming for love nor money but that was rare ime.)