Here's another approach to the dreadful backlog.
Use severity buckets.
When you find an accessibility issue, you don't log it in a list. You put it in a bucket based on how bad it is. Don't overthink your buckets. Go with three: critical, major and minor.
Critical means a user is completely blocked, so you aim to fix it in 48 hours. Major means real friction, so you fix it this sprint. Minor means it's annoying but it has a workaround, so you fix it next time someone's in that code. See fix-on-touch.
That's your system. There's no grooming or prioritisation debates to be had. And your tickets won't rot away for years in a backlog.
There's one thing to watch out for.
That minor bucket can quickly become the place you dump everything. If you want to avoid pressure, all you have to do is call it minor and you don't have to do it. Then suddenly, everything is minor and nothing gets fixed. So this whole thing only works if you're willing to call issues what they are. Be honest. It takes discipline and a bit of courage.
Compared to a backlog, this is much more honest. Buckets force you to make a call on impact.
It's a simple system, which is exactly why it can work.