[hyperpape]

To give you good advice, we'd need to know more about you. You haven't really said that much about why you think you're bad, or what you've tried to do to improve.

You say four things about why you're bad:

  am not even really that great with the "practical" stuff despite having been at it for so long

  I must stress the 'surface level' part

  can recall my last interview tripping up on questions about very basic OOP

  although my solution is more likely to be a complete mess

The first seems relevant, but too vague to get much of a grip on. The second seems of dubious relevance--it can be useful to know about a lot of technologies, but it's neither necessary or sufficient. The third is about interviews. That may hurt you in a practical way (and is something you can practice!), but doesn't actually bear on whether you're a good programmer. The fourth is relevant, and actionable. But why do you think your solutions are messes, and what have you done about it?

Supposing that you do really have a problem, you have to decide what your real issues are. Then you have to identify ways to work on them.

Ideally, you'd find someone you trust who can work with you, who knows something about coaching people, and can give you an accurate assessment.

P.S. All this assumes that you have a problem. I don't think we can assume that. It also assumes you want to stick with programming. But that's not a requirement. There are plenty of good comments about how you could switch fields and be happier. There's no shame in doing that, if you decide it's your best course of action.