Some research should take priority. I'm not saying that you should immediately go back and remove all your unnecessary lookup tables. You can leave them alone until such time as you have to do maintenance on a form that uses one of them. Then convert only that single table. But start IMMEDIATELY using the mini-app for all new lookups and save yourself from having to create all those new tables and forms. Baby steps. Eventually, you may want to fix all your earlier poor choices but there is no reason to stress about doing them now since that will interfere with moving on. Learn as you go. Adopt new ideas for new parts of the app.
I know we all emphasize consistency. Consistency is your friend. The more consistent your methods and coding style are, the less "thinking" and "decision making" you have to perform on the fly every time you start with a new task. If I come up with a method that is superior to the one I was using, I switch to the new method and then if appropriate, modify the old code when I have to maintain it. I almost never just go back and change old code UNLESS it is an efficiency problem. If a query runs for 10 minutes and I come up with a better method, I will take the time to replace the old version with the new, every place it is used because the payback is worth the effort. Obviously, the the new version fixes an error, that is also worth changing. AND keep in mind, that as you get better at this, it is easier for you to figure out when/how to reuse code/queries. So, if I fix old code that is slow, chances are excellent, I can fix it once and have it impact 20 procedures if the code is generic. If the code is not generic, I might take the time to make it generic now!