i guess pretty much everything can be linked to integrable models, as long as you have some differential equations. Prime examples are QM, conformal field theory, statistical mechanics, topological defects etc.
I'm not sure it is important for physics. If we did not know how to solve models exactly using e.g. the Yang-Baxter equation, Bethe Ansatz etc., we would still be able to deduce almost everything we know today about models in general using RG arguments, conformal symmetry etc. etc.