I suppose that by solving a differential equation, you mean to find an exact solution. In many of my research areas, we need to find exact solutions, not just an approximation. Unfortunately, as far as I know there are no other good general methods to solve differential equations of higher order. In fact, even the classical undetermined coefficient method, it only provides approximations in most cases, not exact solutions.
As Bang - Yen Chen has written, there are no general methods for solving exactly such linear differential equations involving arbitrary polynomial coefficients. A general method of approximating the solution is to attach an appropriate first order linear differential system and to apply a fixed point type argument to the latter system.
Yes. I was finding a method to solve LDE involving polynomials as coefficients. But I think solve DE using Power Series isn't a better way as we have to approximate the power series in the solution because " we are not usually able to express power series solutions of differential equations in terms of known functions".
If it is not possible to express the solution in terms of known functions you could'n do it by any other method. In this case you have new functions. But in real life the differential equation generally will be solved numerically..