I hope this question is not mistreated as a duplicate, because that is not my intention.
The Kerr metric suffices to describe the exterior solution to an axisymmetric spacetime mass distribution of mass $M$, but fails to be matched with an interior solution in a covariant form (several works have been done for a Kerr interior solution, however there is still no covariant line element for matching the solution in the boundary of the horizon event, without using scalar fields or any other formalism rather than the Einstein field equations).
For matching the interior solution of an object, like a star, rotating (with axisymmetrical symmetry) there is a not so famous solution, known as the Hartle-Thorne metric, which can be found in a handful of articles made by Bardeen, Hartle & Thorne in the late 60s (early 70s):
$$ds^2 = -e^{2\nu} c^2 dt^2 - 2f(r) \sin^2\theta \,dt d\phi + e^{2\lambda} dr^2 + e^{2\alpha}r^2(d\theta^2+\sin^2\theta d\phi^2),$$
for $\nu, \lambda, \alpha$ are all functions depending on $r$ and $\theta$ (Legendre polynomials $P_2(\cos\theta)$ if we stay at the first order in the expansion). However, this line element is slightly more complicated than Kerr's, and yet Kerr is not a party everybody wants to attend to.
So my question is, how far from reality I can treat the outside part of a star with Kerr? And how can I confront how much the Hartle-Thorne treatment deviates from Kerr's? How can I expand in power series the Hartle-Thorne solution to model, for example, the event horizon in that space time; so that it improves or even matches Kerr's horizon event?