We all know that in order to propagate a trapped optical mode, a simple planar waveguide or step-function optical fiber must have a higher index of refraction in the core than in the cladding.
Planar waveguides and cylindrical (or elliptical) fibers with more complex index profiles can also have trapped or guided modes, however, including some cases in which the index at certain points in their central or "core" region can be lower than the limiting index value far out in the cladding region.
Consider a planar waveguide or cylindrical fiber having an arbitrary transverse or radial index variation across its central or "core" region, which eventually turns into a constant index value far enough out in the (unbounded) cladding region. The guiding properties of such an optical waveguide presumably depend upon some sort of weighted average of the index in the core region, compared to the limiting index value far out in the cladding region.
Query: Is there some kind of general theorem or formula in either of these cases for calculating this weighted average of the index across the core region, compared to the limiting index far out in the cladding region, that will tell us whether or not the structure will support a trapped or guided mode, _without_ actually having to solve for the mode profile itself?
Seems as if there ought to be -- but I've not (in my admittedly limited experience) encountered such a theorem.