I can't answer 'how I get around the problem' because I've never encountered it, but it's an interesting theoretical question.
It might be acceptable for a 16A point load to be applied to a 32A RFC, especially if the diversity applicable or the manner of operation of a permanently-installed appliance can be foreseen and factored in. An oven that requires an average of 15A for a few minutes and then drops back to below 13A average, is not much different from one with a smaller element that draws 12A solidly for a longer warm-up time (It is different, though, because the heating in the circuit cables is proportional to I² but the heating in the oven is proportional only to I.)
We must distinguish between three possible kinds of problem arising from a large fixed load on a general-purpose RFC:
1) Risk of overloading cables, since on a 32A RFC conventionally In>Iz. Normally, this does not happen. If it does, it's very minor, which is why houses don't burn down when RFCs get broken near one end.
2) The load accounts for too great a fraction of Ib to allow the circuit to fulfil its role of supplying general purpose socket outlets. This is not specific to RFCs, it is a simple diversity conflict, that would occur equally on a 32A radial S/O circuit. If the circuit is intended to provide a useful amount of power elsewhere, a 32A circuit with a 16A load is probably better than two 20A circuits with a 16A load on one of them. The bigger the circuit, the better diversity works.
3) The accessories are not suitable for a 16A load. This is simply a practical consideration that is easily overcome by careful selection of equipment.
My view is that it is bad practice to burden a general-purpose socket-outlet circuit with large single loads unless their duty-cycle is known to be modest. I think this is actually a more significant consideration than the point loading of an RFC affecting current distribution.