The optimisation of using the OHLE to the very end is surely small beer compared to the costs if you misjudge it slightly (or forget altogether) and either damage your train or bring the wires down...
Yes, and this has come up in numerous threads, on bimodes, potential BEMUs, and potential electrification add on threads.
Many of the ends of the electrification are historic, just beyond where a fully electric train would reach its shunt limit. This was fine when trains were pure electric only, but ideal wire end locations for multiple mode trains are different and there should be longer run offs (even lengths with unenergised wires etc. for few hundred metres to avoid de-wirements).
Without these extended over runs, an abundance of caution is often used, changing over early. It is the railway using a bodge to get round the problem, rather than fixing the source of the problem.
Whilst there are some bimodes that don't work very well with cold diesel engines, so they get started up early. In most cases it is because the end point of wires (or third rails) have not been extended hundreds of metres to allow seamless transition of power at a location where train is not accelerating.
So in a way a simple answer is the ideal infrastructure modifications were not included when multiple mode trains were ordered, so having to change power modes early to avoid errors and/or problems.