It is interesting that some manufacturers seem to market active antenna as their recommended product for radio mic receive antenna. Is the correct application to make up for cable losses, not to compensate for rf path loss?
Are there any rules of thumb for how far to keep transmit and receive frequencies separated by in a spectrum band plan?
It is, indeed, only suitable to make up for cable losses and/or passive components like splitters, and is unnecessary for simpler systems. The LPDA alone has +5dBi of gain, and if you use a helical, its gain is roughly 8-9 dBi (11-12 dBi gain minus 3dB polarization mismatch loss). If you use high quality 3/8" low loss cable your losses are in the neighborhood of 3.5dB to 4.5dB per 100ft at the middle of the UHF-TV band (585MHz).
The question of band spacing is more difficult to answer. Every urban area in the USA has different areas of the UHF-TV band available for use, so we are forced to be very flexible in our band planning. You really just try to keep them as far apart as is practical and do your best to stick to it. Unfortunately, someone almost always shows up late with a piece of gear that will only tune to a band where you don't want that type. You just have to make it work as best you can.
Having a good spectrum analyzer helps very much to examine the spectrum at potential assignment frequencies to see if they have low enough noise, and to determine if predicted IMD products are really being transmitted by your equipment, although you are rolling the dice on what's happening inside the front ends of your receivers. This is where expensive high-end gear helps, because its receiver channel filtering helps to reduce internal IMD products dramatically.