Well, I was able to artificially raise the CTS line, which is active low, and halt the transmissions, so I don't think the problem is with the UARTE on the nRF52x. Rather, I think the problem may be with how virtual com ports are handled on PC's, and maybe they aren't configured to properly use RTS and CTS for flow control. At any rate, I was able to crank the UARTE transmission speed all the way up to 941176 baud, and at that speed it takes only a very small artificial delay (around 300usec) after each transmitted string to avoid problems, so I'm just going to do that and move on rather than chase down how to get perfect RTS/CTS working on a PC's virtual com port. So, for that reason, this PCB project will still have value when its finished by keeping the "wiring" all within a single PCB rather than strewn about externally with a rat's nest of actual wires.