hi,
Obviously polling for the ready bit is more time efficient

Not always. For example, some my programs use next way.
There is my own _LCDprints() which accepts string pointer as input. This function prepares some data and set ups timer 2 to interrupts at each 0.1 ms. Then it runs timer and quits. It is all. ISR of timer 2 puts chars by itself each 100us till NULL-terminator has been reached.
By this way a program does pay time neither for polling nor in dummy loop. By programming terms, it creates separate task and continues work in own one.

Regards,
Oleg