I think the best thing to do (and I do so with configuration in-ears) is to create a new audio channel strip called eg "master keyboard" and send all channel strips for that bus. The entry would be one of Buses (ex. 1) and the output Output 1-2.
Now create a new audio channel strip called eg "headphones" to enter one of the buses (eg 2) and output Output 3-4.
Now a send from the master keyboard to send the Bus 2 (headphones).
It seems hard but it is really quite easy. Read several times and follow step by step.
I think this is a good way because it has control of everything (volumes, placing inserts, etc.) on each output of your card (master keyboard and headphones and other).
in my case I also have a mic input is one whose output is output 4 (ex) and could be a send to bus 2 (headphones) would be listening, independent levels with a mixture of keyboard and mic.
try it and post your results.