I am a tinkerer rather than a professional at this. I would recommend one of the following people:
Paul Edstein (posts as macropod)
Graham Mayor
Doug Robbins
or Greg Maxey

Specialized "real-time" transcription software is generally much more suited to what you are attempting unless you are using a standard keyboard rather than a steno keyboard.

Also investigate foot switches that attach to the computer. Such a switch could be used to trigger the font change, as well.

I do not know of any way to intercept the Enter key directly. Further, do you not want to use word-wrapping? Word is not well adapted to using the Enter key at the end of every line.

I would suggest rather than every line being time-stamped that every new thought/voice be time-stamped.