Got it. So I'd still calculate fixed-end moments on each side of the flange transition whether at a support or not and calculate rotations at each transition as well as at supports, then finally calculate moments at ends of each beam segment?
You don't need fixed-end moments at the flange transitions, only at supports.
Or you can do without fixed-end moments altogether. In my Conbeam spreadsheet the procedure used for a continuous beam with simple supports is:
Calculate end reactions and moments along the beam with end supports only.
Integrate M/EI twice to find slope and deflection at each support assuming no rotation at left hand end.
Rotate to bring right hand end back to zero deflection.
Apply unit load to each internal support, and calculate resulting deflections at each support location, assuming single span.
Solve simultaneous equations to find support reactions required for zero deflection at all internal supports.
Recalculate: total reactions, shear forces, moments, curvatures, slopes, and deflections along the beam.
Doug Jenkins
Interactive Design Services