Power analysis for stepped wedge trials with delayed treatment initiation and longitudinal measurements