Applying sequential surveillance methods that use regression adjustment or weighting to control confounding in a multisite, rare-event, distributed setting: Part 2 in-depth example of a reanalysis of the measles-mumps-rubella-varicella combination vaccine and seizure risk