Text this: An investigation of discrepancies in outcome reporting and selective reporting bias in interrupted time series studies of health interventions