Expected Variables Present Study Results III: PRESERVE

-UFlorida has unusual longitudinal behavior, especially in the procedure domain, with a large spike in patient proportion in the range of 2013-2015 before returning to a more stable baseline shortly after. This could indicate an influx of historical data around this time, but more investigation will be required. - UNC also has somewhat of a spike for procedure data around 2015, but it is more gradual and not quite as pronounced. - Several sites seem to come online at a specific time point when evaluating antihypertensive exposure after cohort entry, which could indicate a data latency issue or a characteristic of the population causing later cohort entry dates.