We demonstrate the use of topological data analysis (TDA) to visualize the dynamics of organizational routines in outpatient medical clinics over the course of the COVID-19 pandemic. We use Electronic Health Record audit trail data from January 2020 to December 2021 to visualize changes in routines in outpatient clinics in four different medical specialties. By representing the pattern of action each day in each clinic as a weighted, directed graph, we see that some clinics bounced between several distinct patterns of action over the two year period. In contrast, other clinics never bounced back to the pre-COVID pattern of action. In all of the clinics, we see evidence of temporal auto-correlation: the pattern of action on any given day is similar to the pattern of action on days that are close in time. Because it can capture recurrence of action patterns that might be widely separated in time, TDA offers a substantial advance in the state of the art for visualizing the dynamics of organizational routines.