Figure 5: Percent of patients receiving high quality care at baseline and follow-up for each center, by study group and period (weighted averages in bold). The figures depict heterogeneity in both intercepts and slopes, justifying the use of the proposed random effects model.