Use of evidence-based therapies after discharge among elderly patients with acute myocardial infarction

Figure 2: Physician characteristics associated with the temporal rate of increase in the use of β-blockers (A, B) and statins (C, D) from 1992 to 2005. The plotted probabilities of medication use after discharge were derived from the fitted multilevel regression model and represent the predicted probability at an average hospital, for a patient of average age, all of whose risk factors were set to absent, and whose values for the other physician and hospital characteristics were set to their reference levels.