OBJECTIVES: To illustrate a graphical approach for summarizing patterns of drug use and resource utilization surrounding clinical outcomes. METHODS: Partitioned bar charts, displayed on horizontal axis at times of clinical measurements, display percentages of patients taking different drug classes prior to each clinical measure. Superimposed on charts are plots of average clinical values, age, and duration/visits since diagnosis. Bootstrapping is used to provide confidence intervals for clinical and utilization estimates. RESULTS: The clin-o-gram will be illustrated using demographic,prescription, and longitudinal blood pressure measurments on hypertensive patients from the General Electric Electronic Medical Record captured in the Centricity Physician Office system. The descriptive clin-o-gram plots provide clues for subsequent inferential analysis including marginal mean and pooled logisitc regression over vists. CONCLUSION: With increased availability of clinical outcomes comes need to visualize patterns of drug use relative to changes in the outcomes. Statistical models of changing clinical measures as function of changing drug utilization need to be preceded by graphical summaries.