OBJECTIVE: The State Hospital Association (SHA) in a large Midwestern state annually publishes risk-adjusted mortality rates on selected groups of patients, by hospital. For the third year in a row, the SHA reported a higher than expected mortality rate for Non-Surgical Heart cases at Memorial, and the Board of Directors asked for an improvement plan. METHODS: We used the CaduCIS risk assessment methodology to independently validate the SHA s findings, and conducted opportunity analyses to identify a specific sub-population which seemed to be the major contributor to the mortality rate. RESULTS: (1) Most of the excess mortality seemed to occur in medically managed AMI patients. (2)While a few physicians cared for high volumes of such patients, a surprising number of physicians had very low volumes. (3) Risk-adjusted outcomes for the low-volume physicians were significantly worse than predicted. (4)A non-cardiologist was the attending physician for fully 25% of the medically managed patients. (5)A small percentage (4%) of medically managed AMI cases were admitted to the Neuro ICU rather than the Cardiac ICU but the mortality rate in that subgroup was 58% 11% of the total deaths. (6) Thrombolytic therapy was administered to only 1% of medically managed AMI patients. (7) Beta-blockers were administered to only 58% of patients. Improvement strategies included retargeting educational programs, providing ongoing feedback on mortality rates, and modifying clinical protocols. CONCLUSIONS: (1) For the three-month time period following the implementation of improvement strategies for medically managed AMI patients, the mortality rate decreased from 19% to 8%. (2) No AMI patients were admitted to the Neuro ICU. (3) The proportion of AMI admissions directly managed by cardiologists increased to 84%. (4) The frequency of beta-blocker administration improved to 82%.