Data exists that indicate oral positive inotropic agents have a negative effect on mortality. However, evidence suggests that this conclusion may not apply to intravenous inotropic agents. Further, clinical differences in the action of these inotropes may lead to variations in their effect on mortality. OBJECTIVE: This study assesses and compares the odds of in-hospital mortality for patients who received either amrinone, dobutamine, or milrinone. METHOD: The data source was a hospital database of inpatient claims (1995-1996) from 40 U.S. hospitals including information on diagnoses and procedures, costs and discharge, drug therapies, as well as patient and hospital characteristics. Study eligibility criteria require that the patient had a diagnosis of congestive heart failure or underwent a heart by-pass procedure, received at least one unit of an inotropic agent, and received only one type. Baseline comorbidity measures were constructed and used as controls for confounding covariates such as patient and hospital characteristics, condition severity, complexity of drug therapy, insurance status, and treatment type. The likelihood of in-hospital mortality was analyzed using logistic regression. Odds-ratios are assessed and compared for each patient cohort. RESULTS: The logistic model correctly predicted 84% of the 461 observed deaths in 3,070 patients selected for analysis. The results correctly predicted the effect on mortality due to increased age (positive), condition severity (positive), complexity of drug therapy (negative), aggressive drug therapy (negative), and a number of life-threatening comorbid conditions (positive); the three inotrope cohorts indicate a significant difference in in-hospital mortality. Milrinone patients were almost twice as likely to survive the in-hospital stay than dobutamine patients. Further, dobutamine patients were twice as likely to survive than amrinone patients. CONCLUSION: In-hospital mortality varied significantly by inotropic study cohort. Patients on milrinone had the highest likelihood of survival while patients on amrinone had the lowest.