OBJECTIVES: Using Machine-Learning, this research identifies key medications that lead to reduced readmission rates among patients with diabetes, and provides analysis as to why this is the case. METHODS: Gathered by researchers at the University of California Irvine, the dataset encompasses 10 years (1999-2008) of clinical care at 130 US hospitals and integrated delivery networks. It includes over 50 features representing patient and hospital outcomes. The dataset includes information concerning demographics, diagnostics, prescriptions, existence of diabetes and readmission. I used the Complex Tree method for analysis, using readmission as the independent variable, with the other 49 variables represented in Parallel coordinated plots, selected for their multidimensionality.  RESULTS: Out of the 50 attributes of the given database of 101,766 observations, 25 variables medication, change in medication and diabetic medication, performed inside the boundaries +/-3 standard deviations of readmission mean or no readmission mean for the standardized scaling of the Parallel coordinated plot. Out of the 23 medications only insulin and metformin performed inside the boundary of +/-3 standard deviations (std). For validation, categorical analysis was conducted, and only the same variables were responsible for readmission in the first analytical block.  CONCLUSIONS: It appears that metformin and insulin are associated with readmission than any other drug. This has strong implications for developing effective health care plans who regulate diabetes and hospital readmission.