Daniel M. Knowles
M.S. in Statistics Candidate
Dept. of Mathematics & Statistics
University of Southern Maine
Abstract: Reducing re-hospitalizations among Medicare beneficiaries has become a high priority for policymakers and the Centers for Medicare & Medicaid Services. Hospital readmissions are seen as an important indicator of care quality and account for billions of dollars in annual Medicare spending. In this work we search for predictors of 30 day readmission for Medicare fee for service beneficiaries age 65+ whose initial hospital admission is due to acute myocardial infarction (AMI). We develop predictive models that may aid our subject hospital in identifying predictors and stratifying an AMI population as to risk of readmission within 30 days of discharge. We utilize logistic regression and classification decision trees as model methodologies to arrive at a ‘best’ model to predict readmission. We compare final model candidates utilizing common model performance metrics.
For further information about this presentation, please contact Dr. Cheng Peng at 780-4689.In compliance with the Americans with Disabilities Act, if you need accommodations to attend this seminar, please call Laurie McDermott at 780-4608 or Fax at 780-5607.