PS08
Using Decision Trees to Identify Medicare Part B Providers for Audit
Contributed
Noel N. McKetty
First Coast Service Options
,Dr. Donna Mohr
University of North Florida Abstract:
Medicare carriers must efficiently identify providers that should be audited for possible overpayments. Using a historical archive of Part B providers who have been audited in the past, we use Enterprise Miner's regression trees to build predictive models
of the overpayment rate for providers. We compare these results to those of classical stepwise regression techniques. Our goal is to find markers which would identify providers who are likely to have high overpayments.
Biography: Noel N. McKetty has an M.S. in Computer Information Science from the University of
North Florida and is currently Manager of Systems and Reporting at FCSO.
Dr. Mohr earned her Ph.D. in Statistics from Princeton University. She is now an Associate
Professor of Statistics at UNF.