Statistical Issues in Medical Fraud Assessment

28/2013

Statistical Issues in Medical Fraud Assessment

Monday 24th June 2013

Ekin, T.; Ieva, F.; Ruggeri, F.; Soyer, R.

In this paper we provide a survey of the statistical issues in medical fraud assessment. We discuss different types of medical fraud and the type of fraud data that arise in different situations and give a review of the statistical methods that use such data to assess fraud. We also discuss ”conspiracy fraud” and the associated dyadic data and introduce Co-clustering methods which have not been previously considered in the medical fraud literature. In so doing, we present some recent work on Bayesian co-clustering for fraud assessment and its extensions. Furthermore, we discuss potential use of decision theoretic methods in fraud detection and demonstrate an example for evaluating fraud detection tools.