Detecting Fraudulent Suppliers of Incontinence Briefs

Abstract

An unsupervised fraud detection approach that flagged anomalous companies was applied to Medicaid claims data. Durable medical equipment suppliers of incontinence briefs were examined for suspicious behavior using claims data from January 2005 to December 2009. Twelve statistical variables were created and applied to individual companies. For every outlier range a company fell into, it was assigned a point. After examining all variables, the points were tallied. High scoring companies were referred to an investigative unit. The method flagged six suspicious companies, of which half were confirmed by the surveillance unit as identified fraudulent cases. The ability of this statistical approach to detect fraud validates the method as a valuable tool in pre-investigative analysis of durable medical equipment suppliers. This method allows evaluators to scan large populations of insurance claims in search of unusual provider behavior. The net payment amount of the flagged companies totaled $449,100, or about 5.9 percent of total payment made to fund incontinence supplies for Medicaid patients.