Monday, January 16, 2017

The evaluation of trustworthiness to identify health insurance fraud in dentistry

Available online 27 December 2016

Highlights

The Economist says 10% of health care expenditure ($272 billion) was wasted on fraudulent claims in 2012.
Previous studies aimed at other types of fraud are inadequate for solving cross-dentist health insurance fraud.
We devise a social-network-based risk evaluation to detect fraudulent claims in dentistry.
By experiment, using real-world or artificial data, our method can effectively identify undetected frauds in the past in dentistry claims data.
Integrating our method with previous research can provide higher accuracy.

Abstract

Objective

According to the investigations of the U.S. Government Accountability Office (GAO), health insurance fraud has caused an enormous pecuniary loss in the U.S. In Taiwan, in dentistry the problem is getting worse if dentists (authorized entities) file fraudulent claims. Several methods have been developed to solve health insurance fraud; however, these methods are like a rule-based mechanism. Without exploring the behavior patterns, these methods are time-consuming and ineffective; in addition, they are inadequate for managing the fraudulent dentists.

Methods

Based on social network theory, we develop an evaluation approach to solve the problem of cross-dentist fraud. The trustworthiness score of a dentist is calculated based upon the amount and type of dental operations performed on the same patient and the same tooth by that dentist and other dentists.

Results

The simulation provides the following evidence. (1) This specific type of fraud can be identified effectively using our evaluation approach. (2) A retrospective study for the claims is also performed. (3) The proposed method is effective in identifying the fraudulent dentists.

Conclusions

We provide a new direction for investigating the genuineness of claims data. If the insurer can detect fraudulent dentists using the traditional method and the proposed method simultaneously, the detection will be more transparent and ultimately reduce the losses caused by fraudulent claims.

No comments: