Improving Fraud and Abuse Detection in General Physician Claims: A Data Mining Study

Area of Science:

• Health insurance fraud detection
• Medical data mining
• Physician prescription analysis

Background:

• Healthcare fraud and abuse pose significant financial burdens.
• Identifying indicators in general physicians' drug claims is crucial for prevention.
• Targeting high-risk physicians improves the efficiency of fraud detection efforts.

Purpose of the Study:

• To identify indicators of healthcare fraud and abuse in general physicians' drug prescription claims.
• To identify a subset of general physicians more likely to commit fraud and abuse.
• To develop a data mining approach for efficient auditing.

Main Methods:

• A data mining approach was applied to a large dataset of private sector general physicians' prescription claims.
• The methodology involved problem clarification, data preparation, indicator identification, cluster analysis, and discriminant analysis.
• Discriminant analysis was used to validate the effectiveness of the identified indicators.

Main Results:

• Thirteen indicators for healthcare fraud and abuse were developed.
• Over half of general physicians (54%) were identified as suspects of abusive behavior.
• Two percent of physicians were identified as suspects of fraud, with high detection rates (98% for fraud, 85% for abuse).

Conclusions:

• The data mining approach effectively identifies physicians suspected of fraud and abuse.
• This method allows health insurance organizations to streamline auditing by focusing on suspect groups.
• The approach is particularly beneficial for low- and middle-income countries (LMICs) seeking to optimize audit strategies.