Using support vector machines to detect medical fraud and abuse

Area of Science:

• Health Informatics
• Medical Billing Systems
• Fraud Detection

Background:

• Medical billing processes are complex and prone to errors and fraud.
• Existing systems often require significant manual review, leading to delays and increased costs.
• There is a need for automated solutions to improve efficiency and accuracy in medical claims processing.

Purpose of the Study:

• To introduce Quash, an automated system for medical bill processing.
• To evaluate the architecture and efficacy of Quash in bill routing and abuse detection.
• To demonstrate Quash's ability to accelerate the work of medical fraud detection experts.

Main Methods:

• Developed an automated medical bill processing system named Quash.
• Integrated Quash with human auditors and standard bill review software.
• Focused on medical data and billing patterns within the United States.
• Discussed implementation details and preliminary experimental results.

Main Results:

• Quash provides a real-world speed-up for medical fraud detection experts.
• The system is capable of efficient bill routing and abuse detection.
• Preliminary results indicate efficacy in cost containment for medical claims.

Conclusions:

• Quash offers a viable automated solution for medical bill processing and fraud detection.
• The system enhances the capabilities of human auditors, improving overall efficiency.
• The principles of Quash are potentially applicable to other financial transaction environments.