Kirsten Brueggemann

Associate Editor

Loyola University Chicago School of Law, JD 2025

 

The United States spends more money per person on health care than any other country, approximately $4.2 trillion in 2021. Unfortunately, our complex health care system and the large budget make fraud a significant concern for the U.S. Government, payers, and patients. The National Healthcare Anti-Fraud Association estimates that as much as 10% of annual healthcare spending is lost to scams, resulting in billions in losses yearly. To combat healthcare fraud, the Department of Health and Human Services Office of the Inspector General, in collaboration with state law enforcement and other governmental agencies has created special Strike Forces. These efforts have led to substantial recoveries of federal funds and criminal/civil prosecution of individuals or entities involved in Medicare and Medicaid fraud. Besides avoiding unnecessary or fraudulent claims, individual healthcare payers are motivated to prevent fraud due to severe penalties associated with the False Claims Act, Anti-Kickback Statute, Physician Self-Referral Law (Stark Law), and Civil Monetary Penalties Law. How can individual payers detect and try to prevent fraud? The answer is AI.

Current landscape and implementation

Healthcare payers process numerous claims daily and pay healthcare providers. With the threat of expensive litigation, fines, criminal charges, and potential Medicare/Medicaid exclusion, payers are turning to AI to combat healthcare fraud, waste, and abuse. This ensures payment only for necessary and provided services.

AI rapidly analyzes vast data sets to confirm patients receive the service they pay for and also detects evidence of upcoding. AI identifies patterns and spots anomalies in patient and provider histories, helping uncover unintentional and intentional fraudulent activities. In initial AI trials for healthcare fraud detection, fraud identification increased two to three times, with false positives decreasing by ten to twenty times.

The pros and cons of false positives

A false positive occurs when technology flags a claim as potentially fraudulent, but a subsequent analyst investigation finds no fraud. While healthcare payers can never truly eliminate false positives, they can reduce them using AI to yield results. With steep fines and penalties for undetected or tolerated fraud, payers should prioritize implementing protocols and systems that enhance early and accurate fraud detection.

Recent pilot project

Mastercard Health Solutions’ AI teamed up with Milliman Inc. to improve its healthcare systems, address emerging risks, and bolster financial security. This pilot project identified 2,700 high-risk providers and $240 million in potential savings for the payer. The developers built three distinct models to spot suspicious claims and provider patterns. One model built a baseline for each provider by examining volume of claims, average number of patients submitted, and amount of money billed. Another model looked at specific claim details to evaluate whether the claims were appropriate according to the billing codes and medical history. The last model examined whether the providers stuck to coding principles and standards. All three models assigned a risk score to the provider. A higher the risk score, triggered additional actions, such as investigation, requesting supporting records, or even suspending future payments.

Human element

Healthcare laws continually evolve, leading to ever-changing fraud schemes. While AI can teach itself to identify sophisticated fraud schemes unfamiliar to seasoned investigators, it will also flag anything unusual. Therefore, human oversight is vital to verify proper claim processing and payment for legitimate services. One significant challenge is developing a system that understands and processes the human element of data without computing errors. Doctors enter data into a system in a variety of ways. AI must be able to tell the difference between a unique way of entering data and legitimate fraud behavior.

Conclusion

AI is a potent tool capable of swiftly analyzing vast data volumes. It has the potential to revolutionize how government agencies and payers combat healthcare fraud, waste, and abuse. AI can flag and identify data for further review and investigation. However, it requires supervision by analysts who adapt to evolving circumstances.