Fighting Waste, Fraud and Abuse through Analytics: Case Study from Centers for Medicare & Medicaid Services (CMS)

As part of GovLoop’s recent report, Unlocking the Power of Government Analytics, I had the opportunity to speak with staff members from the Centers for Medicare & Medicaid Services (CMS). CMS staff provided expert insights on the importance of analytics to control waste, fraud and abuse. In 2010, CMS and the Department of Health and Human Services (HHS) launched an aggressive campaign to reduce fraud for medical claims. The following post is an excerpt from the report, you can read the entire report below.

Post Highlights

  • Be sure to view GovLoop’s recent report Unlocking the Power of Government Analytics
  • The use of analytics has redefined how the public sector delivers services
  • CMS has leveraged analytics to successfully combat waste, fraud and abuse, view the full guide below or as a PDF

To meet their goals of reducing fraud, CMS developed a new anti-fraud tool using predictive analytics and real-time data to spot abnormalities for fee-for-service claims. CMS has provided a thorough analysis of the Fraud Prevention System Report in a recent report to Congress.

The FPS was required under the Small Business Jobs Act of 2010 (SBJA). In the report to Congress, CMS staff states, “Since June 30, 2011, the FPS has run predictive algorithms and other sophisticated analytics nationwide against all Medicare fee-for-service (FFS) claims prior to payment. For the first time in the history of the program, CMS is systematically applying advanced analytics against Medicare FFS claims on a streaming, nationwide basis.” The Small Business Jobs Act identifies specific functional requirements of the predictive analytics technologies for CMS. This report identifies the following requirements for CMS:

  • Provide a comprehensive view of Medicare FFS provider and beneficiary activities to identify and analyze provider networks, billing patterns, and beneficiary utilization patterns and identify and detect patterns that represent a high risk of fraudulent activity
  • Integrate fully with the Medicare FFS claims flow
  • Analyze large datasets for unusual or suspicious patterns or anomalies before payment and prioritize suspicious activity
  • Capture outcome information to continually refine and enhance the system
  • Prevent payment of fraudulent claims.

Click Here to Download PDF of View Below

The Fraud Prevention System (FPS) developed by CMS uses predictive analytics technology to identify and prevent medical fee-for-service (FFS) claims. Since June 2011, CMS has been using FPS to screen FFS claims. To build the system, CMS worked across sectors to learn best practices from the telecommunications and banking industry.

The report to Congress identifies three analytic strategies that CMS has implemented. They are, anomaly detection models, predictive models, and social network analysis. In our interview, CMS staff identified that through robust analytics adoption, the agency is able to improve the allocation of investigative resources and become more efficient identifying and finding fraudulent claims. FPS has allowed staff to prioritize work, cut costs, and increase workforce capacity within CMS.

The first model that is used by CMS is the anomaly detection model, which is a sophisticated model that defines thresholds of acceptable behavior. This model compares an individual providers behavior patterns and contrasts with that of a peer group. The report states, “Certain behaviors and characteristics that indicate potential fraud may also be indications of acceptable behavior. For example, if a provider bills for many more services than are normally performed by similar providers in a defined time period, the FPS can alert an investigator to inspect the claim prior to payment.”

CMS has also developed rigorous predictive analytics models to spot and identify fraudulent claims. This model uses data collected from previous fraud cases to help predict fraud and allow CMS to investigate suspicious complaints. The CMS report states, “Developing predictive models requires advanced analysis because a fraudulent claim may become apparent only when factors are considered in combination; whereas independently, those factors may not be suspicious.”

Social network analysis models are also used by CMS to identify links to fraudulent complaints. “The ability to link providers through their social networks helps CMS and its law enforcement partners unravel the complex relationships among fraudulent providers and between providers and beneficiaries,” states CMS.

The FPS has saved CMS millions of dollars, and allowed CMS to identify claims and patterns of behavior. This process allows them to understand trends, and since information in occurring in real-time, FPS is able to immediately spot fraud, saving time of investigators and CMS staff.

CMS staff mentioned that for those getting started with analytics, it is essential to clearly define the problem that is being solved, be sure the right data is collected, and identify resource needs for an analytics project.

Want More GovLoop Content? Sign Up For Email Updates

Adobe Analytics delivers the complete picture of digital citizen engagement that helps you track, analyze and optimize information services and citizen-agency engagement across multiple channels-including walk-up, call center, mail, online and mobile. Optimizing your agency’s online presence with Adobe SiteCatalyst while visualizing invaluable knowledge of historical website interactions with Adobe Insight will transform your agency’s online experience and reduce the opportunity for fraud.

Leave a Comment

Leave a comment

Leave a Reply