Using Predictive Modeling to Prevent Healthcare Fraud

December 1, 2011

HealthcareIn my last post, I discussed a few of the changes federal and state governments have made in their ongoing efforts to reduce Medicaid fraud. 

However, there are also significant changes you should be aware of in the way the federal government is identifying and preventing healthcare fraud.

In July, the Centers for Medicare & Medicaid Services (CMS) began using predictive modeling technology as part of its National Fraud Prevention Program.  The technology is based on a customized version of the software Verizon uses for its fraud detection programs.  Northrop Grumman; National Government Services, a WellPoint subsidiary; and Verizon are working together to support the technology.

Predictive modeling uses advanced analytics — including behavioral and statistical analysis — to review volumes of informationto identify fraudulent claims.Yes, this is the same technology that allows your credit card company to flag questionable charges at the point of sale, usually before you even know they were made.

The predictive modeling technology will take CMS from its previous “pay-and-chase” system, which allowed providers to receive Medicare reimbursements first and then required CMS to attempt to recovery the overpayments after the fact.  Now, CMS to have a near real time, comprehensive view of Medicare claims nationwide and, just like your credit card company, it will be able to deny the claim before payment is ever made.

Shortly after CMS announced the new technology, Lewis Morris, Chief Counsel to the Inspector General of Health and Human Services, testified before a Senate subcommittee about the role new technologies would play for the Office of Inspector General (OIG).]  Not only is the OIG using data mining, trend evaluation, and predictive modeling to identify fraud, it is also using it to target their enforcement efforts.

For example, in its Work Plan for 2012, the OIG says that it will use “computer matching and data mining techniques” to select acute care hospitals for focused reviews of claims that may be at risk for overpayment.

While preventing fraud is the focus of predictive modeling and data mining, providers also need to know that under the new rules for suspension of Medicare payments, a “credible allegation” of fraud can be supported by information from claims data mining. For many providers, suspension of Medicare payments will force them out of business very quickly.

Not only is CMS actively using these tools, it is also encouraging the states to follow suit by proposing to allow the Medicaid Fraud Control Units (MFCUs) to spend their federal matching funds for data mining. Florida’s new Medicaid and Public Assistance Fraud Strike Force has already contracted with ERS Group, a Tallahassee economic research firm, to develop statistical models to detect Medicaid fraud.

With an estimated $70 billion in improper payments in 2010 to Medicare and Medicaid, you can be sure CMS will put the new technologies to good use to detect and prevent healthcare fraud.

Resources to help your company or your clients stay in compliance are available on Thomson Reuters Accelus Compliance Advisor.