Stomping Out Health Care Fraud Before it Happens: A Call for Using More Predictive Analytics

The Obama Administration made it a top priority to prevent Medicare and Medicaid fraud, and its efforts have yielded great results. This past summer, the Administration announced that with sophisticated detection methods like “big data” and predictive analytics, the Center for Medicare and Medicaid Services (CMS) was able to prevent $42 billion of improper payments in fiscal years 2013 and 2014.
Still, all is not well. Many legislators are not content with CMS’ efforts in combatting Medicare and Medicaid fraud, and believe that given the sophisticated fraud detection methods CMS has at its disposal, CMS is not doing enough to prevent health care fraud before it happens. Recently, members of several House committees, including the Committee on Ways and Means, which has jurisdiction over the Medicare program, sent a letter to CMS, citing concerns over CMS’ continued “heavy” reliance on a “pay and chase” model, under which CMS investigates claims only after it has already paid the providers.
The problems with a “pay and chase” model, of course, are not only inefficiency and the costs associated with recouping fraudulent payments, but also the threat that remains when delayed detection can result in criminals causing even more damage. As the House committee members stated in their letter,

each dollar lost to fraud is a dollar that is not used to benefit patients.

 
Under the Affordable Care Act, CMS received additional funds for expanding its fraud prevention initiatives. The Fraud Prevention System (FPS), CMS’ “advanced analytics system,” was also implemented in 2011 to address the concerns with a “pay and chase” model with Medicare claims. Through the FPS, CMS is able to identify suspicious patterns by analyzing large volumes of billing information prior to paying the provider. Predictive modeling, unlike data analytics which utilizes  various post-payment analyses in addition to predictive modeling, attempts to apply various statistical and analytical techniques to historical data in order to predict future behavior and events.  It is estimated that the FPS analyzes 4.5 million Medicare pre-paid claims each day.
According to an interview with the Deputy Administrator for Program Integrity and Director of the Center for Program Integrity at CMS, while the push for “big data” efforts has not come without challenges, it has led to more “confident decision making” that has reduced both costs and risks. For instance, FPS was able to identify a Florida home health agency that billed for services that were not rendered. CMS was able to place that agency on prepayment review and payment suspension, refer the agency to law enforcement, and revoke the agency’s Medicare enrollment.
Critics woes have not fallen on deaf ears, however. CMS continues to show its commitment to prioritizing health fraud prevention through a more “proactive” way of dealing with fraudulent payments. Currently, CMS is working on designing a new system to develop “next-generation” predictive data analytics.  The new design is expected to improve the FPS’ utilization and efficiency. It will be interesting to see more figures showing the savings from investments in sophisticated and advanced predictive analytics tools.