The Centers for Medicare and Medicaid Services (CMS) on July 1 will start using "predictive modeling" technology to detect fraudulent claims. According to CMS, this initiative fits with the agency's new approach to fighting fraud, which is to avoid making payments on wrongful claims rather than "pay and chase" the perpetrators after the fact.
Credit companies use predictive modeling to detect patterns of fraud. Health plans also employ it to forecast which of their members are likely to generate large costs in the next year. But using this method of risk scoring to spot suspect healthcare claims is new, at least for CMS.
CMS has contracted with Northrop Grumman to develop its predictive modeling technology. Northrop Grumman is bringing in two other companies as partners: National Government Services and Federal Network Services, a Verizon subsidiary.
Northrop Grumman will deploy algorithms and an analytical process that looks at CMS claims by beneficiary, provider, service origin or other patterns to identify potential problems, and assign an "alert" and "risk scores" for those claims. Those alerts will be further reviewed to allow CMS to prioritize claims for additional review and assess the need for investigative or other enforcement actions.
According to a government fact sheet, the Affordable Care Act provides $350 million over 10 years to ramp up anti-fraud efforts, "including increasing scrutiny of claims before they're paid." Health Data Management, however, reports that the money for the predictive modeling program comes from a provision in the Small Business Jobs Act of 2010.