Medical Mutual of Ohio Seeks to Discourage Fraud Before It Occurs

Medical Mutual of Ohio leverages ViPS' Star Sentinel software to identify potentially fraudulent claims activity as it comes through the door, discouraging fraud before it begins.

Medical Mutual of Ohio (around $2 billion in annual revenue) always has had some success in identifying fraudulent claims activity, as evidenced by the 700 indictments John Shoemaker, financial investigations unit manager, and his team have produced over the past 20 years. More recently, however, the approach Shoemaker and his team take aims to stop fraud before it occurs.

According to Shoemaker, physician providers are behind 70 percent of the Cleveland-based health insurer's fraud cases. By focusing on cases with the highest potential for real fraud while identifying those cases earlier in the claims process, Medical Mutual of Ohio hopes to discourage the provider community from filing fraudulent claims to begin with. "A perfect scenario is [one in which] we never recover a dime because we have designed a system that prevents payment and that warns the provider community and stops them from attempting to bill us fraudulently," Shoemaker explains.

In early March, Medical Mutual of Ohio announced that its financial investigation unit, one of the departments in the company's risk management organization that investigates fraud for the carrier, had uncovered $4.9 million in fraudulent activity in 2007. Shoemaker says his group has seen similar results since 2003, when the company first deployed the Star Sentinel solution from ViPS (Baltimore). The software, he adds, has improved and streamlined the investigation process, allowing investigators to identify and scrutinize leads with the highest potential for fraud and to do so early in the claims process.

Identifying fraud as it is being billed is optimal, Shoemaker continues, not only because it reduces fraud losses, but because it is more effective in changing fraudulent billing patterns and behavior. "The old way of doing things was more reactive," he says. "This has allowed us to be much more proactive, and that's the big change."

The Star Sentinel software primarily looks at structured claims data fields, such as diagnoses, procedures and providers, according to Shoemaker. The resulting data map then is run through more than 300 algorithms -- many related to known healthcare provider fraud schemes and the historical billing patterns of specific physicians.

Ease of use, Shoemaker says, was a key factor in the product's selection because he wanted a tool that would be accessible to all investigators, not just those with a highly developed technical acumen. But, he points out, the "vanilla" system, as Shoemaker describes it, does allow for a certain degree of customization.

"It's a very structured environment, but it allows us as users to do some manipulation to the algorithms that are specific to processing in our environment," Shoemaker explains. "And every payer has little quirks and issues with their own adjudication systems and their own policies and procedures."

Preventing Losses

Traditionally, investigation units put out fires -- insurers identified fraud by acting on tips that often were six or more months old, according to Shoemaker. "By the time [tips] came to investigators, who were using old systems, these cases were two years old," Shoemaker relates. "Millions or hundreds of thousands of dollars were lost. The money was [already] gone, and recovering it was impossible."

Thanks largely to technology advances in areas such as analytics, the claims investigation discipline has evolved since then. Many insurers have shifted their fraud identification efforts to the front end of the process, Shoemaker says. "Almost everyone is looking at postadjudication, prepayment-type systems and trying to catch things before they go out the door," he contends. "We were looking for a system that ... would focus on looking at a short span of history -- three to six months -- and that could identify known schemes and look at the billing patterns of providers to help identify potential problems before they blew up into large problems."

In addition to catching fraud earlier in the claims process, the Star Sentinel software also has helped investigators avoid wasting time on false positives or dead leads, Shoemaker notes, adding that, on average, his unit investigates around 120 cases of potential fraud each year. "There's a big difference between identifying 120 cases that turn out to have no potential [to uncover fraud] and shifting that to work on 120 cases that are aberrant billing that is fraud," Shoemaker relates. "You have a limited number of resources, investigators and analysts, and you have only so much time. So let's work on 120 [cases] that have some real potential versus 120 that are dead ends."