How Predictive Modeling Exposes Claims Scams

Specialty health insurer Universal American (White Plains, N.Y.) had a problem: A common Medicare fraud scheme was making its way from government-administered plans to privately administered ones -- and it was a target.

"The people that were doing the scheme typically would bill the plans for infusion therapy, and once the plans detected them they'd move to the next plan," says James Cooper, Universal American's SIU manager. "That means these schemes can only hit about five [Medicare part B] contractors until it's done. In Medicare Part C, you have literally hundreds of Medicare Advantage plans that you can hit."

Infusion therapy includes any intravenous administration of drugs, including pricey cancer treatments -- which makes it an attractive code for scammers. To detect which claims were legitimate, Universal American turned to FICO's (Minneapolis) Insurance Fraud Manager predictive analytics software. Predictive analytics were key to the company's strategy of identifying problematic claims before they were paid as opposed to afterward, Cooper says.

"If a new provider comes in and billed large amounts of infusion therapy, but the patient doesn't have a history of cancer issues, the software could say, 'This diagnosis doesn't make any sense,'" he explains.

Universal American's use of predictive modeling before payment is a leading strategy in the industry, adds FICO insurance industry principal Russ Schrieber.

[Related: FICO research on predictive modeling in claims]

"We don't think enough of the market is [using these tools] pre-payment," he says. " It's not that the tech is in the way, though, its organizational constraints -- for example, some customers are saying they don't want to disrupt the provider network."

But Universal's strategy is paying off in other ways, Cooper adds.

"By looking at claims that we are denying up front and where we have a high success rate, we can go back to our claims dept and say, 'We keep denying this procedure code at a high rate, what can we do to stop these?" he explains.

The end result is about $2 million extra saved in other detections of waste, fraud and abuse, on top of the total $6 million saved from detecting the infusion therapy scams, he says.

Nathan Golia is senior editor of Insurance & Technology. He joined the publication in 2010 as associate editor and covers all aspects of the nexus between insurance and information technology, including mobility, distribution, core systems, customer interaction, and risk ... View Full Bio