Saying the insurance industry is complex would be an understatement. However, detecting fraud could be easier than companies think. Nirup Menon, information systems and operations management associate professor, found that insurance companies would benefit from sharing their data in order to detect medical claims fraud.
Menon presented his findings in “Information Spillovers and Semicollaborative Networks in Insurer Fraud Detection,” which has been accepted by Management Information Systems Quarterly for publication.
Since many insurance companies spend money on fraud detection, in part due to an Obamacare clause mandating that 80% of an insurance company’s budget should be spent on improving patient care, Menon wanted to learn whether there were spillover benefits from one insurance company’s fraud detection efforts to another company.
“A lot of times the companies don’t have examples of all types of fraudulent records. A lot of them are trying to automate fraud detection, but they need examples and patterns of fraud to do so. That’s what business analytics is about these days,” he said.
Detecting fraud patterns may improve as companies share data and fraud patterns, “this way information spills over from one company to another.” Menon described the concept of information spillover as benefitting from what your neighbors are doing.
“I did find that yes, an insurance company benefits when other insurance companies are also spending efforts on fraud detection. This shows that business analytics should be a collaborative effort, rather than everybody keeping and analyzing their own data,” he said.
There is no downfall to companies sharing data, because “whatever they’re able to recover and not pay out in terms of claims, benefits everybody.” Because of this, Menon was surprised to learn that the level of sharing amongst companies is low.
“I imagined that they’re thinking of sharing data as in actively on each and every case or every patient or provider, but perhaps the sharing is happening more at a level of patterns,” he said. “At conferences, workshops, and with employees going from one insurance company to another in terms of fraud detection, information spillover may be more subtle than what they’re aware of.”
Menon said it would benefit companies to share data in a way they’re more aware of, because they’re likely to find more fraud. “That will reduce their costs, their payouts, which means they would get higher profits and hopefully they can pass on those savings in terms of lower premiums to patients and employers.”