By PATRICIA VOWINKEL, who has worked for national media outlets for more than 20 years
A video showing seven men from a company working on a roof recently turned up on YouTube. It was not the type of thing that would have grabbed most people's attention, but the company's workers' compensation insurer, Boston-based Liberty Mutual Group, found it very revealing. Why?
Liberty Mutual didn't know the company employed roofers--the company had simply described its business as "general carpentry," and that left the company with a lot of explaining to do.
This scenario isn't uncommon. Studies indicate up to 10 percent--perhaps more--of all property/casualty insurance claims are fraudulent, according to the National Insurance Crime Bureau (NICB). The NICB has estimated that workers' comp fraud costs the industry as much as $5 billion annually.
While fraud is a perennial problem, advances in technology are helping to give insurers a new edge. With billions of dollars at stake, carriers and investigators are using an array of tools to combat fraud, including predictive analytics, social media such as Facebook and YouTube, and claims case-management systems.
These tools provide carriers with a wealth of new information that can help control fraud, which contributes to higher insurance premiums for all consumers.
Before an insurer even binds a policy, it has an opportunity to reduce the risk of fraud through predictive analytics, which seeks to identify the potential for problems ahead of time.
Insurers can use predictive analytics to assess public information about customers and then identify the ones who fit the pattern of those most likely to commit fraud.
"What we're seeing is a recognition that the opportunity starts before the policy is even bound," said Michael Costonis, managing director for the North America insurance industry program at consulting and technology services firm Accenture.
Insurers are taking advantage of techniques used for years in the retail industry to learn as much as possible about their potential customers--their buying habits, what sites they visit on the Internet, their credit scores, their purchases. That information can provide important insights and help the insurer evaluate whether that customer fits the profile of a fraudster, Costonis said.
By doing this, insurers can better segment and profile customers before presenting them with a quote or product and can charge them accordingly. "Any policy can be written at the right price," Costonis said.
Identifying potential bad apples in advance is important because by the time a claim is filed, it can be difficult to detect fraud as fraudsters learn quickly how to beat the system of detecting fraudulent claims.
"Insurers are coming to the realization that predictive analytics is the best way to get ahead of the fraudsters," said Russ Schreiber, vice president for the Insurance Market at FICO, a provider of analytics and decision-management technology.
"They're finding the people that commit fraud change behaviors very quickly and that trying to prevent fraud through setting flags and maintaining rules in their claims processing systems puts them several steps behind the perpetrators," he said.
"Predictive analytics, on the other hand, finds fraud much earlier because the technology is looking for aberrant, outlier behavior."
Predictive analytics is also used in claims processing as insurers mine data to see how it matches against patterns that were previously identified as fraudulent.
Insurers can use predictive analytics to examine information presented in claims--such as the name of a body shop, the names of doctors, addresses, phone numbers--and look for relationships that match established patterns.
Claims that appear to fit certain patterns can be forwarded to a more highly skilled adjuster.
The increased interest in this field is underscored by a number of deals and partnerships that have been announced over the last year. IBM, for instance, last year acquired SPSS, a provider of predictive analytics software and solutions, for $1.2 billion.
Accenture and SAS, a business analytics software and services firm, in February said they planned to expand their relationship by jointly developing, implementing and managing next-generation predictive analytics solutions.
The Accenture SAS Analytics Group plans to focus initially on developing predictive solutions for the financial services industry, including banking and insurance, as well as healthcare and the public sector.
In April, FICO and Emdeon, a provider of healthcare revenue and payment management systems, announced an agreement aimed at tackling fraud, waste and abuse in health insurance. Under the alliance, Emdeon will incorporate FICO's insurance fraud manager system into its leading healthcare revenue and payment cycle transaction-processing network to offer an enhanced fraud prevention solution.
Other leaders in predictive analytics for insurance include Valen Technologies and Eagle Eye Analytics. Valen announced a partnership in April with Overland Solutions Inc. to combat premium leakage resulting from misclassified workers' compensation and general liability policies.
Misclassified or misrepresented payroll is when a company claims it has 10 secretaries and no roofers, for example, when it in fact has two secretaries and eight roofers. Or perhaps the company reports that it has 20 employees when it really has 40, with the goal of lowering premiums.
RateRight, Valen's latest predictive analytics product, is designed to identify policies that cover companies where the risk of inaccurate or misclassified payrolls is high.
Eagle Eye, meanwhile, last year announced a strategic alliance with Milliman to provide the insurance industry with the latest predictive analytics offerings.
People like sites like Facebook and YouTube that let them stay in touch with friends and family and share photos and videos.
Fraud investigators like Facebook and YouTube, too, because they have become such a rich source of information. "They put stuff on YouTube that you just can't believe," said Neil Johnson, manager of the premium fraud special investigation unit at Liberty Mutual.
One notorious case involved people who rammed a van into a tree and then posted video of the incident on YouTube, said James Quiggle, a spokesman for the Coalition Against Insurance Fraud.
"People post the dumbest and most incriminating things imaginable," he said. "They don't realize that investigators are trolling social media for case leads."
Of course, investigators have always had access to the news media, but now thanks to social networking media there is a wealth of information at investigators' fingertips--much of it provided by the suspects themselves.
Videos of men working on a rooftop or people ramming their van into the tree, for instance, would have only circulated among a small circle of people and never would have fallen so easily into the hands of fraud investigators before social media.
Predictive analytics can help to identify customers who fit the profile of a fraudster, and social media can help provide investigators with information to help them in their investigations.
Claims databases, however, are another important tool in the fight against insurance fraud. ISO ClaimSearch, for instance, is a comprehensive system for improving claims processing. Each year, insurers and other organizations submit tens of millions of reports on individual insurance claims, and ISO stores those reports in a single database. This provides the industry with data for researching prior-loss histories, identifying claims patterns and detecting suspect claims.
Claims case management also can help in the fight against fraud.
A system from Global Options, for instance, provides users with critical information that helps to speed up case management and quickly resolve questions of fraud. The Web-based system GlobalTrak connects users to databases with information needed to investigate and process claims, such as court records and other public documents. Information that used to take days to obtain is now available within minutes.
By speeding up investigations, insurers can make a determination much more quickly about fraud and whether to cut off benefit payments.
"It saves a lot of time," said Carl Ayestas Jr., corporate compliance officer with third party administrator Cannon Cochran Management Services Inc (CCMSI), which has been using Global Trak for about five years.
"We can go to their site and everything is right there to view and pull as we need," he said. "It helps us to make decisions quicker. No question there is a savings to our clients."
Fraud is the insurance industry's great white whale. Carriers know it's out there and fraud investigators know it's big, but tracking and capturing fraud can be an almost impossible task, Accenture's Costonis said.
The insurance industry is waking up to the possibilities of predictive analytics; and social media, expanded claims databases and more efficient claims case-management systems are helping insurers in their fight by giving the industry information much more quickly.
Insurers may never be able to eliminate fraud entirely, but with advances in technology they might put the great white whale on a diet.
June 1, 2010
Copyright 2010© LRP Publications