Next, it executes a fuzzy logic search to expose name, address, and VIN changes, for example, to help the investigator decide if the information matches up. (A connection to ISO's claims database allows a company to match elements of a claim to those of hundreds of other carriers.) The final part of the search process involves visualization of data-this gives clues as to what to expect in the analysis phase.

Rob Doone, president and CEO of NetMap Analytics, said: “The application helps investigators find areas to search for the most success, from when a claim looks suspicious through the end of the investigative process.”

In the analysis phase, the application uses complex search algorithms, sizing, and spatial relationships to reveal similarities and other factors between entities.

CSC's Fraud Investigator is a database tool that helps insurers identify fraudulent activity. The application imports data from a company's claims database and creates a searchable index with pointers to the original files. Using weighted search parameters, the solution scours the indices it creates to find similarities or relationships between claimants, claims, and other customizable factors, and delivers detailed reports.

For example, if a claim involves an auto accident covered under a policy activated fewer than 30 days from the event, Fraud Investigator can discover that in a search. Once a carrier has determined which variables in a claim might indicate fraudulent activity, it can use Fraud Investigator to head off suspicious notices, refuse payment, and pursue the case in court.

Once a search is executed-a name query, for example-Fraud Investigator will find best matches and echo them to screen with links to all claims involved with or matching the search parameters. According to Dennis Parker of CSC's cost containment area, best matches are what information investigators really need.

“Insurance data are full of mistakes,” Parker said, “but with Fraud Investigator, you can work around that by putting in as many search parameters as you want, instead of searching only names or addresses.”
In the hunt for data relationships, investigators can attribute more weight to any category in a search. System administrators can change category weights on a multi-user or batch level, while users can make single-search modifications only.

Of course, the time needed to import claims information will vary in relation to project size. But according to Parker, it takes two to three days on average to analyze the data, create a schema or 'index blank,' then populate the blank with records. Using the above example, Parker said by processing 74,000 records in 45 minutes, system readiness could be expected in four to five days.

The solution can be used two ways: as an analyst, looking at the big picture and creating diagrams to help illustrate points for investigators; and as an investigative tool, connecting authorized parties to databases to investigate claims submitted by adjusters.

ROI can be found in the short term: A New Jersey-based P&C carrier was able to identify two fraudulent claims in the first month of use, exposing a chiropractic group and lawyers who were creating phony claims by finding runners willing to report soft tissue damage. The company denied the claims and prosecuted. The insurer told CSC denying the two claims paid for the system for an entire year.

Another company used the solution to recoup previous payouts by investigating VINs from Firestone-tire-related incidents. In batch, the VINs were compared to claims paid; in the end, the company was able to subrogate Firestone.

But if an accident involves more than tires, Injury Sciences' WrExpert, a PC-based solution-standalone or on a server-evaluates auto claims and can alert investigators to staged events.

The software searches large amounts of information on a database, looks for similarities or other connections between claimants and the claims, and reports frequencies and relationships. Most importantly, it identifies anomalies between what is reported and what actually happened in auto collisions. To accomplish this, WrExpert searches its library of information about accident events, human anatomy, causes for injuries, and related treatments.

If, for example, stories don't match-one car has too much damage, the other not enough-WrExpert will identify the discrepancy, taking into account variables such as how the collision was caused, traveling speeds, braking, and more.

“This is a tool that takes data and does a reasonableness check,” said Scott Palmer, president and CEO of Injury Sciences. “When things don't add up, it explores other necessary factors to form a conclusion.”

Injury Sciences recently released new features and other enhancements to WrExpert, now up to version 3.0, including the ability to evaluate various side- and angled-impact scenarios and understand potential for injury in related collisions. Palmer said the new features provide an objective insight into point of impact, angle, severity, and what type of injuries-if any-should be expected.

Injury Sciences also hooked up with ProcessClaims.com-an Internet-based forum for insurers, collision repair professionals, and their trading partners-to provide WrExpert's functionality to ProcessClaims.com customers. Through the agreement, users download estimates from the ProcessClaims.com site(www.processclaims.com, of course) into the WrExpert system, which then analyzes the data to determine collision severity and indicate what types of resulting injuries are likely.

Streamlining the claims workflow will help save time and money by allowing field adjusters to move on to other claims rather than working on one for a week or more, only to realize it needs to be passed on to the SIU. So why not automate the process?

Inspire's Empower Claims, set for release this fall, is an enhancement to the company's existing Empower product, currently available to the policy side only. The solution will help take some of the guesswork out of identifying and routing suspicious claims. Inspire is adding fraud indicators to the system workflow distribution application.

Here's how it works. When a loss is keyed into the system, a scoring system based on criteria-set either by Inspire or the customer-routes the loss notice to an adjuster based on complexity and score. If certain characteristics are present in the notice, or a certain score limit is met, the claim will also be automatically routed to the SIU. (This functionality is available to companies who outsource to Inspire or use the software on site.) According to Eric Lundby, vice president of claims administration at Inspire, the solution can be used for homeowners' and commercial policies, and any other type of P&C product.

Lundby said that, as a general rule, the application installs on the policy side in less than a week-the same should be true for claims applications. And Inspire will build an interface for customers to suit their environment and make working with Empower Claims more intuitive.

“Training is minimal because it's just a matter of deciding criteria, and then building them in. It should be transparent,” Lundby said.

Many factors make carriers reluctant to challenge suspect claims in court; the most visible is money. Costs add up when you consider legal fees, investigative expenses, and the prospect of losing more money if a rejected claimant files a lawsuit. Claims fraud detection and prevention solutions give insurers the ability to export detailed illustrations of phony claims, behavioral patterns demonstrated by claimants, and concrete connections between the two-an effective tool for support in litigation. As case complexity disappears, the cost of pursuing fraudulent claims will decrease.

By using what's available, carriers will be able to combat fraud, show criminals that safeguards are in place, and with perseverance, lower costs for consumers.