Firms that provide workers' compensation insurance need to be on the lookout for fraud in a variety of areas, such as employees filing for fake or exaggerated injuries and employers filing incorrect worker codes to save money on premiums.
Increasingly, insurers are turning to a variety of technology solutions to detect and stop fraudulent claims, which typically fall into three main categories, says Karlyn Carnahan, principal at Novarica, a New York-based consulting firm. One is premium fraud, where businesses try to reduce premiums by misclassifying themselves or entering the wrong job codes for employees. The second type is faked injuries, where employees fabricate or inflate injuries to collect payment. And the third involves medical mills, where organized rings of doctors or lawyers fake or pad injury claims, or bill for treatment never provided.
Clearly, fraud is a growing problem, particularly in a sluggish economy. According to a survey of fraud bureaus conducted by the Coalition Against Insurance Fraud in 2009, the bureaus reported that, on average, the number of referrals received and cases opened increased in all 15 categories of fraud included in the survey.
The good news for workers' compensation insurers is that workers' compensation had one of the smallest increases in fraud incidents. Nevertheless, the problem continues to be significant for firms that provide this type of insurance.
DIVINING PATTERNS
The Special Investigations Unit (SIU) at Chicago-based CNA saw a 17% increase in workers' compensation suspicious claim referrals in the second quarter of this year compared with the same period of 2009. That's significant considering that overall claim volume was down about 8% in the quarter, says Tim Wolfe, director of the unit.
"The main reason [we are seeing this increase in workers' comp claim referrals] is because we're seeing more fraud, not on the individual claimant side but more on the medical provider side," Wolfe says. "We're focusing more on that aspect of workers' comp fraud. We've increased our number of investigators on this type of fraud, and we also formed an intelligence unit with a couple of full-time data analysts."
Technology plays a big role in addressing fraud, Wolfe says. The investigations unit in 2007 began using a software application called Analyst's Notebook from U.K.-based i2 Group, which helps identify patterns of fraud on a large scale. The software aggregates data from a variety of claims sources.
"Rather than looking at fraud one claim at a time, we're able to see patterns emerging across multiple claims files," Wolfe says. For example, medical providers might consistently bill the firm for treatment not rendered, overcharge for treatments or bill for maximum time spent with patients.
Looking for these types of activities one claim at a time is a slow process. The software enables the unit to identify patterns of fraud across multiple claim files, so it can run proactive studies to look for medical providers that have been frequently using the types of codes that indicate these fraudulent activities, Wolfe says.
"Once we identify a pattern of fraud, we send alerts to the claims department and notify law enforcement," Wolfe says. "We try and shut down [the activity] before we actually start paying money. In some cases we have saved hundreds of thousands and even millions of dollars."
The CNA SIU also uses a predictive modeling software tool it developed internally in 2008 to analyze data on past workers' compensation claims. "We built the model based on certain data elements that were common to known fraud claims," Wolfe says. "We run new claims through that model" to look for those elements.
The tool isn't perfect, Wolfe says; for example, it has a false positive ratio of 85%, which he says is still too high. But it's inexpensive to run and enables the firm to identify potential cases of fraud among workers' compensation claims.
IN PLAIN SIGHT
EMPLOYERS, a Reno, Nev.-based provider of workers' compensation insurance and services for small businesses, uses database analysis to check Social Security death file information against claims, to see if people are filing claims using the social security numbers of deceased individuals, or making workers' compensation claims for relatives who have died.
Ranney Pageler, VP of the Fraud Investigations Department at EMPLOYERS, says this type of fraud has been on the rise. The process of gathering information on claims and social security numbers is automated, Pageler says, but the department manually compares the data to find matches.
The company has also developed a proprietary data mining application that it uses on every new claim. Information is pulled from multiple systems, including claims underwriting and accounts payable, and the application looks for any activity that might indicate fraud. For example, if a job title of a claimant is not consistent with a particular injury, the system will indicate possible fraud.
EMPLOYERS enhanced its use of database analysis for fraud detection in 2003 and to date the technology has helped lead to more than 125 arrests and 70 convictions for workers' compensation fraud. Pageler estimates that during that time the firm has saved $38 million, much of it from either not having to pay out fraudulent claims or having money returned after fraud convictions.
NEW TOOLS
Web-based social networks sites such as Facebook, Twitter and YouTube are also proving to be quite effective tools for catching workers' compensation fraud, Pageler says.
In one recent case that was prosecuted, a man claimed he could not lift his arm and therefore couldn't go to work. "By going through social media such as Facebook, [investigators found that] the person had posted video of himself arm wrestling," using the supposedly injured arm, he says.
In another case Pageler cites, a man claimed to have injured his knee in a workplace accident. A search of Internet sites revealed that he'd recently posted an item saying he had injured his knee on a skiing trip. The posting also noted that the man's leg was set in a cast, but that he didn't have money for surgery. He subsequently took off the cast, went to work and faked the injury.
Carnahan says she expects to see a growing number of workers' compensation insurers use "automated red flag," programs that indicate possible fraud using business rules based on a variety of claim characteristics and a scoring mechanism.
Other emerging technology solutions Carnahan expects insurers to leverage in greater numbers include network link analysis tools, which enable firms to automatically review thousands of claims to look for signs of fraud. Also emerging are new identity resolution software applications, which enable insurers to identify individual claimants across multiple dimensions. For example, the software would catch individuals who are filing multiple workers' compensation claims using slightly different spellings of their names, or multiple addresses.
Additional promising technologies include predictive modeling, in which companies develop a model to try to best predict the probability of fraud; the analysis of unstructured claims data (for example, data not stored in fielded form in databases, such as adjusters' notes); and voice stress-recognition software, which can be used when taking claimants' statements to determine if they're being truthful.
"These products certainly can't stop fraud," Carnahan says. "What they can do is identify fraud more accurately and enable [law enforcement] to prosecute fraud more effectively and often."
Bob Violino is a New York-based business editor and writer and frequent INN contributor.
