Data mining, rules-based tools, neural networking, predictive modeling, case management systems ... such is the growing smorgasbord of technologies that Pittsburgh-based Highmark Inc. is implementing to reduce fraud. Working with the Minneapolis, Minn.-based Fair Isaac Corp., the independent licensee of the Blue Cross and Blue Shield Association developed a fraud analytical tool, and is now running reports and doing data extraction in minutes and, sometimes, seconds rather than in hours, days or even weeks. The health insurer also implemented a post-pay solution from Fair Isaac that quickly ranked 21,000 providers, enabling Highmark to uncover a significant number of new cases of fraud. Furthermore, in the past year it has referred $9 million worth of fraud cases to law enforcement.
Highmark is eager to test a new pre-pay solution, given that it may only recover 20% to 30% of the funds in post-pay since the indebted entities have often squandered the assets, according to Tom Brennan, director of special investigations for Highmark.
INORDINATE FRAUD COSTS
Highmark is one of many insurance firms now initiating or deepening their outsourcing relationships with anti-fraud vendors in a drive to increase fraud referrals, enhance their reputation, improve loss ratios, lower loss adjustment expenses and boost other savings, which they can plow into added benefits for their customers. Such technology is as important today as ever, with insurers suffering enormous fraud costs, and a growing sense that their victimization may be worsening. With a credit and mortgage crisis buffeting the U.S. economy as well as European countries such as the United Kingdom and Spain, there are signs that mounting financial hardship is spawning more fraudsters. "As the economy tends to slip and soften, we see an increase in crime," says Mark Lowers, president and CEO of Lowers & Associates, a risk mitigation service provider based in Purcellville, Va. "Because unemployment begins to creep up, people are stressed over losses of homes, and a weakening economy pushes people over the edge sometimes."
Of the 45 state fraud bureaus that the Coalition Against Insurance Fraud tracks, about one-third of the states say they are witnessing an increase in the number of open fraud cases involving mortgage and title issues, according to Dennis Jay, executive director of the Washington-based organization. While home arson fires related to foreclosure have seen "healthy increases" in the last 18 months in several areas of the country, the number of automobile arson fires is particularly startling. "This year will likely set a record year for people getting rid of unwanted vehicles, whether by arson or other means," says Jay.
In the United Kingdom, attempts to commit fraud by including material falsehoods on application forms for insurance products (as well as loans and credit cards) rose 13% to total 21,780 cases in the first quarter of 2008, up from 19,239 from the year-earlier period, according to CIFAS, a fraud prevention service based in London.
"The increase in application fraud (often to hide adverse credit history) demonstrates that because people are getting into debt earlier, and because the 'credit crunch' has diminished their access to finance, they are now resorting to fraudulent applications for funds," says Peter Hurst, the company's chief executive.
Fair Isaac, for one, has seen escalating interest in fraud technology in the U.K. "Right now the market is kind of forcing insurers to get more and more on top of this," says Mike Gordon, VP of emerging industries for the technology supplier.
THE LURE OF OUTSOURCING
In large part, insurers are outsourcing fraud prevention and detection technology work to minimize potentially massive losses. For instance, 3% of all healthcare spending-or $68 billion-is lost to healthcare fraud, according to an estimate by the National Health Care Anti-Fraud Association (NHCAA), headquartered in Washington.
The NHCAA considers this estimate conservative, and others place the figure closer to 10%. In any case, an insurer that pays out $5 billion on claims is throwing away tens of millions of dollars on fraudulent ones. "And those costs get passed on to the consumer," explains Highmark's Brennan. Though not quite as severe as healthcare fraud, P&C fraud also is fierce: some $15 billion to $20 billion in fraud is perpetrated against such firms, according to Gordon.
Many carriers find greater expertise externally than internally, as specialized vendors are able to follow scams day-in-and-day out, and keep up with the tactics of fleet-footed fraudsters.
"Criminals are getting more and more sophisticated in terms of how they steal," says Frank Scheckton Jr., president of the Windsor, Conn.-based fidelity and crime division for the Great American Insurance Group. "You need to go to somebody who [fights fraud] as a full-time job because, for the thieves, it's a full-time job."
That's why Great American, which insures stockbrokers, banks and others, outsources risk consultation, risk assessment, vendor screening and other fraud mitigation work to Lowers & Associates. The insurer also turned to Lowers because many of the company's officials are former law enforcement or intelligence service bigwigs that "have a lot of connections, and see a lot of things and hear a lot of things," says Scheckton. Given the compelling vendor options available, outsourcing fraud mitigation is rampant. Few insurers have their own data, access to their own analytic tools or the right people to use those tools, explains Donald Light, senior analyst in the insurance group of Celent, a research and consulting firm based in Boston.
OUTSOURCING FEARS
Still, some carriers balk at outsourcing, largely out of concern that their vendor costs will spiral. Many risk managers report to CFOs and are rewarded by reducing their insurance budget, says Scheckton. Shirking anti-fraud measures is thus viewed as a cost saving. Yet, this tunnel-vision approach proves costly long term as fraud adds up. Great American takes the opposite tack, focusing significant resources on fraud prevention to boost its reputation. "We consider ourselves sort of the Nordstrom's of the crime insurance world, so if you want to pay a higher price, get top-notch service, and get great claims service and loss control, then you should come to us. But we are not going to be the cheapest in the market," says Scheckton.
Security concerns also can dissuade some insurers from outsourcing. "We were very reluctant to ship our data offsite," admits Brennan. "We had never done that before." Yet, Highmark overcame such fears by involving its legal and privacy departments in contract negotiations, and by concluding that Fair Isaac's security and integrity were "well spoken for." Further, Highmark structured its vendor contract so that Fair Isaac's development teams outside the continental U.S. could not look at its data.
MAKING OUTSOURCING WORK
Indeed, carriers can implement a range of shrewd tactics to help make outsourcing fraud prevention and detection efforts effective. For one, carriers should set parameters with their vendors, ensuring that they rein in fraud costs by establishing what constitutes an appropriate charge and how much time the vendor should spend on particular technology work, advises Scheckton. Those on the front lines of fraud mitigation also should make sure that senior management buys into the importance of anti-fraud efforts. At Highmark, anti-fraud work is a corporate strategic initiative, and 23 analysts and representatives handle fraud cases.
Insurers need to seek solutions that go beyond the much-embraced, but less-sophisticated anti-fraud techniques and tools. For instance, many insurers continue to use red flags or rules of thumb to crack down on fraud. They train claims adjusters to look to see if some five to 10 elements are present in, say, a personal injury auto claim, and then have the adjuster refer the case to the special investigations unit (SIU) if it finds two or three red flags. Such static work does not enable carriers to keep up with crafty fraudsters, says Celent's Light.
In contrast, Falls Church, Va.-based Computer Sciences Corp. (CSC), for instance, sometimes runs 18,000 to 20,000 claims during an evening through its Fraud Evaluator system. Each claim is evaluated consistently, and quality referrals are pushed to the top (which are those first brought to the investigators' attention). This can weed out many false positives-a major problem-and enable carriers to reorganize the roles and responsibilities of claims and SIU officials in light of the referrals, according to Lewis Rogers, director of P&C fraud solutions for CSC.
INSURER RESPONSIBILITIES
While vendors can help an insurer detect fraud, the latter needs to figure out ways to better structure its data and organization to reduce fraud. "We can identify the fraud, and we can work with [insurers] to do a good job of detecting it and give them ideas on how to handle it but, ultimately, they have to spend the time and they have to decide: Are they going to handle it, and how are they going to change their SIU," says Gordon. In fact, carriers only should outsource after they have a good sense of the data they control and how they process claims, Gordon says, noting that with a rash of insurers adopting new claims solutions in recent years, many are still organizing their data.
Finally, insurers should be realistic about how much fraud they can scotch. "Every time you plug one of the holes of the dike, another hole opens up because there is someone out there 24 hours a day trying to figure out how to get around the system," says Scheckton. And fraud can be so rife at insurers that vendors such as Fair Isaac find their technology catches too much fraud for an insurer to know what to do with. Insurers need to segment the fraud, prioritizing that which must be squelched first.
Using cutting-edge technology and techniques will not make fraud disappear, fighting it vigorously requires a balanced approach. "I don't think that a computer can be built to compete with the human mind," observes Brennan. "So there will always be post-payment. You are never going to be able to pick up on everything. But with predictive modeling and the pre-pay solution, we should be able to manage things a lot better than we currently do."
HOTTEST ANTI-FRAUD TECHNOLOGIES
While insurers are implementing a wide variety of technologies from vendors to detect and prevent fraud, among the most prevalent are predictive modeling, neural networks, red flags, claims databases, link analysis, profiling and identity matching.
Case management tools also are increasingly in vogue, according to Donald Light, senior analyst in the insurance group of Celent, Boston. Once a potential fraudulent claim is sent to the SIU, it becomes a matter of investigation. The first step in the investigation of potential fraud takes place with the adjuster, who marshals facts, analyzes them and tries to get additional follow-up leads. "From a technological point of view, there are, in the past year or two, more solutions being offered to help the SIU and/or the adjusters manage cases" of fraud, explains Light.
Highmark Inc., Pittsburgh, for one, has implemented a robust case management system, which allows it to scan all documentation so that it is available electronically, says Tom Brennan, director of special investigations for the healthcare insurer. The carrier also is a proponent of rules-based tools, which are being widely adopted. "If you have an issue you can build an algorithm, go into your data and find out just how bad you've been had," Brennan says. Yet, Highmark felt the need to go beyond business rules, and thus embraced predictive modeling, which uses statistical analysis to establish the probability of a claim being fraudulent.
In recent years, link analysis-in which the connection among people, places, and things involved in multiple claims is examined-has gained currency. While ideal for investigating fraud rings, link analysis has not been particularly effective in driving down false positives, says Mike Gordon, VP of emerging industries for Fair Isaac, Minneapolis. But Gordon believes link analysis will be a technology to watch in the next year or two, as a second, more sophisticated generation arrives. "It doesn't really give you kind of a score-per-link relationship" now, says Gordon. "The good thing is that it identifies the relationships, but if you start to score, and you start to think about it, and look at it, and you start to use the models together, then you start to really home in on the [fraud cases] that you want to focus on."
Carriers, too, will increasingly look into "optimization" technologies to better handle fraud, predicts Gordon. For instance, an insurer could use a combination of rules-based systems with analytics and link analysis to allow SIU teams to triage more effectively the fraud claims coming in, to adjust as necessary when the insurer offers new coverages and to start to truly predict what fraud might occur.
Daniel Joelson is a freelance business writer based in Alexandria, Va.
(c) 2008 Insurance Networking News and SourceMedia, Inc. All Rights Reserved.
