For most insurers, there's a natural inclination to settle claims swiftly, particularly to satisfy customers who might take their business elsewhere. But as insurers improve claims processing time, ironically, many are paying a high price.To their chagrin, efficiencies surrounding claims settlement can actually produce negative results. Ultimately, many carriers find themselves dispersing dollars that shouldn't have been paid in the first place, due to claims fraud.
Insurance fraud is increasing more each year, with the abuse particularly impacting health, workers' compensation and auto insurance lines. Fraud ranges from sophisticated crime rings involving multiple parties to cases where individuals or medical health providers act alone to abuse the system.
In response, a growing number of insurers have begun to implement sophisticated fraud detection tools to support manual methods, according to a survey of health care payers by San Rafael, Calif.-based Fair Isaac Inc. However, 80% of surveyed payers engage only in post-payment detection where once claims are paid, they're later difficult to recoup in their entirety.
Pay and chase
"Post-payment investigation and recovery are costly processes, and months or years go by before insurers recover their money," says Joel Portice, Fair Isaac's vice president of health care solutions.
"Many existing systems rely primarily on rules mixed in some cases with statistical analysis. But prepayment and post-payment systems combine rules and statistical analyses with predictive models. They are capable of identifying far more complex data interrelationships."
Executives for Hollywood, Fla.-based Vista Health Plan say they grew weary of so-called "pay and chase" collection tendencies. The company therefore implemented a pre-payment fraud detection solution supported by predictive modeling.
"Having a pre-payment program is critical because once a payment is out the door, you've lost the battle," says William Rushton, director of internal audit and fraud prevention, Vista Health Plan, which has 330,000 members in its network. "With pre-payment fraud, you can recoup 100% of potential fraud. I estimate that with post-payment fraud-or what I call retrospective recovery-we can only recover 10% to 15% of claims payments retrospectively identified as fraud."
In post-payment fraud instances, insurers must rely on complete data; their actions are cost-justifiable only with big-ticket claims; and an insurer must often spend legal and administrative costs to recover dollars. "And, under this methodology you're not effectively controlling fraud on the front end," Rushton adds.
Vista Health Plan could not afford to sit idly by and hope to recover a fraction of claim payments later. Doing business in Florida was a reason why. Fraud occurs with impunity in the Sunshine State due to the transient nature of the fraud perpetrators.
In fact, Washington, D.C.-based National Healthcare Anti-Fraud Association (NHCAA) estimates that in South Florida alone, public and private health insurers have been defrauded of hundreds of millions of dollars in recent years by organized crime rings.
Nationally, the NHCAA believes 10% of every dollar spent on healthcare is lost to fraud, draining the system of $150 billion a year that could reduce consumer costs, reward shareholders and lead to improved treatments.
Facing a host of challenges, Vista Health Plan identified prepayment solutions as the way to maximize fraud reduction practices. Pre-payment analytics can be delivered in real-time, enabling risks to be analyzed accurately even when data is incomplete. In most cases, resolving fraud on small-ticket claims doesn't benefit insurers, but with a predictive modeling solution, action is cost-justifiable with all claims, Fair Isaac reports.
With a predictive modeling program in place to stanch fraud at a pre-payment stage, an insurer might prevent a bogus claim payment from being paid, "but what you have really done is stopped the fraud pipeline, because the bad guys will see they can't get away with it," Rushton adds.
Many industry participants believe to make a significant dent in incidents of fraud, identifying the right mix of technology to fight fraud is crucial.
Insurers have to assume the mentality of "chemist" when it comes to fighting fraud, states James Quigley, director of communications for Washington, D.C.-based Coalition Against Insurance Fraud (CAIF).
"The key is finding the right innovative blend of technology. It could comprise link analysis, link charting, predictive analysis and scoring tools all in combination," Quigley explains. "Some combinations don't work for some insurers, and overall it's costly to identify what the combination should be."
The push to predictive models and data-centric strategies is occurring because the price for data mining technology is falling while the power of computing is increasing, Quigley states. "Data mining is no longer the sole domain of the banking or credit card industries. Smaller insurers are getting hold of these technologies."
Health insurers are among those that need to find the right fraud reduction technologies-and quickly. In this insurance segment, Quigley says, fraud is "bubbling up beneath the surface, fueled by mastermind crime rings."
Vista Health Plan has been exposed to a wide range of potential fraud schemes over the years. "A medical services provider can add services that did not occur or enlist a current member to serve as the vehicle for fraud," Rushton says. "Providers deliver a claim to their insurance company and that starts a fraudulent process. There are no established patterns with fraud. There's not one type of provider that is any more active with fraud than another."
Since fraud detection can be like finding a needle in a haystack, this makes it essential to have predictive analytics in place.
"Rules-based technology can only take you so far," Rushton explains. "Predictive methods are where we're focusing our attention. Predictive analysis looks at the data, compares the data based on theories and conflicts and scores it for further investigation. Rules-based methods emphasize networking with affiliates. It's not data-centric and it's also not performed in real-time."
As part of a pilot project that will become enterprisewide in July, Vista has implemented Fair Isaac's Payment Optimizer to help identify patient and provider fraud, abuse and errors.
An individual who schedules treatment for physical therapy while simultaneously seeking treatment from an ear, nose and throat specialist might represent fraud. The likelihood of these two events occurring together is rare. "The system scores the activity and says, 'you might want to take a closer look at this.' Without predictive modeling, there would be no red flags," he adds.
Other industry observers agree that the push to predictive models and data-centric strategies is happening, due in large part to a decrease in the cost of supporting technologies.
Predictive modeling to combat fraud is "the wave of the future," declares Pat McCann, corporate special investigative unit, process manager for Mayfield Village, Ohio-based Progressive Casualty Insurance Co., which is in the exploratory stage of implementing a predictive analysis solution.
"Using predictive modeling at the first notice of loss lets a claims adjuster identify if this is a suspect claim and whether it needs to be triaged to an SIU or not," McCann explains. "For years, the use of business rules or decision trees helped insurers identify fraud, but predictive modeling takes fraud management to another level."
Progressive created an internal investigations department to analyze data and conduct investigations in which offenders are referred to law enforcement for criminal prosecution. Progressive also employs more than 330 special investigation professionals that focus on external fraud, McCann says.
The Progressive team works in conjunction with law enforcement agencies, the National Insurance Crime Bureau (NICB) and state departments of insurance to help suppress auto theft and insurance fraud. This, in turn, helps keep insurance rates down for Progressive customers, McCann proclaims.
For the past six years, Progressive deployed a data mining tool to conduct visualization and pattern analysis for both open and past claims.
"Using data mining, we can conduct ad-hoc queries, where we enter a phone number or a medical services provider, and the system extracts data from our claims repository," McCann explains. "It slices and dices data returned from the repository to determine if a claim looks suspicious. But the data mining process requires manual intervention while predictive modeling runs on its own."
CAIF's Quigley says that data mining and predictive analysis is supported by software that "stretches the time and space continuum, and (the technology) has become far more 'human' in its composition." These solutions have garnered acclaim for not only detecting fraud, but for ruling it out.
That's because insurers often chase "phantom" fraud. "There are instances where fraud is not being committed-even though it appears to be," explains Rushton. "A medical provider might experience a glitch in its operating systems, or perhaps there's a simple keying error. We can now detect who the fraud perpetrators are and who are the companies that are in need of a systems repair."
But clearly, if a claim looks suspicious, the chances are greater that it involves fraud. "By harvesting knowledge across our company, we've been able to greatly reduce having to even deal with post-payment fraud," says Ann Castro, chief software architect, claims, Columbia, S.C.-based Blue Cross Blue Shield of South Carolina.
"We use a solution that we call ClaimCheck that addresses pre-payment fraud reduction. It can examine coding to determine improprieties such as over-billing by medical providers, duplicate billing-all of these things can be detected before we make a payment. We can also detect whether this is a case of fraud abuse or simply human error or a technology glitch. If there's no abuse, it's no issue and there's no follow up. But overall, we can slice and dice data to see patterns of practice with claims."
Blue Cross Blue Shield of South Carolina has a module designed within ClaimCheck called SmartSuspense that enables insurers to take appropriate pre-payment action for claim fraud.
"We route suspicious claims to registered nurses (RNs) who have the clinical knowledge to know what steps to take. Our RNs access our claims databases and make a determination on whether a claim should be investigated for fraud," explains Castro.
Can Voice Technology Silence Fraud?
Fighting fraud is becoming more meaningful to most insurers, research indicate. The Washington, D.C.-based Coalition Against Insurance Fraud (CAIF), for instance, found that 86% of insurers track the percentage of claims they refer to their SIUs and that 39% refer between one percent to three percent of claims.
CAIF found that 80% of insurers track how much money anti-fraud activities save their companies, while 11% of insurers rate their investigators based on how much money they save the company.
James Quigley, CAIF's director of communications, says health insurers have had a tough year dealing with complex fraud schemes, such as "rent a patient" where a fraud perpetrator recruits low-income or minority patients and pays them a substantial sum of money to visit clinics solely for unnecessary diagnostic tests or surgery. The clinics then bill the insurance company an inflated amount for procedures that were done.
Auto insurance fraud has been marked by crime rings where accidents are either staged or are regarded as phantom accidents. For life insurers, the technology to detect fraud is "very splintered," Quigley says. "Property/casualty claims are all supported by ISO's ClaimSearch database, but life insurers don't have one unified solution to rely on."
With their backs against the wall, insurers will have to find the next breakthrough in technology to add to their "mix."
One emerging fraud reduction technology that's garnering acclaim is voice stress analysis (VSA). "When people lie, their voices change. We would be able to hook this module into our existing ClaimCapture application for fraud," says Will Fulton, president of Charlestown, Mass.-based technology solutions provider First Notice Systems.
"British insurers have had success with VSA," Quigley says. "It's new on the horizon. It functions like a lie detector test, registering stress within a person's voice. It suggests that there might be deception occurring. It's one more tool to refer to an SIU for follow up."
But the downside of VSA is that it's very "Big Brotherish," Quigley adds. "People might consider it an invasion of privacy. And it would not bode well if VSA implicated someone for possibly committing fraud, only later to be found innocent. These are all thing that insurers have to think about when they expand their programs."
Keeping Score On Fraud
Claims scoring has emerged as an effective way to mitigate losses stemming from fraud. Insurers begin by establishing a rules-based system. Scores are predicated on the rules. If a score reaches a certain threshold, an insurer flags the claim and takes appropriate action.
Will Fulton, president of Charlestown, Mass.-based technology solutions provider First Notice Systems, says that "insurers place a weight on custom fraud indicators during call flow to a call center."
First Notice Systems offers insurers ClaimsCapture to detect suspicious claims earlier in the process. The solution is largely targeted to auto insurers. A score is tallied for all fraud indicators and the score is compared against custom thresholds established by insurer.
Red flags indicate to fraud investigators when a claim should be dissected. For instance, a claim that has a post office box listed rather than a street address can be a fraud indicator, and increase a score. So is a claim filed on a policy that was recently bound.
"Insurers are able to modify the rules that add weight to a score. They might find that one rule is adding to a score and bringing back a high number of suspicious claims. On second review, they might determine that certain rules should be deleted because the system is being inundated with fraudulent claims," explains Fulton.
Ultimately, the objective is to "capture, grade and route this information to the right people in an organization. If it takes us 10 minutes to gather data on a potential fraud, an SIU at an insurance company has the information 30 seconds later."
San Rafael, Calif.-based technology provider Fair Isaac Corp. has a solution offered to insurers where, on a scale of 0 to 1,000, a claim scoring 200 would indicate a lower risk than one scoring 800.
Insurance companies can set thresholds, using rules to automatically route all claims scoring above a certain number to investigative staff, according to Fair Isaac.
Insurer Tendencies In Fighting Fraud
Most insurers track percentage of claims referred to their special investigation units (SIU) and measure how well SIUs save the company money. These are among the findings of a survey of 65 SIU managers at April's Insurance Fraud Management conference in Phoenix, sponsored by the Coalition Against Insurance Fraud:
- 86% of insurers track the percentage of claims they refer to their SIUs.
- 39% refer between one to three percent of claims.
- 80% track how much money anti-fraud activities save their companies.
- 90% indicate savings reports to state regulators do not accurately reflect their anti-fraud efforts.
- 11% of insurers rate their investigators based on how much money their save the company.
Register or login for access to this item and much more
All Digital Insurance content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access