By Eric Miller, Highpoint Partners

Even given all the attention on insurance fraud in recent years, it continues to be a growing problem in our industry, with the Coalition Against Insurance Fraud (CAIF) estimating a total fraud cost of $80 billion annually. While many states require insurance companies to establish Special Investigative Units (SIU) to combat fraud, the experts still predict a dismal future and a growing crisis. But, advanced software technologies available today, finally give the insurance industry a fighting chance to gain ground in the ongoing fraud war.

Insurance fraud has most likely been around since the birth or our industry in the seventeenth century, but the problem received little attention until the last 25 to 30 years when a combination of rising prices and organized insurance fraud caused insurers to begin to take a closer look at the issue. But, while we’ve learned much about the breadth and depth of the fraud problem and some measures have been taken to stop the bleeding, why does it seem that for every step forward, we seem to take two steps back as the bad guys get more and more creative with new methods of perpetrating fraud against insurance companies?

As insurance companies have taken proactive steps to put fraud-fighting methods in place through a combination of process and system modifications, the enemy of time continues to be the biggest hurdle to making the needed progress. Even when areas of fraud are detected, the time it has traditionally taken insurers to identify the trend, take the necessary action and implement the software and process changes, has left a gaping hole of time and opportunity; leaving the advantage in favor of the fraud perpetrator.

Part of the solution to this growing problem lies in educating the public and changing the attitudes of "it’s not my problem." However, the real answer is found in fraud-busting methods comprised of a combination of effective investigative techniques and advanced software-based approaches. While anti-fraud technology has been around for some time, the early generations of fraud detection tools often resulted in more time and effort than they were worth. The results are only as good as the analytical tools used and the information being accessed. And, unfortunately, in the valiant effort to detect fraud; bad data, disparate sources of information, and poor tools have frequently resulted in valid claims being flagged as suspect, requiring further investigation.

These suspect yet valid claims, also known as "false positives," further exacerbate the time problem and require unproductive investigation into the claim, requiring unnecessary time and cost investments, leaving a diluted effort in investigating truly fraudulent situations. The result is shaken insurer confidence in the available fraud detection tools and methodologies, with some insurers choosing to turn a blind eye to the "little white lies" and "harmless fudging." But with a million here and a million there…pretty soon we’re talking about real money.

The frequency of false positives often experienced with fraud detection tools and software of the past has not only resulted in the lack of confidence in the "flagged" occurrences, but has also left insurers challenged with a balancing act between investigation of suspicious claims and the perceived harassment of legitimate claimants. But, what’s an insurer to do?

The answer isn’t to stick our head in the sand, hope the bad guys mend their ways, or overlook the soft fraud to focus efforts on the big catch. The solution is in getting smarter about fraud detection and putting technology and software to work. After a few generations of early offerings, advanced analytical tools and software are ready for prime time in the insurance industry’s fight against the growing sophistication of today’s fraud perpetrators.

The first generation of anti-fraud software finds its roots in the 1970s when mainframes were the only available computing power. Not only lacking in processing capability, but expensive as well, fraud fighting methods were limited to rudimentary hard-coded programs. Software programs looking for specific, questionable codes, known suspects or known fraud circumstances, would flag claims that would then be reviewed individually and manually. Proprietary and limited in scope, it didn’t take long for the criminal to get the upper hand in this game of one-upmanship. They quickly learned which situations, such as claims over $100, would raise an insurer’s cause for concern. But the ever resourceful criminal would change their mode of operation by submitting not just one, but multiple claims for $75…no red flag, no suspect claims. The bad guys won again with their continually changing methods to beat the system and remain undetected in their fraudulent activities. Not only were the insurers continuing to lose ground in their war on fraud, but more salt was being rubbed in the fraud wound with the expense and time involved in the effort to modify the hard-coded logic in these far too limiting and inflexible technologies.

With the 1980s and 1990s came the opportunities for increased computing power at significantly reduced costs. Disk space was cheap, processing power was plentiful, and the ability for databases and data sharing showed real promise in the effort to control fraud. With the affordability and power of technology came the wave of relational databases and related tools in the form of query languages as a means of more sophisticated data analysis. However, like earlier tools the analyses were essentially hard-coded and limited by the initiating business analyst. The analyst needed to know what situation they were looking for in order to find the suspected fraud. Early software technologies in the form of rules-based tools proved useful in detecting patterns and uncovering suspected fraud. However, the continued "if, then, else" methods were still too reliant on the human element to define the suspected fraud conditions. The "look for the known problem" approach continued to give the bad guys the edge in our ongoing cat and mouse game.

Even once a fraud pattern was detected, it often required a good year to react and put the necessary process and program changes in place. By that time, the effort was often already invalid with insurers falling farther and farther behind in the fraud race. And since good fraud detection is so reliant on access to large volumes of good, historical data, the silos of data, prevalent in the insurance industry, continued to plague progress even as fraud detection tools and software continued to improve.

Today’s Advanced Tools and Software for Fighting Fraud

While those inflicting fraud have become cleverer and better organized, the industry’s technologies and means of detecting fraud have improved as well. And today’s advanced analytical tools and software finally enable the industry to get ahead of the fraudulent behavior and begin winning the battle.

New technology toolsets and software offer more advanced computational techniques that are able to "learn as they go," adapt to a constantly changing environment and reduce the reliance on the human element. While there are a variety of tools and software available, including HNC’s VeriComp Claimant and Intelligent Technologies’ Fraud-Spotlight; new companies, such as NuTech Solutions, offer an innovative approach. By attacking fraud through concurrent use of a variety of advanced predictive analytics and adaptive optimization techniques such as rough sets, classifier systems, and evolutionary programming, NuTech’s MERIX software combines new fraud detection techniques with the older-generation neural networks and rules-based tools. The results of combining the past (limited but useful tools) with the emerging technologies are significant increases in accuracy with drastic decreases in the occurrence of those costly and time-consuming false positives. 

In addition to the improved accuracy in detecting fraudulent claims, the predictive modeling capabilities of today’s fraud fighting tools provide the ability to automatically identify new fraud trends as they are happening. Due to the adaptive nature of the software, the tools essentially self-educate and change over time as they learn and grow through use. And, since the process is evolutionary, the software improves with use, continually enhancing accuracy and efficiencies through this advanced form of adaptive business intelligence.

More and Better Data Wanted

However, even with improved toolsets and software there is still the problem of less than perfect data, strewn across a variety of silo systems throughout the insurance operation. While today’s fraud software offers methods to organize the data and create a common view, the completeness, consistency and quality of the data continues to be a gap to the perfect solution. But, the beauty of the adaptiveness of today’s software tools enables an insurer to start with what data they have, in whatever condition. The iterative process and adaptive characteristics of the new software tools will leverage the amount of data and improve the quality of the decision process over time—as the system is used. And with this new form of responsive analysis…better results. The key is to not sit back and wait for the perfect answer, but to start with the data the insurer has, and improve over time. Without this proactive approach, those with a criminal mind will continue to wreak their financial havoc on insurance carriers.

In the past, insurance fraud detection was most reliant on the feet on the street to do the investigative work with a minimal amount of support from technology—an inconsistent, expensive and time-consuming proposition fraught with human error and false positives from unsophisticated analytical tools and limited, or even bad, data. With the advent of new and sophisticated software and toolsets the scales are tipping to a greater reliance on technology and this trend will only continue as technology matures and data availability (types and sources), quality and quantity improve.

Putting effective fraud fighting technologies and processes into action is not a one-time event, but rather a commitment to continuous improvement. Just as criminals have gotten smarter and craftier in the past, we can expect the same level of ingenious creativity to continue.

And beware of the Internet. For all its promise, this new medium will surely give birth to innovative new insurance swindles as consumers and criminals transact more and more online with the insurance companies. Fraud rings follow technology in search of new opportunities. Most certainly, the Internet is fertile ground for new classes of insurance fraud; especially fraud utilizing identity theft.

Insurance carriers are spending more on fighting fraud and can expect to spend significantly more. When it comes to technology, the challenge will be to take advantage of new fraud prevention and detection tools for all touch points of the insurance process; from policy procurement, delivery and claims adjudication.

Continuous retuning and refining of fraud detection systems is a necessity with effective and automated feedback loops for a better iterative process. If we’re relying on the human element to complete the loop, we open the process up to more errors, more time and potential gaps; not to mention the potential for internal fraud perpetrators to introduce holes into the process. The more automation — the more integrity and security in the system.

Increasing use of the Internet, a struggling economy and mounting consumer debt will lead to even more fraud. There will be no better time than the present to wage the "war on fraud," and undoubtedly, technology, in the form of advanced software and analytical tools, will be the weapon of choice for the insurance industry.

With better tools, better detection and the opportunity to get ahead in the fraud game, the insurer is finally able to quickly identify fraud trends and now has the ability to react in quick-time to positively affect the outcome. With the use of the toolset available in this new, adaptive software, the time advantage can finally be in favor of the insurer, and "time" is the ultimate prize in the battle against insurance fraud. 

About the Author:
Eric Miller, CPA, CFP, is senior principal with Highpoint Partners, LLC. and a member of the Association of Certified Fraud Examiners. Eric can be reached at ericmiller@highpoint-partners.com or 704-370-0117.