Some insurance companies and SIUs already have turned to fraud technology in order to help detect and investigate fraud. Specialty units within some SIUs now include information specialists and intelligence analysts who use data mining and other tools to support adjusters and field investigators. An insurer's fraud program can be retooled with sophisticated fraud technology solutions, but a good antifraud program requires a business strategy and tactical business processes to be integrated with the technology in order to be successful.

Predictive Modeling

The latest and most advanced fraud detection solution that has shown promise by demonstrating strong results is predictive modeling. Predictive and rules-based systems have emerged and are in production today with several insurers. They have the capability of flagging questionable claims at first notice of loss and throughout the claim lifecycle, making it a front-end claim fraud detection tool.

Generally speaking, there are two types of scoring models. One is predictive modeling and the other is a rules-based system. Predictive models are built from a company's historical claim data and fraud-investigation experience or known outcomes. Other modeling methods can be used if a company lacks historical tracking of fraud results and can be built for various lines of business. Rules-based systems typically use traditional industry red flags associated with claim fraud. Data from a claim is run against a set of predefined rules and violations are flagged. Rules-based systems are assumption driven and do not necessarily relate to changing industry or company-unique trends and patterns of fraud activity across claims.

Once the model is established, claim data is run through it and scored. If the score hits a predefined threshold, the claim is flagged. Indicators of potential fraud may be subtle, obscuring them from common recognition. It may be the absence or inclusion of certain information or the relationships and combinations of attributes present on a claim that the data model derives.

The potential benefits of predictive modeling can be significant. It provides a global and more consistent screening process across the claim organization. It helps to identify claims with the greatest propensity for fraud much earlier in the claim lifecycle. Claims that are scored, flagged, and then investigated for fraud typically have a much greater success rate. They also produce a higher return in the fraud-mitigation process.

Data Analytics

A major obstacle in developing actionable intelligence is meaningless information. The information may come from many different data sources and technology tools exist that can pull the data together. Unlike predictive modeling, these tools typically are back-end solutions used by data analysts and are transparent to a claim adjuster. They support the investigative process when fraud is suspected and some even offer proactive fraud-detection capabilities on the back-end. The ability to mine millions of claims and other records can have a significant impact on a company's fraud-fighting efforts and results.

Data visualization or link-analysis technology can join and transform seemingly unrelated pieces of data into meaningful information. Data can be pulled together from disparate sources and used to perform analysis. Visual-analysis software can be used to reveal patterns, trends, and relationships contained within complex data sets. Unlike statistical analysis, which deals mostly with aggregated results and reports, proactive and reactive analyses can be done to explore direct and indirect connections in data. These patterns and relationships emerge from the data and are presented graphically.

Many companies have been using public records and other information database sources in the investigation process. Supplementing the fraud detection and investigation process with external data could improve a company's ability to detect and investigate fraud, but the use and type of external data must be carefully evaluated for effectiveness. A company should always leverage and perform analyses on internal data before seeking external sources. The use of external data with link-analysis technology has been around for some time, but introducing and determining the value-add of external data in the predictive modeling process is still very much in its infancy.

Anti-Fraud Solution

An effective fraud solution requires a holistic approach to the problem. Support at all levels of the organization and a skillful blend and integration of technology, people, and the right business processes are needed. Using the right mix of technology in concert with human talent — claim adjusters, investigators, and analysts — is critical for a successful fraud program. It leads to better information and intelligence that delivers results.

Technology has the potential to provide substantial benefits to a company's fraud program through the use of electronically enhanced detection and investigation tools. These tools can significantly expand the ability to identify and investigate potential fraud over manual methods currently used today. Fraud technology alone is not a solution and a quality fraud investigation is still where the rubber meets the road; no amount of technology can serve as a substitute for this. It provides a level playing field and ensures consistency across the organization. Fraud technology does work, and matching good information intelligence with the right resources can make all the difference when it comes to fighting insurance fraud.

Put aside the speculative statistics on just how much the problem costs the industry and most everyone will agree that it is huge. Aggressively combating insurance fraud makes good business sense because it impacts the bottom line, affects a company's competitive position in the market place, and influences policyholders. Having a strong anti-fraud program can serve as a deterrent to those individuals looking to commit the crime and perhaps lead them elsewhere. Even small incremental changes in the fraud identification and investigation process can have a large impact on the fraud mitigation rate. For example, a quarter percent increase in the number of fraud cases referred to an SIU and investigated could result in millions of dollars more in migrated losses. Less than a 10 percent improvement in identifying quality SIU referrals could result in millions more.

Decades later, the next generation of fraud fighting is now.

David J. Rioux, CIFI, is assistant vice president and manager of the Corporate Security Department and Investigative Services for Erie Insurance. He can be reached at [email protected].