Text Mining an Emerging Tool in Anti-Fraud Technology

September 29, 2014

Insurer use of anti-fraud technology is rising, according to a report released this month by the Coalition Against Insurance Fraud. The study reports that as of 2014 nearly all insurers (95 percent) said they use anti-fraud technology, compared to 88 percent in 2012.

Evidence that fraudulent activity is increasing bolsters the growing business need for technology-based solutions. More than half the insurers surveyed said that suspicious activity has risen over the last three years and that it flows through the entire claims cycle.

Fraud schemes today have not only increased in number, but shifted away from auto theft and more towards bodily injury and suspicious medical provider activity, according to the study. Insurers, needing to adapt to these tactical changes, are increasingly adopting advanced analytics to combat the evolving nature of fraudulent activity.

Advanced technological tools may help limit the rise in fraudulent activity, according to the report. The survey found that while 81 percent of insurers use basic technology tools, far fewer employ more advanced technology, such as link analysis (50 percent), predictive modeling (43 percent), or text mining (43 percent).

The most common challenge faced by 53 percent of insurers when considering implementing advanced anti-fraud technology is the lack of IT resources. Among other obstacles, SIU’s inability to process the large volume of potentially fraudulent claims was cited by 6 percent of insurers.

The increased presence of organized crime rings is a noteworthy development in the evolution of fraudulent activity. Text mining is one anti-fraud tool that can be utilized to expose these crime rings at the beginning of the claims process.

Text mining can connect information that is otherwise unstructured data buried in multiple, separate places, i.e., adjuster field notes, email, medical records, and police reports. By using text mining, insurers can explore this data to discover previously unknown concepts and patterns. For example, insurers use text mining to discover scripted comments in claims notes and call center logs, allowing them to notice incidences where multiple, allegedly unrelated, claimants say the same thing.

The study concludes that an anti-fraud strategy incorporating advanced technology and tools will result in a much higher fraud detection rate, thus ultimately cutting overall costs for insurers.

The report, titled “The State of Insurance Fraud Technology,” is published jointly by the Coalition Against Insurance Fraud and SAS. Click on the link to read more.