Detecting insurance deceit and reducing losses

Česká pojišťovna uses SAS® Fraud Framework for Insurance to spot and prevent claims fraud, saving hundreds of thousands of dollars annually

There are countless ways to defraud an insurer. Stage a car crash. Fake an injury. Burn down a house. Make a false health insurance claim. Even fake your own death to collect life insurance. With fraudsters trying these tactics every day, rip-offs represent an estimated 10 percent of all claims for European insurers.

These costs are more than just illegal uses of insurance funds. Insurance firms are wasting the money from honest customers to pay the “bad guys.” To avoid this, Česká pojišťovna, the largest insurance company in the Czech Republic, uses analytics technology to avoid writing suspicious policies and uncover improper claims.

By using SAS Fraud Framework for Insurance to analyze all new policies and detect suspicious new contracts, Česká pojišťovna has saved tens of millions in Czech crowns each year.

Zdeněk Dragoun
Fraud Analyst

Uncovering suspicious activity leads to substantial savings

“By using SAS Fraud Framework for Insurance to analyze all new policies and detect suspicious new contracts, Česká pojišťovna has saved tens of millions in Czech crowns each year,” says Zdeněk Dragoun, Fraud Analyst in Česká pojišťovna’s Fraud Detection Department. “Additionally, we have been able to uncover cases of insurance fraud totaling 20 million Czech crowns (US$820,000) annually that would have otherwise gone undetected.”

Previously, the insurer investigated 1,000 new claims each day, relying on experienced field technicians and claims adjusters to select and monitor individuals from claims groups with high fraud potential or follow up on tips. These techniques required an army of adjusters – and success was often a matter of being in the right place at the right time.

Today Česká pojišťovna takes advantage of a hybrid analytic approach that combines multiple advanced statistical methods, including rules-based detection, predictive analysis and social network analysis. All of these methods increase the efficiency and effectiveness of the company’s fraud detection analysis.

Testing against rules to trigger investigations

Rules-based detection tests each transaction against a predefined set of algorithms or business rules to discover known types of fraud based on specific patterns of activity. These systems flag claims that look suspicious due to their aggregate scores or relation to threshold values. Experienced adjusters then investigate flagged claims more thoroughly.

Česká pojišťovna has developed nearly 300 rules – and continually adds now ones – based on its experience investigating thousands of claims in various market segments. “For example, rules trigger an investigation when claimants make frequent claims or add or increase homeowners or auto insurance coverage shortly before submitting a claim,” explains Dragoun. The insurer has even developed rules that look at seemly unrelated information to uncover larger fraud rings and even organized crime.

Using predictive analysis to produce fraud propensity scores

Predictive modeling uses past patterns to anticipate which situations are likely to be fraudulent in the future, reducing the need for hands-on accounts management. For example, physicians and clinics may bill insurance companies for services never rendered or for procedures that weren’t medically necessary. They may even charge too much for certain services or charge multiple times for the same service. Predictive modeling can uncover suspicious bills and billing patterns from these health care establishments.

Česká pojišťovna has employed data mining tools to build predictive models that produce fraud propensity scores. Adjusters simply enter data, and claims are automatically scored for their likelihood to be fraudulent. Česká pojišťovna takes advantage of a data mart containing 1,000 variables and 40,000 observations to score 100,000 cases each day. Originally, the company used six models to perform this scoring. It now takes advantage of combined models that incorporate connected decision trees, linear regression and neural networks.

Discovering social connections

Social network analysis enables Česká pojišťovna to uncover links between individual subjects associated with the insurance company (individual people, cars, addresses, telephone numbers, etc.). Social network analysis tools can be tuned to display link frequencies that exceed a programmed threshold. Large volumes of seemingly unrelated claims can be checked and then patterns and problems identified.

Analysis of social relationships gives Česká pojišťovna’s investigative team a visual map and statistics to help them better understand relationships in risky communities. Social networks can reveal fraud such as multiple claims within a short period of time from related parties, such as members of a single family or the classic ring associated with staged accident scams. Not only does the company use this capability to avoid paying fraudulent claims, it also uses it to check new connections to historical fraud to avoid perpetuating fraud schemes.

In the future, Česká pojišťovna plans to continue to expand its use of SAS to find new ways to fight fraud. The company is investigating a number of new technologies, including text mining, geographical analysis using GPS coordinates, and even analysis of images, photographs and voice.

Česká pojišťovna

Challenge

  • Avoid writing suspicious insurance policies.
  • Detect fraudulent insurance claims.

Solution

SAS® Fraud Framework for Insurance

Benefits

  • Saved tens of millions in Czech crowns annually by better detecting suspicious new contracts and clients.
  • Uncovered cases of insurance fraud totaling 20 million Czech crowns annually that would have otherwise gone undetected.
The results illustrated in this article are specific to the particular situations, business models, data input, and computing environments described herein. Each SAS customer’s experience is unique based on business and technical variables and all statements must be considered non-typical. Actual savings, results, and performance characteristics will vary depending on individual customer configurations and conditions. SAS does not guarantee or represent that every customer will achieve similar results. The only warranties for SAS products and services are those that are set forth in the express warranty statements in the written agreement for such products and services. Nothing herein should be construed as constituting an additional warranty. Customers have shared their successes with SAS as part of an agreed-upon contractual exchange or project success summarization following a successful implementation of SAS software. Brand and product names are trademarks of their respective companies.

Back to Top