CNA detects, prevents insurance fraud beyond expectations

Within two years of implementing SAS® Fraud Framework for Insurance, one of the largest US insurers had opened 101 new provider investigations with an exposure of $18 million. At the same time, savings generated from the implementation of four predictive models reached over $6.4 million.

To many criminals, insurance companies make easy and attractive targets for fraud - from single individuals misrepresenting the extent of their losses to large networks perpetrating elaborate schemes. The FBI estimates that fraud costs insurers as much as $40 billion each year.

For CNA, the US' eighth largest commercial lines carrier with annual revenues of more than $9 billion, fighting fraud is mission-critical. According to Tim Wolfe, Assistant Vice President of CNA's Special Investigations Unit, elements of fraud may be found in as many as 10 percent of the claims that CNA processes.

"We train our adjusters to identify the red flags," he says, "but, we reached a point where we understood that we had an opportunity to do a better job in both identifying likely fraud and avoiding the wasted expense of investigating false positives."

State-of-the-art fraud combat

CNA operates in four key segments that are most attractive to fraudsters: commercial property, commercial auto, general liability and workers' compensation.

At one end, workers’ comp cases include disability and medical claims for fabricated or exaggerated workplace injuries. At the other, networks of affiliated providers submit claims for improper, excessive or nonexistent services to rack up thousands of dollars in unearned reimbursements.

"There are new and highly sophisticated schemes emerging all the time, so we were constantly vulnerable to new threats," says Wolfe. "Our Executive Vice President of Claim, George Fay, who has a strong background in military intelligence, challenged us to find state-of-the-art ways to improve our fraud detection results."

Our employer customers are as eager as we are to root out fraud. One or two bad claims can put a small company out of business. So, they really appreciate that we have this program in place on their behalf and ours.

Tim Wolfe
Assistant Vice President, Special Investigations Unit

20 percent hit rate

The company chose the SAS Fraud Framework for Insurance. With SAS, CNA built predictive models and now runs weekly analyses against its structured claims data as well as text notes from adjusters.

"We have an excellent working relationship with SAS," says Wolfe. "They took the time to learn from us and truly understand the nuances of claims fraud at CNA so that we could build effective predictive models for each line of our business.

"Each Monday morning, after a weekend data run, SAS provides our staff with a percentage of claim alerts that score high for fraud potential. Right now, we're reviewing about 100 alerts a week, and we're finding that we are averaging a 20 percent hit rate – about one in five alerts that we review is a good case for investigation."

One quarter, 15 new cases

Next, CNA uses SAS® Social Network Analysis to find broader patterns and connections among providers indicative of fraud conspiracies.

"We have a separate team investigating the provider networks, and SAS is having an important impact there," says Wolfe. "Not many providers can deliver that visual representation and the links to the individual entities that are potentially perpetrating these larger-scale frauds."

"These investigations can sometimes take months or years to reach fruition. But in just the first quarter, we initiated 15 different investigations. It is expected that these fraud rings will prevent up to $20 million in fraudulent claims. SAS found more viable cases than we'd anticipated."

$6 million savings right away

"Industry research indicates 10 percent of all claims contain an element of fraud," says Wolfe, and as of the end of 2010, CNA was seeing just 3.7 percent of its claims referred as potential fraud. "That’s considerably below the industry average, and we wanted to find out how much we were missing that wasn’t identified by our adjustors."

Within two years of full implementation of the models, CNA had seen the ratio of cases flagged for potential fraud rise to 8.1 percent, resulting in recovered or prevented fraudulent claims totaling over $6.4 million, directly attributable to SAS.

And investigations now operate more efficiently because CNA can focus on high-likelihood cases instead of false positives.

"The implementation of such groundbreaking technology and its impact inspires confidence and gratitude among our customers," Wolfe says.

"Our employer customers are as eager as we are to root out fraud," he adds. "One or two bad claims can put a small company out of business. So, they really appreciate that we have this program in place on their behalf and ours."

The purpose of this article is to provide general information about CNA and its current Claim strategies. Given the unique nature of CNA strategies, they may or may not be appropriate for use by other organizations and may be subject to change without notice. In addition, this article may contain views expressed by Mr. Wolfe that are his own and may not necessarily reflect those of CNA. CNA is a registered trademark of CNA Financial Corporation.


Increase the rate of accurate fraud detection and successful fraud prevention.


SAS® Fraud Framework for Insurance


  • Within two years, CNA recovered or prevented fraudulent claims totaling over $6.4 million.
  • 101 new provider investigations launched with more than $18 million exposure.
  • More accurate detection and fewer false positives improve efficiency and value of investigations.
  • Greater confidence among key stakeholders.
Hasil yang diilustrasikan dalam artikel ini khusus untuk situasi, model bisnis, input data, dan lingkungan komputasi tertentu yang dijelaskan dalam pernyataan ini. Setiap pengalaman pelanggan SAS berbeda berdasarkan variabel bisnis dan teknis, serta semua pernyataan tidak akan dianggap khusus. Penghematan, hasil, dan karakteristik kinerja aktual berbeda, tergantung pada konfigurasi dan kondisi masing-masing pelanggan. SAS tidak menjamin atau menyatakan setiap pelanggan akan mencapai hasil yang sama. Satu-satunya jaminan untuk produk dan layanan SAS adalah yang tercantum dalam pernyataan jaminan tersurat dalam perjanjian tertulis untuk produk dan layanan tersebut. Segala hal dalam pernyataan ini tidak akan ditafsirkan sebagai jaminan tambahan. Pelanggan telah membagikan keberhasilan mereka dengan SAS sebagai bagian dari pertukaran kontrak yang disepakati atau keberhasilan proyek setelah keberhasilan penerapan perangkat lunak SAS. Merek dan nama produk merupakan merek dagang masing-masing perusahaan.

Back to Top