Cracking Non-compliance

WorkCover NSW recovers millions in premiums with SAS Data Mining

WorkCover NSW is the New South Wales (Australia) government organization responsible for managing workplace safety, injury management, and workers compensation systems. Its philosophy is "prevention is better than cure" and works with industry, the workforce, and insurers to promote a culture of safety through public awareness programs, education and other community activities.

Issues of non-compliance in all areas of workers compensation have added to the workload of WorkCover NSW for many years. Traditionally, investigating areas where suspected non-compliance is occuring has been resource-intensive.

However, several projects are now being implemented that are enabling the organisation to not only track down where underpayments and non-compliance are occuring, but are also allowing it to identify high-risk industries where safety is an issue.

Like many data intensive organisations, WorkCover has accumulated huge stores of data that are located across incompatible systems. One of WorkCover NSW's main roles is to provide compensation money to injured workers. The organisation holds 350,000 policies and receives around 100,000 claims every year. Such a large number of policies and claims makes investigating suspect claims a logistical nightmare if carried out manually.

According to Kris Seeneevassen, manager of WorkCover NSW's Data Mining Unit, the problem was that the data was not being viewed as a corporate asset.

"This meant we were not using it to its full potential," he says. "However, we are now using SAS solutions to develop analytical data mining approaches that leverage our stored data."

Ascertaining High-Risk Employers
There are four projects at development and operational stages within the Data Mining Unit. The first is an Employer Segmentation project that involves segmenting employer groups according to risk.

"We are analysing the type of controls particular employer groups have in place to identify which are the high-risk groups," Seeneevassen says. "This is an area we have not been efficient in and with the help of SAS, we expect to see vast improvements within the year."

WorkCover is also using SAS Enterprise Miner to develop an Employer Premium Payments model to identify employers who are not paying the correct premiums to the organisation.

"This is not to say employers are necessarily doing this intentionally," says Seeneevassen. "Many organisations are unaware of what they should be paying or they simply do not understand the process. However, underpayment of premiums affects everyone, and ultimately everyone ends up paying more."

The solution uncovered particular patterns in WorkCover's databases that enabled it to detect employers that were likely to be underpaying their premiums. A direct mail campaign was implemented that informed employers of their obligations and this was followed up with a series of targeted audits. Another group of identified employers was not sent a letter.

Those receiving the letter increased their premium payments by an average of AUD$5,000 (US$3,300). The overall result was that the data mining project led to a total saving to WorkCover of AUD$3 million (US$1.98 million). The feedback from the project also provided WorkCover with information it could use for future direct mail campaigns.

The third project is known as Employer Notification. "At any one time there is probably around 160,000 notifications from employers in our system," Seeneevassen says. "These contain information from employers about claims they have submitted.We are looking at SAS' Text Miner solution, to develop a process that will identify which employers are posing the highest risk."

Total Cost of Claims Calculated
The fourth project is a Case Estimation Model and Seeneevassen says this project will help WorkCover to determine estimates for what a claim will cost the organisation over its life. With this information, better use can be made of available resources.

Choosing SAS was an easy choice, according to Seeneevassen. "To carry out any serious number crunching, you have to use SAS solutions," he says. "It offers the most state-of-the-art technology around."

The organisation expects to see a range of benefits once its projects are fully operational, including the ability to better understand risk. Already the Employer Premium Payments model has yielded excellent results. Since applying this model, there has been a doubling in collected premiums for two consecutive years. Net revenue of audits has increased fivefold times.

WorkCover has created a database that incorporates policy information along with data fed to it from the nine independent workers compensation insurance companies in NSW. With the help of SAS, many instances of under-payment and non-compliance will be identified and suspect companies will be served with audit notices.

Seeneevassen says the value in a project is not always about cost savings. "We also want to prevent the severity of incidents," he explains. "We are identifying an employer list of the top 100 poorest performers and want to work with them to help overcome problems in their workplace by putting controls in place."

Back to Success Stories