Social Security to step up fraud detection with predictive analytics
- By Stephanie Kanowitz
- Apr 22, 2014
Two major fraud incidents in the past year have spurred the Social Security Administration to step up its fraud detection and prevention efforts through the use of better analytics. The result is a fraud prevention unit dedicated to building data analytics to help SSA workers detect and prevent scams.
“SSA will apply analytics tools that can determine common characteristics and meaningful patterns of fraud based on data from past allegations and known cases of fraud,” said BJ Barrett, an SSA spokesman. “We will apply these tools when reviewing business applications or existing data on beneficiaries for potential fraud or other suspicious behavior.”
Announced March 31 by Carolyn W. Colvin, acting commissioner of SSA, the predictive tools will increase the agency’s ability to identify suspicious patterns of activity in disability claims and to prevent fraudulent applications from being processed, Barrett said, adding that specialized units such as a fraud prevention unit in New York will assist the project by conducting a detailed analysis of cases identified by the tools.
“With these diagnostic tools, we anticipate increasing our ability to identify questionable patterns of activity in disability claims and prevent fraudulent applications from being processed,” Colvin told the House Ways and Means Committee’s Subcommittee on Social Security during a Feb. 26 hearing. “During the remainder of [fiscal year] 2014, we will develop and begin testing some of these tools.”
“SSA has a long, successful history of developing online applications, electronic tools, and predictive models to efficiently process benefits claims, enhance decisional quality, and target limited resources toward those program integrity reviews most likely to return savings to the taxpayer,” Colvin added.
The New York team is made up of 20 expert disability examiners who are currently involved in the re-review of disability medical decisions resulting from recent indictments there and in Puerto Rico.
Since Jan. 7, officials have made 106 arrests in a disability fraud scheme based in New York involving retired members of the city’s police and fire departments. “The total amount of fraudulent benefits allegedly taken from SSA through this scheme by the 130 beneficiaries and four facilitators identified now exceeds $30 million,” according to SSA’s Office of the Inspector General.
Last August, more than 70 people were arrested in connection with a Puerto Rican-based scheme in which the group received payments they didn’t qualify for. Those involved face more than $3 million in forfeitures, the Washington Times reported.
“SSA’s current experience with disability fraud through our Cooperative Disability Investigation Units indicates that every $1 spent on disability fraud prevention will save our agency $16,” Barrett said. “We’re optimistic that we will experience a similar ROI for our data analytics program.”
Several states have pursued analytics-based fraud detection. Massachusetts’ MassHealth, uses predictive modeling to provide real-time risk assessments to combat scams in Medicaid payments. The system has helped MassHealth recover $2 million in improper payments, according to Joan Senatore, director of the Massachusetts Medicaid Fraud Unit.
Similarly, three states – Georgia, Indiana and Louisiana – use LexisNexis Solutions’ algorithms to screen tax returns against a database of public information to combat tax fraud, NBC News reported in March.
Despite the two recent scams at SSA, preventive measures already work well, Colvin told the subcommittee. “While any level of fraud is unacceptable, the low level of disability fraud in our programs speaks to our efforts; the best available evidence shows that the level of actual disability fraud is below 1 percent,” she said.