MITRE Corporation names analytics company a finalist in benefits program eligibility challenge
With $141 billion lost to improper payments each year, the US government needs innovative ways to ensure the right benefits get to the right people. When the MITRE Corporation called for creative solutions to the problem, analytics leader SAS answered with an analytical approach for verifying the eligibility of benefit applicants. SAS was one of four finalists selected from 29 industry and university entrants in the MITRE Challenge on Strengthening Eligibility Verification for Federal Benefits Programs.
“Approximately 25 percent of improper payments occur because agencies do not verify with certainty that the individuals paid were eligible,” said Gordon Milbourn, MITRE’s Payment Integrity Portfolio Leader. “This year’s MITRE Challenge was designed to discover innovative, cost-effective solutions that government agencies can use to improve the verification of benefit eligibility in federal programs and thereby save money.”
According to MITRE, improper payments due to fraud or error have increased 33 percent since 2013. The 2018 President’s Management Agenda listed “getting payments right” as one of its 14 cross-agency priority goals.
The SAS® solution helps recipients get back to work while ensuring benefits get to the people who truly need them. By providing administrators a holistic view of a person’s life over time, the recipient can receive the right assistance, from the appropriate agencies, at the best time, instead of an assortment of disparate benefits.
A real problem for a fictional agency
The challenge involved a fictional new benefits program, Program G, which helps people get back to work.
Program G provides supplemental, short-term benefits to working families during periods where benefits from other programs are inadequate for basic living needs. It also provides cash and vouchers to cover expenses for clothing, relocation, transportation, and other items and services required during a work search.
The presiding agency wants to fortify Program G’s enrollment and eligibility determination process against lost funds due to fraud, applicant error and agency error. At the same time, the agency wants this process to be efficient, effective and user-friendly, and to protect privacy. Instead of identifying fraud and waste after the fact, the agency seeks to deter fraud and prevent errors through improved application processes and stronger verification techniques.
Read More: What’s the Next Game AI Will Solve?
Eligibility must be verified not only at the time of application but monitored monthly in case “life changes” that affect eligibility occur, such as when the household income level rises above the eligibility threshold. Proposed solutions were required to consider and conform to existing privacy regulations, and to be realistic and replicable.
“What we proposed could be implemented at a benefits agency today,” said John Stultz, Government Fraud Solutions Specialist for the Fraud and Security Intelligence practice at SAS. “We pulled from years of working with public and private sector customers to reduce improper payments and prevent losses.”
Putting a legacy of fraud fighting to work
SAS is a leader in fraud analytics, and has worked with government agencies, banks and insurance companies for years to combat fraud, waste and abuse.
Instead of verifying eligibility per benefit program, SAS proposed a solution that takes a “recipient-centric” approach. The proposed open, interoperable analytics platform strategy provides a benefit program office the ability to access data from virtually any available source, hardware platform, operating system or data format, instill governance, and operationalize data environments that can be used for risk modeling and automated alerting based on life-changing events.
By adding contextual information from outside data sources, in accordance with privacy regulations, Program G improves data reliability and timeliness. The agency can better understand when changes to a recipient’s life may have altered their needs.
Artificial intelligence and machine learning can detect and prioritize risk associated with a life changing events like losing a job, leaving a household or being incarcerated. Alerts could identify high-priority individuals who, given a life-changing event, should receive assistance such as a rent voucher or transportation services support.
The solution also supports an infrastructure for collaborating with other federal and state agencies, commercial firms and community partners so the agency can objectively and independently verify whether applicants meet all criteria, before benefits are paid.
Dashboards can empower a benefits program to quantify performance and understand risks to positive outcomes. Data quality techniques strengthen the accuracy of verification and improves fraud detection. For example, entity resolution gives program managers a consistent single view of a recipient. They can know that the William F. Jones in one database, the Billy Jones in another and the Bill F. Jones in a third are the same person, based on additional data about the person.
“Ultimately, the solution would make Program G’s benefit enrollment and eligibility determination process less vulnerable to lost funds due to fraud, applicant error and agency error,” said Stultz. “Using proven software and techniques, the SAS solution can accurately identify risk, reduce false positives and ensure that the vast majority of Program G’s valid transactions continue without delay.”