Industry Perspective: Accenture executives detail how data-driven decision-making helps human services agencies improve operations and control costs.
Editor’s Note: Christopher Gray is director of Accenture Global Intelligent Processing and Compliance. Julie Booth is managing director of Accenture North American Human Services.
Struggling with increasing demand for services amid widespread economic constraint, human services organizations face a major challenge of finding ways to minimize costs while improving services and ensuring accurate benefit distribution.
By using analytics, forward-thinking human services organizations are rising above this challenge. They are preventing, detecting and mitigating transactions where there is error, fraud or abuse. And they are using information gleaned from analytics to significantly reduce operating costs and drive business results.
These approaches represent a dramatic break from the status quo. Human services agencies have traditionally used analytics to identify and correct noncompliance only after a transaction had been completed — for example, using analytics to identify cases for investigation. In this “pay, then chase” model, organizations spend already scarce resources to pursue fraudulent or erroneous payments that should never have been issued.
Now, human services solutions that encompass innovations in process synchronization and predictive analytics allow agencies to work more proactively than ever. Rather than detecting and correcting noncompliance after the fact, they are staying steps ahead with analytical insight.
Human services organizations around the world are already seeing the benefits that analytics can bring:
A large social security agency in Europe uses new analytic approaches to tackle fraud and error within the benefits and contributions systems. This social security agency, as part of a wider strategy to combat loss caused by fraud and error, has implemented a new antifraud model and surrounded it with a new analytics system to identify potential fraud and loss. The agency estimated prevention of €42 million of fraud during a one-year pilot.
The U.S. Social Security Administration (SSA) has begun to use text analytics in processing disability benefits applications. To more effectively manage the applications process, which traditionally suffered from long delays, the SSA produced a scoring model against which all applications are now automatically analyzed (using text analytics). This approach has helped greatly reduce application time for these cases, as well as reducing the staff time and costs involved in the process. The system has been particularly successful in managing the renewals cycle, automating low-risk renewals and removing the need for reassessment from doctors for low-risk individuals based on their applications and histories. This approach boosts capacity while helping control costs.
A large Canadian social services organization implemented predictive analytics to develop the risk model to direct the appropriate treatment of cases. Additionally, to address resource capacity limitations, this organization used analytics to target activities that would yield the greatest results. Through the use of predictive models, the organization realized a boost of 400 percent in its return on investigative activity, ultimately improving the overall integrity of the program and reducing erroneous payments.
A large U.S. city used analytics to achieve a 55 percent improvement in identification of business fraud (new, emerging and hidden). Another local government social services agency in North America implemented a fraud prevention program that has yielded annual savings of 4 percent on a $2.5 billion income assistance program.
The increased availability of data today provides a great opportunity to use analytics. As such, human services IT systems are collecting greater amounts of electronic data with interfaces to external information sources. By capitalizing on available public- and private-sector data and advances in technology, human services organizations can better understand the characteristics and motivations of different client types and quickly tailor their responses.
The key is to embed predictive analytics at the heart of business operations, so agencies can assess risk in real time during core transaction processing. By predicting, identifying and preventing noncompliance before the transaction is complete, organizations achieve direct savings. Moreover, audit and investigation resources are free to focus on complex and needy cases, and collectors are free to pursue the highest value cases.
The right application of analytics-driven compliance can help human services organizations:
Spot problems. By applying a predictive analytics lens across compliance activities, organizations identify high-risk and potentially erroneous or fraudulent claims at speed, and can focus on those clients and interactions that require the most attention.
Focus efforts. Continuous measurement, monitoring and review of client behavior patterns and the predictive models help determine the right strategies and resource allocation, matching the organization’s response to the needs of its clients.
Cut costs. Organizations can generate savings quickly through technology accelerators and real-time analytic data management. Leading human services agencies using predictive analytics are realizing more than 200 percent lift in their return on investment.
Boost business results. Predictive analytics can identify changing client behaviors and needs, giving organizations key information to enhance value and improve business results. Human services organizations using analytics are experiencing improved program integrity, leading to program savings of approximately 4 percent from the prevention of overpayments and erroneous payments.
Enhance quality of service. Predictive analytics enable organizations to proactively differentiate their response and service to clients.
Assess and adapt. The right compliance framework can help organizations assess the quality and fitness of any existing models, rules and analytic data, and augment current models to address other areas of concern, such as case selection criteria, identity theft and fraud.