Fraud and Abuse Detection Basics
Fraud and Abuse Detection
In the current environment where Fraud and Abuse is rampant throughout all types of industries and government programs, having a Policy for Fraud Detection with a clear and concise plan in place to detect improper transactions is critical. Daily we hear on the news or read in the papers how someone has either been caught defrauding agencies or companies out of millions of dollars. Almost everyone knows someone who has had their identity stolen.
It is still the Wild West in IT and software systems for Fraud Detection. Just when a system or a company thinks they know how to catch the bad actors, they change how or what they are doing to steal from us all.
How do we even start to look for the criminals and their bad actions? There are several proven techniques you can start with:
- Comparison of data. Check and compare different data sources for consistency of information. For example if a Business selling food reports to the state that they have sold 3 million dollars of groceries in the prior year but they have redeemed 6 million dollars in food stamps then we have a problem.
- Social media scrutiny. One state EBT Program is checking their local Craigslist ads for Formula being offered for cash. The state then cross references the contact information (cell number, email address etc.) with the state database of food stamp recipients. One current scheme is to purchase formula with food stamps and then resale it at an attractive discount for cash.
- Location and Distance. Does where a service being provided to a plan participant make sense? If the state is paying for child care so that a person can return to school or work programs is the child care center within a reasonable distance from where the family lives and works?
More advances techniques are available once a decision has been made to proactively detect bad actions. Predictive Analytics, Forecasting, Scoring Transactions and Unstructured Data Searches are some additional techniques. To accommodate these types of intensive detection schemas the underlying technologies are crucial for program success. Areas critical to address are:
- Speed. Huge Data Warehouses with multiple government departments can hold massive amounts of data. Speed has often become an issue when doing a query. The SAP HANA Platform is designed in such a way that the information needed can be in memory and responsive to inquiries.
- Integration. Comparisons of data from disparate systems can be problematic. Many government and commercial customers are run on a SAP Platform which would reduce the complexity of integration.
- Accuracy. In using technology to predict abusive behavior it is critical that the algorithms used are accurate and dependable. If the data is clean, responsive and well integrated the mathematics have a much higher likelihood of success.
In closing if I were an official whose program integrity was my responsibility I would want technologies that were proven and state of the art to assist with detection of fraud and abuse. I would also want to partner with a firm that has experience in catching bad actors. I would call my SAP Account Executive.