12:30 PM
Fraud Detection and Prevention Starts With Good Data
Different healthcare carriers are implementing different fraud detection and prevention technologies according to their business needs. These include homegrown applications and various types of rules-based applications that are available from vendors. There are other technologies available now using neural nets or predictive modeling, which is the state-of-the-art technology for healthcare fraud. With the rules-based technology, if you know you have a problem or issue, you can go in and build an algorithm to see the impact the issue has had on your business.
The predictive modeling technology is data-driven, so it's built on your data. You build the model for it, and it then gives you the information that you can customize any way you want. It actually goes in and finds problems you don't even know about. Highmark has implemented predictive modeling from Fair Isaac (Minneapolis), and we have a medical model and a pharmacy model.
We started an internal project about three and a half years ago. We took a look at what we were doing and found we needed to work better, faster. We looked at our processes and found the things that were bogging us down in our case development were getting access to our data, doing our data extracts, running those reports and doing the analysis. That would take hours, days and, in some cases, weeks. We teamed with our Informatics group, which built an analytic tool for us called FIRST. With this application on every SIU member's workstation, we are running reports and doing data extracts in minutes and even seconds, where it would have taken hours or days before.