If you are responsible for fighting fraud, waste and abuse (FWA) at a health plan and you are using rules-based approaches alone, you could do better.
The National Health Care Anti-Fraud Association (NHCAA) reports that between $68 and $226 billion is lost annually to FWA claims overpayments. What's more, the FBI even estimates that as much as 10 percent of healthcare spending can be linked to abuse alone, with the federal government losing more than $70 billion to improper Medicare and Medicaid payments in 2010.
In spite of how much coverage healthcare FWA has received in recent years, however, many health plans are making little headway toward solving the problem for two reasons. First, they are applying antiquated detection technologies to increasingly complicated FWA schemes. Second, they have been confined to use only their in-house data to identify claim overpayments, limiting their ability to identify and even predict as of yet unknown sources of abuse.
Analytical rules applied across multiple data variables, timeframes and claims repositories provide health plans with a more complete view of provider coding and billing practices. Unlike traditional clinical edits or fraud-based rules based on a single payer’s data, they can expand modeling capabilities further to identify provider-specific outliers.
As we speak, scientists and technologists are combining rich stores of data with sophisticated detection modules that enable health plans to view provider behaviors from as many different angles as possible.
For example, payers may desire to track individual provider patterns, such as the number of hours they are recording on a typical day. Using only single-payer data, it would be difficult, if not impossible, to locate egregious billing anomalies. However, the combined data from multiple payers, with analytics and rules applied, can identify a physician’s daily billed time.
With an analytics approach, health plans are alerted to outliers, such as a physician who recorded 26 hours for a single day of work across several patients’ health plans. Monitoring this statistic over time, payers will learn if the anomaly was an honest mistake or whether there is legitimate cause to investigate the issue as fraudulent or abusive behavior.
When payers add the human element via investigation experts and clinical professionals, they are achieving remarkable success in distinguishing between true overpayments as opposed to legitimately costly procedures delivered by a hospital or physician for a sicker population of patients. This not only prevents them from wasting resources on frivolous recovery efforts, but it can help to avoid creating acrimonious provider relationships.
Aiming for a multidimensional strategy
With advances in analytics and predictive modeling, many wonder what’s to become of rules-based detection technologies. Most experts conclude that to achieve the greatest payment integrity success they should employ a full-featured strategy that includes claims scoring predictive analytics with traditional rules-based solutions. Not only does this strategy provide the best chance for achieving ROI by preventing overpayments, but would-be offenders may be more likely to abstain from making improper claim submissions.
With a full-featured strategy that combines multiple technologies with rich clearinghouse data, health plans can achieve greater success in moving FWA detection much earlier in the process, before a claim is paid, preventing them from chasing dollars that can quickly evaporate.
Analytical models are designed to help payers detect claim or provider outliers, and can detect new FWA schemes. An analytics component could be configured to identify a provider whose patients cost significantly more than that of their peers in a similar specialty, for example, creating a platform for further investigation. Unlike rules-based solutions that are looking for specific anomalies, this detection model can identify outliers that were previously undetected.
Designed to recognize behavior patterns, predictive models are an even more refined tool for identifying FWA scenarios. Predictive modeling incorporates sophisticated algorithms that can foresee potential fraudulent or abusive claim activities. Models often include an artificial neural network that can sort through masses of complex interactions and variables, taking into account many different data points.
Powering these detection efforts is big data. With terabytes and even petabytes of information, health plans have a valuable resource to help them identify overpayments, yet they typically only have access to data generated by claims delivered by their provider organizations or internal processes. If they could access multi-payer clearinghouse data that includes rich details about not only medical encounters, but also dental visits, filled prescriptions, and member demographics, they would have an incredibly powerful tool for helping to reign in overpayments.
Bart Masters is the payment integrity analytics statistician for Emdeon, a provider of revenue and payment cycle management and clinical information exchange solutions in the U.S. healthcare system.
