Medicare Compliance & Reimbursement

Compliance:

Part B MACs Now Using 'Predictive Modeling' of Claims to Catch Fraud

Fortunately, software alone won't be responsible for nabbing claim issues.

Don't ever accuse CMS of being unsophisticated, at least from a software standpoint.

The agency announced last week that Part B MACs have been using predictive analyses to scan your claims and potentially detect fraud, according to MLN Matters article SE1133. "As of June 30, 2011, CMS is streaming all Medicare FFS claims through its predictive modeling technology," the article notes. "As each claim streams through the predictive modeling system, the system builds profiles of providers, networks, billing patterns, and beneficiary utilization. These profiles enable CMS to create risk scores to estimate the likelihood of fraud and flag potentially fraudulent claims and billing patterns."

When the system alerts the MAC to unusual billing activity, the carrier will thoroughly review the claim before releasing payment to the practice. However, CMS notes, MACs will continue to use human analysts in concert with the software programs.

How it will work: "It's definitely a bit more complicated than before and what it means is that, using predictive analytical algorithms, CMS will be able to assign a score to each claim, real-time, that assesses the probability that that claim (or claim line) may be fraudulent," explains Frank Cohen, MPA, MBB, principal and senior analyst with The Frank Cohen Group, LLC.

The computer system prioritizes claims and providers that generate the most alerts and highest risk scores. Then, analysts review these prioritized cases by checking claims histories, conducting interviews, and performing site visits when necessary, the article notes.

In cases where the analyst finds that the practice is billing appropriately, the payment is then released to the provider. If, however, the analyst finds fraud evidence, the case is referred for further action, payment is denied, and in some cases, Medicare billing privileges are revoked.

Automatic Denials Aren't Happening

Although the system sounds very automated, the extra step of human review means that the software won't automatically deny claims solely based on system responses. Instead, the computer analysis will simply flag claims for further review.

CMS acknowledges that the system is not yet perfect, noting in the article that the agency is "working closely with clinical experts across the country and of every provider specialty" to refine the algorithms that will best reflect the complexities of medical treatment and billing.

The bottom line: Although the changes may sound like something out of a sci-fi movie, you need not worry about willy-nilly denials. As CMS works out tweaks in the system and perfects the analytical algorithms used, you'll probably see no difference in the way your claims are paid.

To read more about the predictive modeling system, visit www.cms.gov/MLNMattersArticles/Downloads/SE1133.pdf.