Monday, September 21, 2015

Predictive Modeling, a thing of the future….or is it here now?

Did you know that the Centers for Medicare and Medicaid Services (CMS) uses predictive analytics to analyze all Medicare fee-for-service (FFS) claims? This is done in part to detect claims that may be considered fraudulent.

The predictive analytics system utilizes algorithms and models that will examine claims submitted to Medicare so as to determine if the claims should be flagged for suspicious billing. Moreover, this is done in a “real time” environment. This process is being incorporated by CMS into the claims processing part of the equation.

This idea came about through the Section 4241 of the Small Business Jobs Act of 2010 (SBJS) as it mandated that CMS implement a predictive analytics system to analyze Medicare claims for areas at high risk of fraud. This is a similar process to one credit card companies are currently doing in the pre-payment arena.

As of June 30, 2011 CMS has been running all Medicare FFS claims through this predictive modeling system. This process builds profiles on providers, networks, billing patterns, and beneficiary utilization. The outcome of this process is that CMS can now create risk scores to determine the likelihood or fraud in the billing process. In addition, they can flag potentially fraudulent claims and/or billing patterns.

The risk scores will clearly and quickly identify any billing activity that is unusual and group them by providers, beneficiaries, and networks with the highest risk scores. This system, a very high-tech system, is not designed to replace the expertise or experienced analysts as the analysts still review prioritized cases and history along with identifying any innocuous billing and they will record this activity directly into the predictive analytics program and the payment is released as usual. If an analyst finds any activity that is fraudulent they will alert the CMS Center for Program Integrity, MACs, or Zone Program Integrity Contractors to enact targeted payment denials. When the alert involves egregious fraud, the billing privileges of the provider will be revoked.

The risk scores alone to not initiate any administrative action and the providers will not be able to appeal these risk scores. Currently, CMS is not denying any claims based solely on the alerts generated by this system. The predictive model is still being refined and they are developing more advanced algorithms that line up more closely to the complexities of medical treatment and billing.
With all of this said, CMS is dedicated to ensuring prompt payment to the providers as this is a statutory requirement, but in urgent circumstances CMS will leverage its authority to waive the prompt payment process and conduct a more detailed review on the provider.

Overall, by enacting a predictive modeling system, CMS is staying up with the times and doing its best to combat improper billing practices and to protect the Medicare Trust Fund.

Submitted by Kevin (Michael) Harrington, Full-Time Faculty at SJC

No comments: