Using Data Mining as a Component of Audit Defense
During the last several years, the Centers for Medicare & Medicaid Services (CMS) have been focused on preventing and detecting Medicare fraud and abuse more than ever.
With the installment of the permanent Recovery Audit Contractor (RAC) program and the transition to Zone Program Integrity Contractors (ZPICs), methods of preventing fraud and abuse have become increasingly more organized and targeted. Several of these methods revolve around technical innovations used as a way to increase efficiency and accuracy when dealing with potentially improper claims submitted to Medicare. Of these technological innovations, one of the most prevalent is the use of data mining.
CMS has compiled an extensive computerized database of claims and services billed to Medicare by providers. RACs, ZPICs and other audit contractors use this database to analyze coding and billing practices of Medicare providers across the country. Contractors in doing so identify specific data (i.e. the most frequently billed codes, practice locations, etc.) and use it to create a model physician profile for a particular provider type.
These profiles then are used as reference points to compare against individual providers' coding and billing practices. This information allows contractors to identify unusual billing patterns as well as identify outliers by comparing providers' practices to those of other providers within the same specialty area. Upon identifying any billing abnormalities or outliers, contractors may take action, such as the initiation of an audit of a provider's services, in an effort to retrieve any overpayments issued by Medicare. For example, RACs conduct automated reviews using data mining techniques to identify payment errors (i.e. duplicate claims for the same service) whereby upon discovery, they send an overpayment demand letter to the provider. However, RACs also conduct complex reviews when they determine that a payment error likely has occurred, leading to requests for medical records and further review of claims.
The audit contractors' use of data mining is a key element of the potential audit risk facing many Medicare providers. If and when an audit occurs, the provider often is faced with significant overpayment return demands and the prospect of weathering the five-step Medicare appeals process in an effort to defend against claim denials. This process can be ominous and time-consuming. As such, it is imperative that providers proactively prepare to defend claims on a clinical basis as well as develop legal challenges as part of an audit defense strategy.
Much like audit contractors, providers also can take advantage of a form of internal data mining. Specifically, as providers challenge denials through the Medicare appeals process, they can collect key data elements about the specific claims at issue and the outcome of appeals at the various stages of the appeals process. This data from previously audited claims then can be applied in the defense of current claim appeals.
This approach can be particularly effective for institutional providers, including hospitals and health systems, which process a significant volume of audit claim appeals. For instance, when appealing claim denials, providers can track specific elements of RAC-approved claims and successful appeal results at the redetermination, reconsideration and ALJ stages, then compare those findings with their present appeal case(s). This comparison may provide additional support and a persuasive analogy for a current appeal in the case that payment was approved in an identical or substantially similar past case, giving the provider no reason to believe that its current appeal would not be approved. Such data mining also may serve to bolster a provider's waiver of liability defense, which allows payment when a provider did not know nor reasonably could have been expected to know that payment would not have been made for services at issue. While appeal decisions in other cases are not binding when it comes to ALJs, they may provide persuasive support for Medicare coverage based on both the medical necessity of the services and the waiver of liability defense. Providers are advised to redact any protected health information (PHI) on ALJ decisions or forwarded information tied to similar appeals.
An example of providers using data mining tactics can be highlighted by RAC denials of inpatient hospital admissions, which often is a dicey proposition because the distinction between "inpatient" and "outpatient" is often vague and difficult to ascertain. Hospitals can look at the characteristics of past RAC approvals and successful appeals for individual patients, then draw analogies to claim details connected to appeals in progress. The provider should isolate key elements - including length of stay, type of diagnosis (primary or secondary), type of services, common risk factors, etc. - and then use data mining to compare this information to claims approved in the past.
Providers are advised to make these data mining considerations early in the appeals process and track favorable appeal decisions at each level of appeal. Due to the requirement of early presentation of evidence at the reconsideration stage of appeal, this evidence should be submitted prior to the issuance of a reconsideration decision.
The burden of challenging audit denials through the Medicare appeals process can place a significant strain on providers, making it vitally important for them to put the necessary tools in place to defeat claim denials. The use of proactive data mining can be an extremely valuable defense tool, aiding providers in successfully appealing RAC, ZPIC, and other audits - and ultimately getting their claims approved.
- Next >>