Michael Spake (michael.spake@mylrh.org) is Senior Vice President of Compliance and Integrity at Lakeland Regional Health in Lakeland, FL.

Detecting healthcare fraud and abuse is challenging even to a well-seasoned compliance auditor. The nature of healthcare services, unlike other service industries, is complex due to the uniqueness of the individual patient and their medical condition. In addition, variability among physician practices in the same specialty across different communities is significant (e.g., there are marked variations in the frequency of major procedures across different regions of the country and end-of-life hospital days).^[1] This variation, coupled with the complex delivery of medical services, makes potential fraud and abuse practices difficult to identify, because variations in billing practices do not automatically indicate billing abuses or inappropriate medical use. As a result, there are plenty of opportunities for a provider to adjust medical use and/or billing practices to fraudulently enhance revenue without detection. Because of the opportunities presented, compliance teams must evolve new tools to identify fraud.

In the past, many instances of potential fraud were only detected through laborious chart audits that sometimes could only detect fraud months after the questionable service, bill, or transaction had been rendered and/or paid. As a result, fraud or even unintentional overpayments could go on for many months before being detected, even in some of the most rigorous auditing systems. However, today the use of electronic health records blended with statistical analytics can quickly identify patterns that may lead to the detection of fraud and abuse or erroneous billing patterns faster. One particular statistical tool is a simple distribution analysis that can assist in the examination of evaluation and management (E/M) coding patterns, Medicare severity-diagnosis related group (MS-DRG) pairings, and even potential narcotic diversion and overuse and physician compensation. A simple distribution analysis can visually demonstrate to a compliance auditor when any of these practices do not align with a bell curve or comparable benchmark and instead cluster around questionable frequency levels.