Data mining helps uncover MS-DRG mistakes

The increasing complexity of CMS data mining tools to monitor compliance and overpayments should encourage hospitals to develop their own data mining initiatives, according to an article in the Sept. 29, 2008 issue of Report on Medicare Compliance.

The article describes three factors that are increasing the need for hospitals to create their own data mining efforts:

1. January 2009 marks the end of CMS’ Program to Evaluate Payment Patterns Electronic Report (PEPPER), which provided free data on 13 risk areas that could be used by hospitals to monitor how their coding compares to peers.
2. The mandatory assignment of Present on Admission (POA) indicators is increasing the complexity of the coding process, and is providing a data element that CMS can use to more closely monitor a hospital’s coding and clinical practices.
3. CMS has increased the complexity of its data analysis capabilities, now only with the Recovery Audit Contractor (RAC) program, but also with the development of the One Program Integrity System Integrator (One PI) data-analysis tools that are used for fraud and abuse detection.

Combined, these elements are providing CMS with a substantial advantage when it comes to data analysis. In response, the article recommends that hospitals create data mining techniques to analyze their own data to identify potential compliance errors. The article provides specific examples of how hospitals can use their MS-DRG claims data to monitor areas that are likely to attract scrutiny, such as pneumonia discharges where it is vital — although tricky — for clinical documentation to support the patient’s length of stay.

Another example within the article highlights the effective use of "POA mini-audits" to identify coding errors and incomplete clinical documentation practices.

Access the full article here.