During the demonstration project, RACs chose not to apply the process of extrapolation, and while they have yet to use it during the permanent program, the fact remains that it is still a potential threat.
Understanding how the extrapolation process works will prove to be valuable to providers in the end.
Extrapolation is the use of sampling a specific set of statistics to project an overall error rate to determine overpayment amounts made to a facility. For RACs to use extrapolation, there must be a determination of sustained or high level of payment error, or documentation that past educational interventions have failed to correct the payment error, according to the Program Integrity Manual.
A RAC must then use a trained statistician, or someone with statistical expertise, to choose a valid random sample in the same manner that CMS carriers are instructed to do. The new issue review board must then approve the methodology for the extrapolation process to begin.
Unfortunately, the use of extrapolation can result in an inflated level of overpayment determinations. For example, if the RAC is reviewing how a facility billed drug units for Neulasta, and has determined that the facility has an 80% error rate on reviewed claims, it would be considered a high error rate, according to Debbie Mackaman, RHIA, CHCO, regulatory specialist for HCPro, Inc.
"Based on that error rate, the RAC could develop a statistically sound sampling and project how many times the error 'probably occurred' on all paid claims from October 1, 2007 forward, and based the total overpayment amount on the sample rather than the actual claims," she says.
If the results of an extrapolation are not appealed by a facility, the RAC is entitled to keep its contingency fee based on the sampling. Because RACs have a limit on the maximum number of medical records that can be requested per 45-day period for complex reviews, providers need to be aware of RACs potentially using extrapolation to determine overpayment amounts.