Statistical Sampling and Extrapolation
Knoxville Academy of Medicine Bulletin – December 2007
By Diana L. Gustin and Jessica Kulkarni

The use of statistical sampling and extrapolation by CMS’ program safeguard contractors (PSCs) often results in allegations that a Medicare provider has been astronomically overpaid.  Statistical sampling and extrapolation of the sample results are used to establish an error rate.  This error rate is then applied to the “universe” of claims made by a provider.  (The “universe” is the period of time to which the sample will be applied).  In other words, Medicare might review $4,000 in claims and then calculate a $400,000 overpayment based upon extrapolation of the sample results to a specific time period.    

In 1990, the use of statistical sampling and extrapolation to calculate an overpayment was challenged in the case of Chaves County Home Health Services, Inc. et al. v. Sullivan, U.S. District of Columbia, U.S. District Court for the District of Columbia, (Feb. 12, 1990) Civil No. 86-2691.  The Court ruled that the Secretary of Health and Human Services had considerable discretion under the Medicare Act to use any reasonable means to recover overpayments, and given the fact that a case-by-case review of each individual claim was not feasible, the use of the statistical sampling method was reasonable.  The case was appealed to the U.S. Court of Appeals for the District of Columbia Circuit.  The Appeals Court upheld the decision, stating that because the plain language of the statute did not directly address the sampling issue, deference to the Secretary’s interpretation was required.

In light of the Chaves decision, physician and provider organizations began to lobby Congress to stop the use of the statistical sampling.  Excessive overpayments threatened the viability of small practices.  The American Medical Association and the American College of Physicians urged reform, advocating the elimination of extrapolation of alleged overpayment amounts to other non-audited claims in first time assessments of an alleged overpayment to a physician.

As a result, a limitation on the use of extrapolation was signed into law on December 8, 2003, as part of the Medicare Prescription Drug, Improvement and Modernization Act of 2003, Public Law 108-273.  (Title 42 U.S. Code Section 1395ddd.) Section 935 of this legislation limits the use of extrapolation, clearly prohibiting the practice unless the Secretary can show one of two requirements:  (A) sustained or high level of payment error; or (B) that documented education given to the provider has failed to correct the payment error.  If no education is given to the provider prior to the sample, extrapolation would only be permissible if the Secretary can show a sustained or high level of payment error.  The question remains as to whether the sample itself can be used as evidence of a high level of payment error, or if the procedure must encompass a two-step process.  The limitation on extrapolation will be most effective if the first sample is used to identify problems and offer education to providers to correct the deficiencies.  This would set the stage for a second sample where extrapolation would be permissible if the documented educational intervention failed to resolve the billing error(s).  Legal interpretation of the limitation on the extrapolation is pending in several cases working their way through the administrative appeal process.  The results could change the impact of the ­Chaves decision for all providers faced with post-payment audits based on statistical sampling and extrapolation.

Disclaimer: The information contained herein is strictly informational; it is not to be construed as legal advice.