A Primer on Medicare Statistical Extrapolations
Medicare audits often include a statistical extrapolation to estimate the full extent of an alleged overpayment. Medicare contractors are authorized to review the merits of only a small “sample” of submitted claims and extrapolate the results of that review to a large “universe” of claims to estimate the overpayment amount. This practice can lead contractors to allege overpayments of hundreds of thousands or millions of dollars despite only reviewing and denying a small handful of claims. Medicare contractors generally have broad authority to use a wide array of statistical methods when extrapolating the overpayment amount, which can lead to grossly overestimated overpayment determinations.
When conducting the statistical sampling and extrapolation, the contractor will select the period for review and establish the universe and sample frame. The sample frame is the large group of claims from which the sample is randomly selected, and the universe is the group of claims over which the results of the review are projected. The sample frame and universe may or may not be identical. The universe and sample frame may be defined by specific codes, dates of services, beneficiaries, or some combination thereof. From here, the contractor will select a random sample from the sample frame, review the claims within the designated sample, and extrapolate the results of the review of the sample to all claims in the established universe.
A statistical extrapolation is subject to appeal, similar to any Medicare overpayment determination. However, there are several issues unique to appealing a statistically extrapolated overpayment. First, it adds increased importance to appealing the denial of each claim in the sample. While an individual claim may not represent significant monetary value on its own, it may represent tens of thousands of dollars after it has been statistically projected over a large universe. Second, there are special procedural rules for appealing an extrapolation. For example, providers are generally prevented from arguing before an Administrative Law Judge (ALJ) that the extrapolation was flawed unless they included specific language in their request for ALJ review.
Further, there are a number of technical arguments that can be made regarding how the contractor performed the extrapolation. One common statistical issue involves the use of an unstratified sample frame, where the contractor will lump many different claims into one sample frame without accounting for the differences in value or other aspects between the claims. On the other hand, using a stratified random sampling would generally ensure that the sample frame accurately reflects the entire universe of claims, so failing to stratify the sample frame will likely lead to inaccurate, and often exaggerated, overpayment values. Similarly, contractors will often fail to normally distribute the claims found in the sample frame, which will further skew the extrapolation.
While Medicare does not require its contractors to use statistical best practices and gives its contractors significant leeway to make errors in their methodology, the cumulative effect of many otherwise allowable errors may be enough to render a statistical extrapolation invalid. Therefore, appealing an extrapolation often requires the help of an expert witness to review the contractor’s work and identify any mistakes. Other payors, such as commercial insurers and Medicaid, may also utilize some form of statistical extrapolation in their audits and these may be subject to different rules.
For over 35 years, Wachler & Associates has represented healthcare providers and suppliers nationwide in a variety of health law matters. If you or your healthcare entity has any questions pertaining to a statistically extrapolated overpayment or healthcare compliance, please contact an experienced healthcare attorney at 248-544-0888 or firstname.lastname@example.org.