Audits of healthcare providers are inevitable. Negative findings and allegations of overpayment are almost certain too. Because of voluminous records to be reviewed and scrutinized, auditors use extrapolation, reviewing only a snippet of all claims submitted. If any of the claims reviewed do not comply with applicable regulations and generally accepted practices – or contain clerical errors – such discrepancies are extrapolated across the total number of claims submitted. For example, if a PBM reviews a sample containing 100 claims and finds that 10 of them contain errors, it may seek recoupment of 10% from all claims submitted (not just these 100 claims reviewed but thousands of claims submitted by pharmacy during the audit time-frame).

Extrapolation has been widely criticized as an auditing technique because it does not accurately determine error rate. Many pharmacy advocacy groups – such as NCPA – consider extrapolation as a means to inappropriately enhance PBM’s audit revenues. As a result, a number of states have passed legislation to prohibit or limit extrapolation in healthcare audits.

States prohibiting extrapolation:





Kansas (unless Medicaid or Medicare audits)

Maryland (unless pharmacy agrees to extrapolation)

Minnesota (unless required by state or federal regulations)


New Mexico

South Carolina





PBMs’ use of extrapolation in California:

In auditing California pharmacies, PBMs rarely use extrapolation. It might be due to a fact that California has a provision in its Business & Professions code Sec. 4438 providing that if a PBM “uses extrapolation to calculate penalties or amounts to be recouped, the pharmacy may present evidence to validate orders for dangerous drugs or devices that are subject to invalidation due to extrapolation.” Therefore, it becomes imperative for a pharmacy being audited to determine if PBM uses any extrapolation methods, and if so, exactly what they are. If extrapolation has been used, pharmacy should present evidence that the claims assumed to be erroneous are actually clean claims. It might be necessary to present actual accounting of all claims paid to rebut the extrapolated amount.


Extrapolation in Medicaid and Medicare:

With respect to medical assistance programs like Medicaid and Medicare, it is well established that federal and state auditors may use an extrapolation method to calculate overpayments if the number of claims is “voluminous.” State regulations, however, may limit how and when extrapolation may be used in such audits. For example, New York provides that a provider may present counter-evidence to the extrapolation method and an expert testimony to invalidate the extrapolation method. For an opinion, describing extrapolation methods that may be used in Medicaid, see Bulmahn v. New York State Office of Medicaid Inspector General (N.Y. App. Div. 2013) 106 A.D.3d 1504.

In Bulmahn’s opinion, the state’s extrapolation method was invalidated because the state did not consider an amount underpaid to the pharmacy. The extrapolation method used in this case was invalidated because the administrative law judge gave no credit to the testimony of pharmacy’s expert that the failure to consider the underpayment resulted in an inaccurate determination of the amount the state had overpaid the Pharmacy.

As to extrapolation in Medicare, the Centers for Medicare & Medicaid Services has set forth in detail the method for extrapolating overpayments made by medical assistance programs in the Medicare Program Integrity Manual (MPIM). The MPIM specifically provides that “[i]n simple random or systematic sampling the total overpayment in the frame may be estimated by calculating the mean overpayment, net of underpayment, in the sample and multiplying it by the number of units in the frame.”


It becomes critical for a pharmacy to determine whether extrapolation is used to calculate overpayments. Next, research your state’s regulations on extrapolation and whether it is prohibited or limited in use. If Medicare or Medicaid claims are reviewed, determine if underpayment is included in the extrapolation methodology.