7.4 Audit Samples

At times, audit samples are collected, preferably from the same container as the original sample collected by the importer, sampling service, or private laboratory. They are tested by an ORA laboratory to verify the analytical results that purport the product to be in compliance with the FD&C Act. The package reviewer may request collection of an audit sample. Because resources are limited, not every request will necessarily result in a collection. The factors below are suggested criteria for the recommendation for collection of audit samples. Collection should be made when substantial and significant incidences are observed in the following analytical areas.

Private Laboratory Observations 

• If analytical problems are found during an assessment visit of a private laboratory made by any district. Samples should be related to the types of analytical problems observed.
• If this is the first submission of a private laboratory or analyst for a particular type of analysis.
• In general, for at least 10% of the work performed by a private laboratory, to be increased if findings warrant. This includes both samples collected due to factors on this list and random samples.

Analytical Packages

• For analytical packages that are unacceptable upon initial review. However, no audit sample should be collected until corrections have been made so that the package is found to be acceptable. For example, if a package is incomplete, no audit samples should be collected until the missing information has been supplied and reviewed.
• For types of analyses and commodities for which reports with errors, technical problems, or missing data have been submitted during the past year.
• For reports with abnormal or unexpected results or data—such as reports indicating no growth of microorganisms or negative filth or decomposition results. These can be in a product, shipper, or country combination that often has these elements, or exceptionally "clean" chromatograms for products that usually exhibit product peaks.
• For multiple borderline analytical results of a given type, or other potentially suspect findings associated with borderline results. For example:
  • 0.9 ppm methyl mercury in swordfish;
  • 11.5% chromium in surgical instruments;
  • 1 of 18 cans of tuna for decomposition;
  • 3 of 18 subsamples of shrimp for decomposition;
  • large numbers of class 2 shrimp;
  • excessive filth in one subsample but none in other subsamples;
  • pesticide residues close to the tolerance levels; or
  • aflatoxin close to the tolerance levels.

Detention without Physical Examination (DWPE)/MOU's

• At least one audit sample should be collected and analyzed upon a request for taking a product, shipper, or country off DWPE.
• As called for by MOUs or compliance programs and assignments.

Sampling

• When there are past findings of erroneous, incomplete, or suspect sampling procedures by the private laboratory, importer, or third-party sampler.
• For samples collected by the importer rather than by the private laboratory or an independent sampler.

Evaluation of Audit Sample Result. If the ORA laboratory finds an audit sample violative for an entry that the private laboratory had previously found to be in compliance, it can not necessarily be concluded that the discrepancy is due to a problem on the part of the private laboratory. Possibly the analyte was not homogeneously distributed, or the sampling was not done correctly (for which the private laboratory would not be responsible unless it had performed the sample collection). Each case is evaluated individually. If it can not unambiguously be determined where the problem lies, then additional audit samples or an on-site assessment visit may be warranted.