FDA Requirements on Reporting and Interpreting OOS Results

GMP News No. 879: FDA Requirements on Reporting and Interpreting OOS Results

GMP News
31 January 2007

FDA Requirements on Reporting and Interpreting OOS Results

In our

GMP News of 12 October 2006  we informed you about FDA's new final OOS Guidance.

The section on "Reporting Testing Results" of this guideline deals with reporting and interpreting test results. Here, two important aspects are averaging and outlier tests.

1. Averaging

With regard to averaging, the FDA distinguishes between appropriate and inappropriate uses.

Averaging can by all means be an appropriate method. One of the examples given in the text is optical rotation, another one refers to HPLC. Averaging a number of consecutive, replicate injections (usually 2-3) is possible. In this case, the average is the reportable result and is also regarded as the final analytical result. However, this approach has to be laid down expressly in the test method. Thus the term "reportable result", which has come up more and more often in discussions on OOS in recent years, can now also be found in the FDA guidance.

What is important in averaging is the need to specify limits for the variability of the replicates in the test method.

Areas where the FDA considers averaging to be inappropriate are e.g. determining the uniformity of blends/mixtures or the uniformity of dosage form content.

2. Outlier Tests

Very rarely, an analytical test result can be quite different from the other results obtained in a series of measurements using a validated method. In this case an outlier test can be conducted. However, the guidance defines several prerequisites for this:

  • the use of an outlier test has to be laid down in advance in an SOP on data interpretation
  • the SOP has to specify which outlier tests are applied
  • the relevant parameters of the outlier tests have to be defined beforehand
  • the minimum number of results necessary to obtain a statistically significant assessment has to be laid down

In case of validated chemical tests with relatively low variance, an outlier test can only be a statistical analysis of those data obtained from testing itself and from retesting. An outlier test can never identify the cause of an extreme observation. For this reason it should not be used to invalidate a suspect result.


Get to know the most important statements and requirements of this new FDA Guidance in detail - during the GMP Education Course

Out-of-Specification Results in Barcelona, Spain, on 14-15 March 2007.

Frederick W. Blumenschein, FDA, will introduce you to this new FDA Guidance in a video conference directly from CDER's Office of Compliance and answer your questions in a Q&A session. Do not miss this opportunity to get first-hand information.

Dr Günter Brendelberger
On behalf of ECA

The complete Guidance for Industry "Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production" can be found here:

Go back

GMP Conferences by Topics