Samples need to be taken from the production line at intervals that will allow a sufficient statistical6 evaluation of the results. Each of the CPPs are tested, thus samples for each of the CQAs need to be taken several times during the run. Under normal operating conditions samples are taken at the beginning, middle, and end of each run. However, during validation executions samples need to be taken more frequently (eg, every hour).
Because of the extra samples needed during PPQ execution the QC lab or the contract lab to which the samples are to be sent for analysis must be prepared for the extra load. This includes not just the testing of the samples but also storage of the samples as necessary.
Sampling Plans7
There is no single approach to determining the appropriate sampling plan; however, it must be scientifically accurate to provide valid results of the testing. Plan design should include a statistician or at least someone well versed in statistical methods. Designing a sampling plan that has the appropriate resolution to describe the process variability is an important part of building confidence in the process. Whichever approach is selected it must have a clearly defined rationale behind it. There is no right or wrong answer, but whatever sampling plan is developed must be defensible based upon the level of resolution necessary to see variation in the process as well as being statistically accurate. Thus, a large number of samples are taken at short intervals during production.
Recording the Results
As with any protocol execution the results need to be recorded using Good Documentation Practices (GDP). Recording of numbers is very important and should follow corporate SOPs (Standard Operating Procedures) on rounding, truncation, and use of significant figures. All original data must be retained. Copying over of results so that the document looks neat is often a cause for transcription errors and more.
The Validation Report
When the execution is complete, that is the specified number of runs have been made and all of the samples collected and tested, the data entered, lab results are complete and confirmed as correct (ie, all calculations have been verified) the validation final report can be prepared. In addition, all deviations (see chapter: Dealing With Deviations) have been logged and the appropriate investigation executed. Remember, the protocol cannot be closed while any deviation is open. The report should follow the company SOP on validation reports and should contain the following:
