Test Method Validation

Like all processes, the Quality Control testing must also be validated; Test Method Validation. The steps are similar to Process Validation, IQ, OQ and PQ. For tests like visual inspection, only the PQ is applicable. A difference in the validation method for attribute and variance testing is made.

Validation Process

The Installation Qualification and the Operational qualification of test equipment are similar to process validation. It is essential that the equipment is incorporated into the quality system, that software and firmware are validated, and that the equipment performs the analytical test as expected (on standard samples). During operational qualification, the correct settings of the machine for a specific type of sample are determined.

You want the QC test results to say something about the quality of the product and the capabilities of your production process. So you need to know the process, equipment, and operator-related variance (see drawing).

The number of samples and the statistical calculation depends on the type of test and its effect on the sample (see drawing).

PQ – Variance GR&R

Now the number of samples and the statistical method is defined, the samples are tested, and the results are entered into statistical software. Different evaluation methods are possible; the most common is the ANOVA method. The contribution of other factors to the variance (contribution of repeatability, reproducibility, and Gauge R&R) is evaluated.

With variance Gauge R&R testing, the results are compared to pre-determined acceptance criteria. The tests pass the acceptance criteria, and the PQ is finished. If the criteria do not pass the acceptance criteria, an investigation into why it does not pass is performed, and corrective actions are implemented. The GR&R testing is then repeated until the acceptance criteria are passed.

PQ – Attribute GR&R

Attribute testing results in a PASS or FAIL (think of visual inspection). It is essential that these attributes are not arbitrary and all operators score similarly. The Attribute Gauge R&R is done by having multiple operators score multiple times a set of samples following a work. This set of samples must contain a similar number of PASS as FAIL. The operators score the samples (pass-fail) compared to a master list. Using statistical techniques, the effectiveness of the attribute test is determined. An effective measurement system should demonstrate low variability within operators (repeatability), low variability between operators (reproducibility), and high agreement between operators or systems (master list).

When pre-determined acceptance criteria are not met, an analysis of the data will tell if an individual operator requires retraining or the work instruction requires review. Re-testing is done until the acceptance criteria are met.