Method validation and verification in testing laboratories

Analysts and technical staff who bring methods into routine use, quality managers who sign the resulting records, and technical managers accountable for the scope a laboratory claims.

It assumes you work in a laboratory and have run methods. It does not assume any statistics background — the statistics are built from the ground up around what each figure is actually for.

The syllabus below is the standard shape of the course. It is adjusted to the matrices, methods and instruments you actually run, which is agreed before delivery rather than discovered on the day.

• The distinction

  When a standard method needs verification and when a method needs full validation

• Parameters

  Selectivity, linearity, range, accuracy, precision, LOD, LOQ, robustness

• Acceptance criteria

  Setting criteria before the work, not after the data arrives

• Experimental design

  How many replicates, at how many levels, over how many days, and why

• Modified methods

  What a change to a standard method obliges you to re-establish

• Documentation

  The validation report and protocol an assessor expects to be handed

Courses are written around what a delegate can do afterwards, not what was presented to them. These are the outcomes this one is built to produce.

1. 01Decide correctly whether a given method needs verification or validation, and justify the decision in writing.
2. 02Design a validation study that answers the question with the smallest defensible amount of bench work.
3. 03Set acceptance criteria in advance and explain where each one came from.
4. 04Write a validation report that an assessor can follow without asking for the raw data.

The most common validation finding is not a bad experiment. It is a good experiment with no protocol written before it started, acceptance criteria that appeared after the results, and a report that records what happened without saying what was being demonstrated.

Once a method is in routine use, its validation is the only thing standing between a reported result and the question "how do you know?". This course is about being able to answer that question two years later.

Delivered over 2 days, at your site or at the laboratory, in English or Arabic. Sessions are built around real records, real data and real failures rather than slides of clause numbers, and the people delivering them spend the rest of the week doing the thing being taught.

Group size is agreed in advance — practical sessions are capped so that everyone works rather than watches. An attendance certificate is issued to each delegate on completion.

Ask about any subject not listed here. If it is not something we teach well, we will say so rather than build it for the sake of the booking.