Interlaboratory Testing: Why Proficiency Schemes Matter

Today we shall discuss what forms the very foundation and an indispensable hallmark of a high-quality analysis.

I have often had to hear something along these lines: "I don't trust analyses — every laboratory gives different results for the same sample." Or this: "There are three great lies: a barefaced lie, statistics, and laboratory analyses."

To my great regret, this is very often true on the Ukrainian market.

For many years, in my own workplace, I have fought for the right to say with confidence: "Laboratory testing has no business belonging to the category of lies."

I want to teach you, very simply, how to understand whether a laboratory can be trusted or not.

The very first and principal criterion of the reliability of analyses is whether the laboratory participates in interlaboratory proficiency testing for the specific type of analysis for which it issues results.

What is interlaboratory testing?

Interlaboratory testing (or a "round robin") is when an organisation authorised to carry out such work sends one and the same sample for analysis to many laboratories.

We take part in proficiency schemes involving more than 20 laboratories.

For the sample it receives, the participating laboratory does not know the correct values of the parameters it is measuring. It performs the analysis and sends its result to the organising body. After some time, the laboratory receives a report from the organiser containing:

● the correct results,● the permissible deviation for the obtained result,● and a conclusion as to whether the laboratory fell within the permissible deviation for the result or not.

This exercise is the principal means of confirming the correctness of one's own analyses. And when an accreditation body audits a laboratory (in granting it accreditation), this is the chief criterion it examines.

In Ukraine, I have come across two kinds of interlaboratory testing:

1. When a laboratory takes part in reputable comparisons (those involving more than 10 laboratories) and when these are arranged by a qualified organiser that is itself accredited to conduct such an exercise. The majority of such organisers are located outside the former Soviet states.

2. When two or three laboratories arrange a little private "get-together", analyse one sample, and simply swap values among themselves.

In my view, one can judge the level of a laboratory (or its approach to its work) by which of these two kinds it chooses.

If a laboratory does not take part in interlaboratory testing at all, I will not presume to judge the standard of its analyses. In that case, the phrases I quoted at the outset become entirely justified.

In good laboratory practice, a laboratory not only takes part in interlaboratory comparisons once a quarter or every six months, but also, in its day-to-day work, tests a reference sample for which it knows the exact values. Why? In order to guard itself against a possible error that may arise at any stage of the analysis: when it has been overlooked that a reagent has passed its expiry date (just as our medicine cabinets may hold tablets that should have been binned a year ago), or when an instrument has been calibrated late or incorrectly.

If a laboratory constantly analyses such a sample, it will detect the error before issuing a result to the client.

It would not go amiss to ask for the laboratory's proficiency-testing report. A conscientious laboratory will provide this information. If, however, such a report proves difficult to obtain from a laboratory, there is a strong likelihood that it simply does not exist.

I sincerely wish for the standard of Ukrainian laboratories to grow and to become one that helps businesses achieve success.

Anastasia Shuparska, Head of the Halomlab Laboratory
For more on building a quality-management system, see the blog of our partners UKRAGROTEST

13 December 2020