Pop Quiz: Is this a 4:1s violation?

Posted by Sten Westgard, MS

Is this a 4:1s rule violation? What's your guess? Or what are your questions?

Great question. The answer depends on which type of "Westgard Rules" you are implementing - as well as the Sigma-metric of the method.

If you are using classic "Westgard Rules", those set forth in 1981, you first need to see a 2s violation. That's the warning rule going off, then that allows you to check for rejection rules like 4:1s. Interestingly, there is a 2s violation, but on the other control. So that warning rule on the second control allows you to examine the other rules. In that case, your scenario finds that YES, the 4:!s rule has been violated.

If you are using a more modern "Westgard Rules", as we recommend, you don't wait for a 2s violation. There are no warning rules, you just continuously check the rejection rules. In that case, also, YES you have a violation of the 4:1s rule.

Our most recent version of "Westgard Rules" incorporates Six Sigma. In that scenario, you adjust the number of "Westgard Rules" you need based on the performance of the assay. In a 5 or 6 Sigma assay, the 4:1s rule is not necessary, and you wouldn't worry about the data you illustrate. In a method that is 4 Sigma or lower, the 4:1s rule is necessary, and therefore you DO have a violation that should force a rejection of the runs.

So, in other words, there are different ways to interpret this, and it depends on the context of the performance of the assay and the implementation of QC you have designed.

Comments

[Hassan Bayat (Sina Lab; Qaem Shahr, Iran)]

There is another consideration: Do we use 'look-back' approach or not.

If look-back is not used, i.e. R=1, then only the control results from each QC event must be evaluated independent from previous events. In that case, the answer is NO; a violation of the 4:1s rule hasn't happened (on day 2, only 3 results pass 3s at the same side).

On the other hand, if R=2, and therefore we interpret the control results from present and previous QC events collectively, then YES, the 4:1s is violated.

It must be considered that in the case of a R2 look-back approach, the patient results from last run shouldn't be reported until next QC event is passed. This way, if R=n, reporting patient results must be postponed until upcoming n-1 QC events are passed.