Q & A: What is a "control range"?

Posted by Sten Westgard, MS

An interesting question came in through email (we always want to hear from you!):

"1. Can we apply Westgard multirules to hematology control?

2. I read in one paper that the hematology control range is calculated as mean +/- 2.5 SD, is this correct and if not how can I calculate own laboratory hematology control range?"

The answer, after the jump...

It's extremely common to get the first question. Since Dr. Westgard wrote the multirule paper and discussed chemistry analytes, there is a strong association of the rules with chemistry tests. However, multirule QC is not exclusive to chemistry. It can be used in many different areas, including hematology. It is not uncommon to see "Westgard Rules" applied to hematology tests.

If you think about it, multirule procedures are just about numbers, and those numbers don't have to originate from chemistry tests. We've seen applications of multirules to water quality testing, carbon-dating, even rocket engine testing. If you've got numeric results for quality control, there's a good chance you can use some type of multirule QC procedure. I like to say that the control rules are ecumenical about the source of the data.

Now, as to the second question, we find that a "control range" often depends on definition. The definition mentioned above is unusual to us. It is better not to focus on the "control range", but on specific control limits.

Using a QC Design approach, you can select the appropriate QC procedures and define exactly which control limits are important. For excellent assays, you may only need 3s control limits. For problematic assays, you may need "Westgard Rules" and limits drawn at 1s, 2s, 3s.

The best approach (QC Design) is to do the right amount of QC for each assay, which means taking into account the performance of the method and the quality required by the clinical use of the test. That specific combination of controls and control limits will make up the "control range" that you use.

Again, thank you for this question. We encourage all our visitors to send in their questions, their problems, and their data.