Blood Tests - accuracy and precision; why trends are important

[AussieNeil]

With more queries recently about blood test results, I though it appropriate to explain a couple of significant reasons why looking at trends is important. Anyone wishing to record their blood tests can make use of the templates referenced in this post: healthunlocked.com/cllsuppo... The templates include lots of useful information explaining the relevance of each result to someone with CLL.

There are many influences over the final results that appear on your blood test print-out, such that you can get a wide range of test values from the same blood sample. For some interesting background on this, read SeymourB's long reply in this post: healthunlocked.com/cllsuppo...

In the bulls eye illustration above, what we want to achieve is hitting the target right in the centre (the true blood test result), time after time. To do that we need both precision (repeatability) and accuracy. Accuracy is determined by the design of the instrument and its regular calibration to a standard. Precision is determined by how well controlled the factors are that affect repeatability.

Without understanding these test instrument limitations, we can become concerned over apparent changes in our blood test results that are due to accuracy and precision limitations in the test equipment - our actual blood test value could well be the same or only slightly different from the last test result. Monitoring for trends overtime takes out the influence of these variances from the true result (and we'll hopefully be less anxious).

Dr. Susan Leclair has over 40 years of professional clinical hematology and oncology lab experience and has provided many excellent Patient Power videos where she explains how to interpret blood test results. Here she explains the limitations in blood testing that can cause variations: patientpower.info/video/is-...

Dr. Leclair's other Patient Power videos can be found here: patientpower.info/bio/susan...

I've taken the illustration from The Skinny on Precision and Bias: Aiming Our Sights on Precision: aashtoresource.org/universi...

Wikipedia also has a reference on these important limiting factors

en.wikipedia.org/wiki/Accur...

Neil

[Jm954 Administrator]

Can I also add another factor - sorry!

The reference range using one manufacturer's analyser and kit, or even the same analyser and a different manufacturer's kit can be very different.

This is less so with CBCs as there are internationally agreed ranges that the analysers are calibrated against but with other tests the reference range can vary wildly so it's important to check the result against the reference range stated on the report before panic sets in.

Labs often have more than one analyser doing the same tests and it's important that they check the inter-analyser variability. I had my CBC done in the clinic on a small stand alone analyser and the result in my Hb differed between 131 and 123. Not too important at this level but could have been if it had been 84 and not 90.

So, as Neil says, trends are more important than any individual result but outliers, unexpected deviations from what was expected, need further investigation and most often the first thing to do is to repeat the test.

[CAGIULIANO]

Thank you, very interesting and true in almost all controlled results of many sciences

[keepfit123]

What about 17p, 11q, 13p and Trisomy12? Can't you still be negative for these even though you have certain percentages.

[AussieNeil Partner Administrator in reply to keepfit123]

With those tests it's a percentage threshold; above it you are positive and below, negative. Unfortunately there have been cases where there wasn't international agreement on the threshold. Probably more importantly, the results can vary based on where in the body the sample was taken.

Neil

[shoshanaz in reply to keepfit123]

If the percentages are near the threshold, then it may depend on the doctor's conclusion, not the lab. The lab is looking at a margin of error. The doctor is looking at treatment options.

Initially I was told I was not 17p-, having a threshold just under the lab cutoff. Now, with the same results, was told that I was 17p- because of the presence of some, even a low percentage of such cells. The reason for that is a concern that some treatments which are ineffective on 17p- may have a consequence of raising that percentage.

So you can be negative for those traits per the lab, but still be considered positive in terms of future treatment options.

I would expect that the test will be repeated prior to treatment, as the numbers can and do change. It is harder to identify "trends" with only one or two tests. So there is that, as well.

[Cllcanada Top Poster CURE Hero in reply to shoshanaz]

Remember its a moving target, particularly with FISH. Many research hospitals use different numbers for thresholds of genetic FISH markers, particularly Mayo clinics...

In FISH the sample is only 200 B cells, out of milions, I think we can assume that the results are approximate at best...

Its a sampling...

~chris

[cllady01 Former Volunteer]

So, we are each standing in the middle of our own Teeter-Totter/seesaw, balancing as best we can with what we know up to this point.

Sorry, the image came and had to be described.

[Hidden]

Thank you. This is also very helpful, along with your most welcome replies. Actually, you told me more than my Oncologist/Hemotologist ever did.

[keepfit123]

Can anybody say what the actual cut off percentages are for the four categories regarding UK haematology clinics?

Cheers,

keepfit123

[Cllcanada Top Poster CURE Hero in reply to keepfit123]

I have no idea... ask a pathologist perhaps... it might vary as well by probe manufacturer..

This discusses it in part

ncbi.nlm.nih.gov/pmc/articl...

10 years ago the CLL Consortiuum hospitals tried to get standardizaion.. not sure how well they did.

ncbi.nlm.nih.gov/pmc/articl...