Close Server: KOPWWW05 | Not logged in

Welcome to Health Care POV | sign in | join
Press Start: Lead an Empowered Life as a Clinical Laboratorian

Is Proficiency Testing Enough to Prove Accuracy?

Published July 4, 2015 6:07 PM by Glen McDaniel

 Proficiency Testing (PT) is just part of life for most laboratories. Laboratories are sent challenges (unknown samples) which they test like patient samples and then submit their results. An agency like CAP then grades the result against those submitted by peers using the same test system. Your lab is considered as doing a good job if your results are close to those of the peer group.


Statistics like standard deviation and SDI give the impression the results generated are being checked for degree of accuracy. But are they? Well not really, the mean is used as a surrogate for the correct value and the variation from that mean is then regarded as the degree of inaccuracy (bias).


But PT materials are not patient samples and often additional material is added to preserve and stabilize the samples. How do these additives change how samples react? To assume that closeness-to-peer indicates “correct” accurate results is to be overly optimistic-and wrong.


Additives - and even shipping and storage conditions-  introduce matrix  and other effects meaning the PT sample used to give you that “atta boy” is in fact not directly comparable or commutable to a fresh sample from your patient.


So what if we just made allowance for that knowledge, knowing PT is not perfect but it’s the best we can do right now?  However that assumption raises another issue. How can we be completely sure that our test system (say an analyzer with a certain reagent) is in fact producing accurate patient results. Calibrators and quality control material often have matrix effects as well. Then there is the additional issue of comparability among various test systems.


Suppose research shows that a clinical condition has certain cut points or diagnostic targets. Given the degree of uncertaintity and variation how can a clinician be confident making clinical decisions based on such results? We are already familiar with harmonization for cholesterol and with the use of INR (versus PT)  to guide warfarin anticoagulation. In both cases there is much less inter-laboratory variation than there used to be.


That brings us to the areas of accuracy and harmonization. Accuracy is “trueness of result” while harmonization is a system of ensuring results for an analyte are truly comparable; even performed by different methods, at different sites, on different platforms.


CAP has recently introduced Accuracy-Based Surveys for various analytes like lipids, Vitamin D, Creatinine, testosterone and Hemoglobin A1C.


Participation is a totally voluntary process and laboratories might not want to subject themselves to extra scrutiny and the possibility of “failing” even if results are within the peer group. One advantage of the Accuracy Based Program (ABP) is that manufacturers can re-evaluate their products (reagents and test system) and even revisit their calibration traceability.


Wouldn’t it be wonderful if your PT results indicated how you were doing in an absolute sense, and not simply compared to your peer group who may have the same bias? Wouldn’t it be great to have tests systems giving equivalent results across platforms and laboratorians and clinicians having the confidence that their “numbers” are both accurate and comparable and can be used to confidently guide patient care decisions?



Interesting concept. I agree that is something we take for granted. Every once in a while there is no concensus among peers. Sometimes it is a numerical value and sometimes it is even qualitative or positive/negative. CAP will then claim it was due to storage or handling of PT  material by lab. But sometimes it's just an unstable PT to start with.

Thomas V., Lab Supervisor July 5, 2015 5:33 PM
New York NY

leave a comment

To prevent comment spam, please type the code you see below into the code field before submitting your comment. If you cannot read the numbers in the image, reload the page to generate a new one.

Enter the security code below:


About this Blog

Keep Me Updated