EXTERNAL QUALITY ASSESSMENT
Are we providing suitable analytical performance specifications for EQA?
Ceri Parfitt and Annette Thomas recount a recent study to review the strengths and weaknesses of different models for determining analytical performance specifications. The study used historical data to determine what was achievable in a real-world environment to establish clinically appropriate APS for several serum chemistry analytes.
In terms of external quality assessment (EQA), analytical performance specifications (APS) are defined as a range of values around the target which is considered acceptable for the performance of that test.1
APS provides a simple tool to allow a rapid, standardised assessment of EQA results in both numerical and graphical report formats.
A result outside the acceptable
range should alert the laboratory that their assay may produce results that are at risk of detrimentally affecting clinical decision- making. APS provides a simple tool to allow a rapid, standardised assessment of EQA results in both numerical and graphical report formats. Laboratories and point- of-care test (POCT) users must ensure that the analytical quality atained for that test is appropriate for the needs of the clinical service and the clinical utility of the test. Use of the distribution standard deviation limits around the peer group mean may imply acceptable performance for laboratories, but these may not be clinically relevant, resulting in increased risk of patient misclassification. It is therefore essential that EQA performance specification also reflects the clinical need and utility of the test.
Defining performance goals – a hierarchical approach In 1999, the World Health Organization (WHO), International Federation of Clinical Chemistry (IFCC) and International Union of Pure and Applied Chemistry (IUPAC) met in Stockholm to develop a hierarchy of five approaches for establishing performance goals.2
These were refined at
the EFLM Milan strategy in 2014 as three models:3
Model 1 – APS based on clinical outcomes Model 1 directly links variation in analytical performance to clinical outcomes, identifying where errors in measurement could lead to an incorrect clinical decision. Although this is the optimal model available, high-quality data supporting use is rarely available in practice. Typical evidence may include decision threshold studies, outcome-based studies, clinical guideline thresholds and clinical modelling studies. However, establishing causation requires collection and analysis of huge datasets. In addition, patient outcomes are typically multifactorial, with clinical decisions relying on several factors in addition to laboratory results. In practice it is applicable only to a few tests since it is difficult to show the direct effect of laboratory tests on medical outcome.4 Model 1 APS is most appropriate for
44
WWW.PATHOLOGYINPRACTICE.COM May 2026
situations with clear decision thresholds and high clinical impact such as hs- troponin and HbA1c.
Model 2 – APS based on biological variation Model 2 compares variation in analytical performance to inherent biological variation of the analyte, including both within-subject variation and between- subject variation, allowing identification of results which are clinically meaningful. This model is appealing, as it is based on clinical factors and is independent of current laboratory performance, but relies on the availability of good-quality studies on biological variation for individual analytes. More recently this has been addressed by the development of the EFLM Biological Variation Database (htps://
biologicalvariation.eu/). Unfortunately,
AdobeStock /
peopleimages.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52
