How good should we be? Setting analytical specification goals for laboratory tests

2.15pm – 3.45pm BST, 9 June 2026 ‐ 1 hour 30 mins

Parallel session

Chair: Rav Sodi

A key role for clinical laboratories is to provide accurate results to support and enable clinical care. It is often cited that up to 70% of all clinical decisions are predicated on pathology results. Therefore, it becomes imperative that the clinical laboratory sets quality goals that ensures that the analytical service provided is fit for the intended clinical purpose. Ever since the seminal meeting in Stockholm in 1999 when the white paper Strategies to Set Global Quality Specifications in Laboratory Medicine and then the conference held in in Milan in 2014, pathology thought leaders have discussed ways to improve the quality of laboratory tests by setting analytical performance specifications (APS). It is generally accepted that the most optimal way of doing this would be base the APS on clinical outcomes, but this is rarely practicable nor feasible. A more pragmatic approach is to use biological variation data to set APS rather than base quality goals on expert opinions whose provenance is rarely known. It is now over 10 years on since the Milan conference and whilst progress has been made in standardisation and harmonisation, there appears to be a sense in the Laboratory Medicine field that there is still uncertainty with regards to assessing the quality of biochemical assays objectively. There have been numerous national incidents in recent years where erroneous results have been issued with drastic clinical consequences. The question of how good an assay should be should resonate with all of us who are working in clinical laboratories. In this session we aim to move these discussions forward. We have key international experts who will share their views on a selection of tests and this session will provide attendees with the knowledge and skills required to set achievable quality goals in their laboratories.
Learning outcomes:

  1. Understand the concepts of analytical performance specifications (APS) and total allowable error (TAE).
  2. Be able to apply these concepts in clinical practice.
  3. Use biological variation data to derive analytical performance goals for your laboratory.
  4. Be aware of recognised databases such as the EFLM Biological Variation website from where to obtain required informtion.


Practical uses for analytical performance specifications: an example involving HbA1c, Eric Kilpatrick

Analytical Performance Specifications (APSs) are agreed performance limits intended to inform how well a test should perform in order to meet a patient’s clinical needs. Unlike in many other countries, the use of APSs in the UK is not widespread and the limits applied are often inconsistent. However, once in place, APSs can become a cornerstone for assuring test quality. Quality control (QC) strategies designed to meet these APSs can objectively determine fundamental aspects of QC such as the control rule to use and the number/frequency of QC samples to be run. They can also help remove the subjectivity around deciding whether a poorly performing assay can continue to be used clinically or not.HbA1c measurement quality has come under scrutiny recently and provides an example of how APSs can assist in ensuring that test performance meets clinical requirements.
Learning outcomes:

  1. To understand how analytical performance c. (APSs) can help assure test quality.
  2. To provide examples of how APSs can be derived and implemented, including HbA1


Improving the approach to cardiovascular risk assessment: performance of modern equations for LDL-Cholesterol vs direct LDL-C assays, Tahir Pillay

Accurate assessment of low-density lipoprotein cholesterol (LDL-C) is central to cardiovascular risk stratification and decision-making for treatment, yet routine practice still relies heavily on imprecise and unstandardized direct assays and the legacy Friedewald formula. In this presentation, we review methods for LDL-C including  the   “gold standard” method  for quantifying LDL-C and explain why this complexity has driven widespread use of calculated LDL-C, alongside the use of direct biochemical assays. We contrast the performance and limitations of the Friedewald equation with modern equations such as Martin–Hopkins, extended Martin–Hopkins and the Sampson NIH2 equation, including performance at high triglyceride levels and very low LDL-C in the era of PCSK9 inhibitors. Drawing on new data and synthetic-data simulations, we will compare the performance of contemporary equations and direct assays and outline practical recommendations for laboratories seeking best practice in assessment of  LDL-C  and cardiovascular risk assessment.
Learning outcomes: 

  1. Describe the principles, advantages and limitations of reference, direct and calculated methods for LDL-cholesterol measurement.
  2. Compare the performance of the Friedewald formula with newer equations (e.g. Martin–Hopkins, extended Martin–Hopkins, Sampson) across different triglyceride and LDL-C ranges.
  3. Evaluate how the choice of LDL-C method affects cardiovascular risk classification and treatment decisions, especially in patients on intensive lipid-lowering therapy.
  4. Formulate practical, evidence-based recommendations for implementing and reporting calculated LDL-C in routine laboratory practice


Poor serum B12 (total B12) method harmonisation and unwarranted variation in diagnostic cut-offs for the diagnosis of vitamin B12 deficiency, Dominic Harrington

Serum B12 (total B12) is the most commonly used laboratory test for the evaluation of B12 status. The clinical utility of total B12 rests largely on the application of an appropriate diagnostic cut-off for the diagnosis of B12 deficiency. NICE guidance (NG239), published in 2024, specifies a threshold of 180 ng/L (133 pmol/L) as indicative of B12 deficiency, with the caveat that local validated reference ranges may be applied as an alternative to the recommended diagnostic cut-off.  This caveat is a response to the poor harmonisation of total B12 assays.  In addition to poor harmonisation, audit has revealed a greater than three-fold difference (65 – 239 ng/L) in the applied diagnostic cut-off for the diagnosis of B12 deficiency across the UK, with the greatest intra-method variation seen in users of the Beckman DxI total B12 assay (65 – 180 ng/L, n= 38 laboratories). Drawing on this data, we will consider the impact of poor harmonisation of total B12 assays and unwarranted variation in applied diagnostic cut-offs for the diagnosis of vitamin B12 deficiency when using the total B12 test. 
Learning outcomes:

  1. Variation in the application of the serum B12 test for the evaluation of vitamin B12 status.