The Crucial Role of Reliable Biological Variation Data in Defining Analytical Performance Specifications

BV estimates have many applications in laboratory medicine, most importantly to set analytical performance specification (APS) for imprecision, bias, maximum allowable measurement uncertainty and allowable total error limits (ATE) that specify the maximum amount of error – both imprecision and bias combined – that is allowed for a measurand. There are two components of BV: within-subject BV (CVI) which refers to the random variation in concentration or activity around the homeostatic setting point of a measurand in an individual, and between-subject BV (CVG), reflecting the variation between homeostatic set points in a group (1). Some measurands e.g. electrolytes are characterized by very narrow within-subject BV estimates while other measurands are associated with wide BV estimates. Usually, BV data are generated prospectively by collecting a series of specimens from a cohort of individuals, with BV components estimated by statistical methods such as ANOVA or Bayesian analysis. APS/ATE based on BV components are derived from CVI and CVG estimates in a three-level model that includes minimal, desirable and optimal performance (2). This approach allows the user of BV data to choose an error limit that matches their analytical system.  

However, the usefulness and accuracy of these limits are impacted by the quality of available BV data. Over the last 40 years, many BV studies have been published for a wide number of measurands. The first literature-based overview of BV estimates was compiled by Dr. Carmen Ricos et al in 1999 and was kept updated at the Westgard website until 2014 (3-4). At the time, this database provided a useful resources of BV data, but the quality of the studies from which presented BV estimates were calculated required further review. To address this issue, the European Federation of Clinical Chemistry and Laboratory Medicine Task Group on the Biological Variation Database and the Working Group on Biological Variation (TG-BVD and WG-BV) published in 2018 the Biological Variation Data Critical Appraisal Checklist (BIVAC), which is a tool to critically appraise BV publications with regard to study design, preanalytical handling, analytical methods and statistical analysis (5). Using the BIVAC, the EFLM TG-BVD has for the last decade systematically reviewed published BV studies and compiled this data in the EFLM Biological Variation Database (BV database). This database which is freely available to users worldwide provides the most updated and largest number of the individual BV estimates, including the results of the BIVAC assessment and a detailed Minimum Data Set derived from each study, as well as global BV estimates derived from meta-analysis of BIVAC compliant studies. Furthermore, it also delivers automatic calculations of APS including ATE and reference change values (RCVs) for over 190 measurands. The database also lists BV data for many other measurands, where there are not sufficient data fulfilling inclusion criteria for meta-analysis. Data from BV publications that have been deemed noncompliant per BIVAC are excluded from the meta-analysis and not displayed in the database. 

The EFLM BV database is continuously being updated as new BV estimates are published in the literature. However, users of the data should be aware that some of the currently available BV estimates may be based on very limited data sets and therefore should be applied with caution. There is a continuous need for reliable BV studies to be delivered as new measurands emerge and are available in clinical practice. As a result, the WG-BV and TG-BVD in 2024 released a step by step guidance framework for delivery of BV data with highest quality (6), which is provided on the EFLM website.

In conclusion, the EFLM Biological Variation Database provides a rich, detailed resource of BV data that is continuously updated and where users can assess the details of each dataset to decide what data are suited for their particular purpose. The EFLM BV database is freely available and also provides automatic calculation of APS, including ATE and RCV for use in laboratory practice worldwide.

References

1. Fraser CG. Biological variation: from principles to practice. AACC, 2001. 
2. Fraser CG et al. Proposals for setting generally applicable quality goals solely based on biology. Ann Clin Biochem 1997; 34: 8-12 
3. Ricos C et al. Current databases on biological variation: pros, cons, and progress. Scandinavian Journal of Clinical Laboratory Investigation 1999;59:491-500.10. 
4. Perich C et al. Biological variation database: structure and criteria used for generation and update. Clin Chem Lab Med. 2015;53:299–305. 
5. Aarsand AK et al. The Biological Variation Data Critical Appraisal Checklist: A Standard for Evaluating Studies on Biological Variation. Clin Chem. 2018, 64(3): 501-514 
6. Bartlett WA et al. A standard to report biological variation data studies - based on an expert opinion. Clin Chem Lab Med 2024