CCA-13356; No of Pages 6 Clinica Chimica Acta xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization Neil Greenberg ⁎ Neil Greenberg Consulting Services, LLC, 40 Chipping Ridge, Fairport, NY 14450, USA

a r t i c l e

i n f o

Article history: Received 11 June 2013 Received in revised form 30 December 2013 Accepted 30 December 2013 Available online xxxx Keywords: Standardization Traceability Metrology Harmonization Calibration SI

a b s t r a c t An increasingly important quality objective in laboratory medicine is ensuring the equivalence of test results among different measurement procedures, different laboratories and health care systems, over time. In recent years, interest in sharing a single patient's clinical laboratory data, regardless of where the measurements are performed, has moved out of the domain of the scientific community and spilled over into the domain of regulators, lawmakers and the general population in many parts of the world. For all parties involved in the dialog, establishing and maintaining a clear understanding of the essential concepts that are vital to achieving global equivalence among clinical laboratory measurements have therefore become a priority. Concepts that are critical to this discussion include standardization, traceability and harmonization. This report provides an updated discussion and practical definitions for these terms and others that are linked to metrological principles. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Increased analytical accuracy is a hallmark of improved quality in a laboratory measurement system, and is strongly associated with improved health care. Because multiple measurement methods and procedures may be available for a given measurand, harmonization of test results, especially in reference to internationally agreed measurement standards, creates the opportunity for sharing results among different health care systems as well as across geographic boundaries and time. In addition, clinical laboratory measurement harmonization is a critical issue to be addressed in successfully designing clinical trials intended to demonstrate the efficacy of new and improved therapeutic agents [1] and in translating clinical research into routine practice. The benefits of improved performance and harmonization of reported values for different clinical laboratory measurement procedures for a given measurand are illustrated by quality enhancements in routine clinical laboratory serum cholesterol measurements realized over the thirty-year period from 1970 to 2000 [2,3], which coincided with a profound reduction in the mortality rates for coronary heart disease

Abbreviations: CHD, coronary heart disease; IVD, in vitro diagnostic medical device; EU IVD Directive, Directive 98/79/EC of the European Parliament; ISO, International Organization for Standardization; VIM, International Vocabulary of Metrology; GUM, Guide to Expression of Uncertainty in Measurement; BIPM, Bureau International des Poids et Mesures; JCTLM, Joint Committee for Traceability in Laboratory Medicine; WHO, World Health Organization; RMP, Reference Measurement Procedure; ICHCLR, International Consortium for Harmonization of Clinical Laboratory Results; NMI, National Metrology Institute. ⁎ Tel.: +1 585 317 8791. E-mail address: [email protected].

(CHD) in the US [4]. While many of the improvements achieved in diagnosis and treatment of CHD are attributable to a variety of factors not directly related to the clinical laboratory, the economic yield of the investment in harmonization of blood cholesterol measurements in the US alone is very conservatively estimated to be in the range of several hundred million to several billion US dollars annually [5]. The need for harmonization of clinical laboratory measurements has also been emphasized as an essential element of key legal constructs that regulate the delivery and commercialization of laboratory medicine products and services in many world markets. The regulatory focus on harmonization is exemplified by the EU IVD Directive [6], where it is an essential requirement that “…the traceability of values assigned to calibrators and/or control materials must be assured through available reference measurement procedures and/or available reference materials of a higher order.” Implementation of this EU regulatory requirement was further supported by the publication of an ISO standard concerning traceability of values assigned to calibrators, ISO 17511:2003 [7]. A similar regulatory expectation is defined in the U.S. Code of Federal Regulations 21CFR Part 809.10 (b)(12) [8], where it is stated that manufacturers of commercial IVD reagents must provide information to users regarding “…specific performance characteristics … accuracy, precision, specificity, and sensitivity.” As a result of the recent launch of the International Consortium for Harmonization of Clinical Laboratory Results (ICHCLR) measurement harmonization initiative under stewardship of the American Association for Clinical Chemistry (AACC) [9–11], discussion of the concept of harmonization among different measurement procedures for the same measurand has achieved increased visibility among the various stakeholders, such as routine and reference laboratories, IVD reagents,

0009-8981/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2013.12.045

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045

2

N. Greenberg / Clinica Chimica Acta xxx (2014) xxx–xxx

systems and calibrator manufacturers, and accreditation and regulatory bodies. In the broadest sense, harmonization of measurements refers to any process that enables the establishment of equivalence of reported values produced by different measurement procedures for the same measurand, i.e. the quantity intended to be measured [12]. In laboratory medicine, harmonization of results for a given measurand can be achieved by two main methods, where the selected method is predetermined based on availability of suitable reference measurement procedures and reference materials for the target measurand. Method 1, commonly referred to as standardization, is the conventional approach, and is dependent on application of a well-understood reference measurement procedure and commutable reference materials to establish the calibration for each available routine measurement procedure. The second method, Method 2, to be considered only when reference measurement procedures are not available, is based on application of a factor or other data manipulation strategy to manufacture arbitrary equivalence of otherwise disparate results generated by the various (non-standardized) routine measurement procedures. The stated focus of the ICHCLR harmonization initiative is the achievement of harmonization of reported values among multiple routine measurement procedures for the same measurand when there are no reference measurement procedures (RMPs) available [9–11], and the development of practical solutions for the establishment of equivalence of reported values among different measurement procedures that are not traceable to SI. Realization of these initiatives will enable the practical implementation of common cutoff values and/or reference intervals, as well as retrospective application of harmonized clinical decision values to legacy clinical trials data that were used to establish the clinical sensitivity and specificity characteristics for many of these types of measurands. The complete rationale for the ICHCLR Harmonization Initiative has been described elsewhere [10,11]. Three key words in the current dialog regarding improving the equivalence of reported values among different clinical laboratory measurement procedures for the same measurand are therefore standardization, traceability, and harmonization. With the co-existing technical and legal contexts of these terms that are so central to defining the quality of laboratory measurements, the definitions of these and other related terms such as “higher order reference materials,” are at times confused and misused among the diverse range of participants involved in the discussion. 2. What is standardization? Standardization is probably the most widely used of these three key words, as it is a term that is well rooted in many technical and non-technical disciplines and languages worldwide. For purposes of this discussion, the most appropriate supporting definition is found in the International Vocabulary of Metrology (VIM) [12], which defines a measurement standard as the… “realization of the definition of a given quantity, with stated quantity value and associated measurement uncertainty, used as a reference.” The term quantity as used above describes a measurable property of a given substance or material, where the magnitude of the measured property can be described with a numerical value and a reference (i.e. a measurement unit.) Examples of the kinds of quantities that comprise the Système International (SI) “base units of measurement”[13] and which are particularly important in the field of laboratory medicine include mass (kilogram, kg), length (meter, m), amount of substance (moles, mol), time (second, s) and thermodynamic temperature (Kelvin, K). Note that these SI base units are also essential to the definition of other important measured SI quantities, called derived SI quantities [13], which are obtained from mathematical equations supported by various SI base unit quantities, e.g. volume (cubic meter, m3), mass density (kilogram per cubic meter, kg/m3), amount of substance concentration (mol/m3).

The expression, “realization of the definition of a… quantity” concerns the establishment (by agreement) of either a physical material or a highly reproducible physical phenomenon or constant, which by definition and/or by agreement is the ultimate standard, i.e. it physically embodies the measurement unit, or a submultiple or multiple thereof. For example, the basic metric unit for mass, the kilogram, exemplifies a definition of a particular quantity for which the “realization” is based on a physical material. The prototype kilogram (a cylinder made of a platinum–iridium alloy) is maintained by the Bureau International des Poids et Mésures (BIPM) in Sèvres, France. To effect the global promulgation of the standard kilogram at the regional and national level in support of scientific and industrial applications of mass measurement, the prototype reference kilogram is used to make metal alloy copies that are maintained by various national metrology institutes (NMIs) worldwide. Given the above definition of a measurement standard, standardization is a harmonization process in which the values assigned to hierarchically lower order standards (e.g. a value assigned to a commercial stainless steel reference mass of approximately 1 kg, intended for calibration of commercial field balances) are systematically determined either by a direct comparison to the highest order reference standard available (i.e., the prototype platinum–iridium alloy kilogram at the BIPM), or indirectly, by comparison with an intermediate (lower order) reference standard, such as a platinum–iridium “copy” prototype kilogram maintained at an NMI. In the fields of analytical chemistry and laboratory medicine, for the practical implementation of a harmonization process that can be characterized as standardization, trueness is transferred by way of a systematic step-wise process from highest order available (traceable to SI) standards to successively lower-order routine/commercial calibrators. Such assigned-value transfer processes may include the use of various hierarchically intermediate-order reference standards as calibrators in multiple measurement stages, as long as the provider of such intermediate-order reference standards makes available adequate information concerning their assigned values and uncertainties, including a description of the applicable calibration hierarchy and higher order references (material standards and measurement procedures), with a cumulative accounting for uncertainties propagated from any higher order references. The accompanying information for any applied intermediate-order standards should also state what is known about the compatibility of such standards with the available measurement procedures, especially with respect to the performance of these reference materials in comparison to the types of samples usually intended to be measured (e.g. patient samples in the field of laboratory medicine) for the quantity of interest. Further comments on this point are given below in the discussion on traceability. 3. What is traceability? The concept of calibration traceability, although certainly not a new one, has received increased emphasis in the field of laboratory medicine in recent years, primarily because of the very specific language in the essential requirements of the EU IVD Directive [6], declaring that “the traceability of values assigned to calibrators and/or control materials must be assured…” Although there are many kinds of traceability in a broad array of disciplines (e.g. document traceability, accounting traceability, design traceability, information source traceability, etc.), the type of traceability that is the focus of the essential requirements of the IVD Directive is metrological traceability. Metrological traceability is defined in the VIM, clause 2.41[12] as the “property of a measurement result whereby the result can be related to a reference (a standard) through a documented unbroken chain of calibrations, each contributing to the measurement uncertainty.” The International Laboratory Accreditation Cooperation (ILAC) Policy on Traceability of Measurement Results [14] states that the essential elements of a traceable measurement include (1) an unbroken chain

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045

N. Greenberg / Clinica Chimica Acta xxx (2014) xxx–xxx

of comparisons that starts with an agreed reference; (2) calculated uncertainty of the final measurement that accounts for all cumulative elements of uncertainty throughout the chain of comparisons; (3) fully documented procedures and outputs from each element or process across the chain of comparisons; (4) evidence of the competence of the laboratories conducting each process, preferably through evidence of accreditation; (5) the highest order elements of the chain of comparisons should, where available, be primary standards linked to realization of SI units; (6) defined calibration intervals, the length of which are based on intended use and performance requirements, sources of uncertainty, and knowledge of the stability of each of the measuring elements and devices in the traceability chain. In simple terms, metrological traceability refers to the lineage of the calibration of a measurement procedure. If a measurement is claimed to be metrologically traceable, it is implied that the validity of routine measurements can be assured because there is a clearly defined set of relationships linking the routine measurement procedure and its calibration to the highest available reference (and the relevant SI unit) for that quantity. Metrological traceability therefore requires that there be an established hierarchical relationship defined by a series of reference materials and reference measurement procedures cascading in sequence from the highest order available reference material and/or measurement procedure to lower order measurement procedures and materials, and ultimately to the final reported value of the measurement. The validity of a traceable calibration, and therefore the assurance of the validity of routine measurements, is dependent on the successful implementation of each step in the defined calibration hierarchy. A critical aspect in achieving the successful implementation of any calibration transfer process is establishing the ability of a given reference material (calibrator) to appropriately emulate the performance of the intended (patient) samples with each measurement procedure for which the calibration standard is to be deployed. This characteristic of a reference material (calibrator) is known as the commutability of the material. Commutability is defined (VIM, 5.15 [12]) as the …“property of a reference material, demonstrated by the closeness of agreement between the relation among the measurement results for a stated quantity in this material, obtained according to two given measurement procedures, and the relation obtained among the measurement results for other specified materials,” (i.e. patient samples). The commutability characteristics of a calibrator (i.e. a reference material, as in the VIM definition) are observed when comparing differences in measurement results for a panel of patient samples among two or more different measurement procedures for the same measurand, especially when these measurement procedures are based on different measurement principles, and ideally where at least one of the procedures is a reference measurement procedure. If the chosen calibrator is commutable, the calibrator will demonstrate performance in these comparisons that falls on the line of identity in a difference plot of the results for the patient samples, in each two-way method comparison. For the successful implementation of a traceable calibration hierarchy, it is obligatory to establish that each reference material (calibrator) is suitable for use with each applicable measurement procedure in the chosen calibration transfer hierarchy, if the goal is to achieve a realization of the fundamental SI unit at the level of routine measurement procedures. The commutability of a reference material with respect to chosen measurement procedures in a given calibration hierarchy is clearly an important feature in the determination of its suitability for use. A more detailed discussion regarding appropriate methods for establishing the commutability characteristics of a reference material is provided in CLSI document EP30-A [15]. Another very important element of metrological traceability, as described in the ILAC Policy on Traceability [14], is the concept of the propagation of measurement uncertainties in a cumulative fashion, from the top of the metrological hierarchy to the lowest order measurement procedure and its reported values. Accurate estimation of the

3

uncertainty associated with the final measurement therefore requires a consistent approach and full communication among all parties participating in the various steps in a “traceable” calibration hierarchy. An invaluable resource for in-depth guidance on the estimation of measurement uncertainty is the Guide to the Expression of Uncertainty of Measurement (GUM), ISO/IEC Guide 98–3:2008 [16]. Additional practical guidance is available in CLSI EP29-A [17] and in Eurachem/CITAC Guide CG4 [18]. A higher order reference can be characterized in terms of the type and quality of certifications associated with the particular reference. According to VIM, clause 5.14 [12], a certified reference material (CRM) is a “reference material accompanied by documentation issued by an authoritative body and providing one or more specified property values with associated uncertainties and traceabilities, using valid procedures.” Similarly, VIM clause 2.7 [12] defines a reference measurement procedure (with the label of “reference” intended to imply a higher order measurement procedure) as a “measurement procedure accepted as providing measurement results fit for…assessing measurement trueness of… values obtained from other (i.e. lower order) measurement procedures…in calibration, or in characterizing reference materials.” A very useful resource for identification and selection of available higher order reference materials and reference measurement procedures in the field of laboratory medicine is the database maintained by the Joint Committee for Traceability in Laboratory Medicine (JCTLM) [19]. The JCTLM was created in 2002 under a Declaration of Cooperation between the International Committee for Weights and Measures (CIPM), International Federation for Clinical Chemistry and Laboratory Medicine (IFCC), and the International Laboratory Accreditation Cooperation (ILAC) to support implementation of the EU IVD Directive. The JCTLM establishes and maintains lists of available higher order reference materials, reference measurement procedures and reference measurement laboratory services in the field of laboratory medicine based on a determination of compliance with the relevant international standards (ISO 15193 — Requirements for Reference Measurement Procedures [20], ISO 15194 — Requirements for Reference Materials [21], or ISO 15195 — Requirements for Reference Laboratories [22]). An important precaution with respect to the listings of higher order reference materials provided in the JCTLM database is that, although the intended use of a particular material may be stated, information with regard to the commutability status of a listed reference material at various levels throughout a potential calibration hierarchy for a given measurand is often unknown and/or not provided by the sponsor or provider of the reference material. As stated above, if the goal of a metrologically traceable calibration hierarchy is to achieve a realization of the fundamental SI unit at the level of a routine measurement procedure, successful implementation of the calibration hierarchy requires that each reference material (calibrator) within the defined calibration transfer scheme be determined to be suitable for use with the applicable measurement procedures. There are a variety of models described for the establishment of traceable calibrations in the field of laboratory medicine, several of which are documented in detail in ISO17511:2003 [7]. The applicability of each of the described calibration traceability models is dependent on (1) availability of higher order references (measurement procedures and/or materials) and (2) the availability of commutable intermediate (lower order) matrix-matched reference materials suitable for the transfer of assigned values to routine (field) calibrators. Within the ILAC Policy on Traceability [14], Note 1, it is acknowledged that, “…due to the nature of some tests (measurement procedures), it is not possible, realistic or relevant to expect traceability of measurement results to be demonstrated.” This caveat relates to the important qualifying language embodied in the essential requirements of the EU IVD Directive [6] that “traceability must be assured through available reference measurement procedures and/or available reference materials of a higher order.” This statement has very practical implications, since it is well known that in the field of laboratory medicine,

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045

4

N. Greenberg / Clinica Chimica Acta xxx (2014) xxx–xxx

although routine clinical laboratories may offer measurements for perhaps 1000 or more different measurands in various biological tissues and fluids, the number of clinical laboratory measurands with JCTLMlisted higher order reference measurement procedures and/or reference materials as of this writing numbers well below 200. The lack of availability of higher order reference measurement procedures and/or reference materials in the field of laboratory medicine was also noted in ISO 17511:2003 [7], and was addressed to some degree with information regarding several possible models for the establishment of a traceable calibration, in ISO 17511 clauses 5.4, 5.5 and 5.6. ISO 17511, 5.4, presents a model for addressing calibration traceability in situations where there is only an international conventional reference measurement procedure (not higher order), but there is no international conventional calibrator and no metrological traceability to SI units. ISO17511, 3.12, defines an international conventional reference measurement procedure as a measurement procedure yielding values that are not metrologically traceable to the SI but which by international agreement are used as reference values for a defined quantity. Similarly, ISO17511, 3.11, defines an international conventional calibrator as a calibrator whose value of a quantity is not metrologically traceable to the SI but is assigned by international agreement. ISO 17511, 5.5, presents a model for addressing situations where an international conventional calibrator exists, but there is no international conventional reference measurement procedure, and there is no metrological traceability to SI units. An example of this type of international conventional calibrator is the WHO 2nd International Standard (2003), Hepatitis B surface antigen, subtype adw2, genotype A [23]. An important precaution with regard to the calibration traceability model given in ISO 17511, 5.5, is that, in addition to a lack of traceability to SI units for values assigned to international conventional calibrators, frequently there may also be a lack of stated uncertainties for the arbitrarily assigned values. Additionally, many international conventional calibrators have not been studied for commutability, and should not be presumed to be commutable or suitable for use with available routine measurement procedures, unless data and/or published studies are available to demonstrate their suitability. Once again, it cannot be over-emphasized that if the goal is to achieve a realization of any defined units at the level of routine measurement procedures (including arbitrary units, international conventional units, and fundamental SI units), it is obligatory to establish that each reference material (calibrator) in a defined calibration transfer process is suitable for use with each applicable measurement procedure in the chosen calibration transfer hierarchy. Finally, ISO 17511, 5.6, deals with cases where, due to a lack of available references (higher order reference materials and/or reference measurement procedures), the only reference that exists is within the domain of one particular commercial IVD reagent kit manufacturer. At present, this situation frequently occurs in laboratory medicine, and is characterized by the absence of metrological traceability to SI units. In these circumstances, the manufacturer of the commercial reagent and its accompanying commercial calibrator for the measurand of interest usually implements an arbitrary internally established calibration and value assignment scheme for sequential production batches of the reagent and calibrator. Such calibration schemes are often supported by a stable master calibrator lot maintained by the manufacturer. In these cases, the master calibrator lot comprises the only existing realization of the arbitrary measurement unit for the selected measurand, which itself may also be defined only in the context of the particular commercial measurement system. The defined measurement unit does not relate to any external reference (material or measurement procedure), and it is usually never transferred beyond the domain of the single manufacturer. While a metrologically traceable scheme can in theory be defined in these cases, such calibration traceability will likely fall short in terms of supporting the ultimate objective in laboratory medicine of establishing equivalence of reported values for a particular measurand among

different measurement procedures, and among different laboratories and health care systems over time. 4. What is harmonization? The calibration traceability cases described in ISO 17511, 5.5 and 5.6, define the current state of the art for many, perhaps the majority of, very important measurands in the clinical laboratory (i.e. only measurement references that exist are either an international conventional reference material or some unique reference material established and maintained within the domain of a single manufacturer, but there are no RMPs and no traceability to SI units.) Brief lists of examples of measurands without RMPs have been cited elsewhere [10] and include substances such as thyroid stimulating hormone (TSH), human chorionic gonadotropin (HCG), prostate-specific antigen (PSA), troponin I, natriuretic peptides. Characteristics common to many of these measurands, often proteins or peptides with complex molecular structures, are that (1) they exhibit molecular heterogeneity, (2) they are frequently present in body fluids as heterogeneous mixtures due to varying degrees of molecular modification caused by post-translational in vivo processes such as glycosylation, and (3) the clinically relevant measurand(s) may be some or all of the various molecular forms present in the patient sample. In the absence of an RMP, various in vitro diagnostics reagents and systems manufacturers have responded to the demands of laboratory customers by providing platform-specific measurement procedures, typically immunoassays, often with unique configurations of reagents and antibodies, proprietary calibration schemes, and widely varying measurement principles. This development has contributed to at times substantial difficulty in comparing and interpreting results for a single measurand on a panel of samples measured with different immunoassay measurement procedures, as has been previously highlighted for example in the cases of TSH [24] and HCG [25]. Under these circumstances, a significant challenge remains with regard to the objective of achieving equivalence among results from different measurement procedures. For a single measurand, there may be perhaps 20 or more different commercial measurement procedures for the same measurand available to clinical laboratories, all of which may be calibrated according to different and unrelated standards. The product calibrations for these commercial assays are often supported only by a noncommutable reference material (or a reference material that has not been validated for commutability) such as certain international conventional calibrators or other independent and unique proprietary reference pools that may be sustained only within the domain of a single manufacturer. For these classes of measurement procedures the actual measurand may also be poorly defined and/or different among the different routine procedures, especially where variable specificity for different epitopes of the measurand (for example, due to use of different antibodies in different immunoassay procedures) contributes to divergent measured values in certain sets of patient samples. The Oxford Dictionary of American English [26] defines (to) harmonize as a verb meaning (to) “make consistent.” Harmonization, or consistency of measured values among different measurement procedures for the same measurand, can be achieved by two major approaches. Method 1 covers traditional hierarchical metrological schemes supported by higher order standards (reference measurement procedures and/or materials) traceable to the SI (i.e. standardization.) Method 2 uses intermethod comparison schemes coupled with mathematically derived corrections for observed systematic differences. Method 2 should only be considered when it is not possible to apply Method 1, especially in cases where there is no RMP. Indeed Method 2 is recommended by the ICHCLR Harmonization Initiative [9–11] as a useful strategy for achieving traceable calibrations when harmonizing measurement procedures for measurands with no available RMP. Method 1 harmonization applies a process that begins at the top of the metrological hierarchy, starting from the highest order available

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045

N. Greenberg / Clinica Chimica Acta xxx (2014) xxx–xxx

references (materials and/or measurement procedures), successively passing “trueness” along to lower order measurement procedures and calibrators. For Method 1 harmonization processes to be valid and to satisfy requirements, the commutability of reference materials used to transfer calibration values must be understood and suitable for their intended uses, at each level throughout the chosen hierarchy of traceable calibration steps. If commutability requirements and characteristics have not been satisfied, or data are not available in the public domain, the traceability and validity of the values reported at the level of routine measurements can be called into question. Method 2 harmonization is a process that begins with correlation and method comparison studies conducted by measuring shared panels of human samples among any number of routine (commercial) measurement procedures that are at the bottom of the metrological hierarchy, from the SI viewpoint. The values reported by the various routine (commercial) measurement procedures in Method 2 may be initially traceable to a number of different and independent calibration standards (not traceable to the SI). The process, if successful, ultimately reconciles these differences in reported values through the use of empirically derived mathematical adjustments, ultimately reporting the values from all of the routine measurement procedures (included in the harmonization process) in terms of a singular unified arbitrary unit. It is extremely important, when using Method 2, to ensure that empirical systematic corrections applied do not allow for significant residual (patient sample dependent) differences, in order to achieve the stated goal of establishing equivalence of the measured values among different measurement procedures. The occurrence of significant (large, patient sample dependent) residual differences (i.e. unique non-systematic differences in some of the values observed among the different measurement procedures) can be more simply described as the “scatter” around the line of identity (or normalizing function) in a simple patient sample panel comparison study, among two or more measurement procedures for the same measurand. To the extent that large differences (i.e. differences approaching or exceeding predefined quality requirements) occur for a large percentage of the samples tested in such a study, the calibration correction or adjustment determined may be subject to an increasingly large uncertainty, compromising the validity and reproducibility of the correction to be

applied. Thus, the quality requirements and measurement performance goals defining “equivalent” values must be clearly established. Routine measurement procedures that do not achieve the quality requirements, due to excessive residual differences remaining after correction for systematic differences, should not be claimed to be “harmonized.” For review, Table 1 provides a high level overview of the applicability, mode of implementation, and key differences between Method 1 and Method 2 for achieving equivalence (harmonization) in measured values among two or more measurement procedures for the same measurand. In summary, harmonization according to a hierarchical standardization scheme (Method 1) embodies the principles of a traditional metrologically based measurement system, and includes built-in sustainability features especially where there is an available hierarchy of higher order and lower order references, including materials and measurement procedures. Lack of available higher order standards (especially reference measurement procedures) is a major obstacle to further development of traceable calibration schemes for a large proportion of the hundreds of measurands of interest to the clinical laboratory. Achievement of harmonization among measurement procedures in the absence of higher order standards requires the adoption of more empirical processes (i.e. Method 2), which may result in greater measurement uncertainty as well as increased risk for non-sustainability of the resulting harmonized calibrations. Nevertheless, some compelling examples of the feasibility of the Method 2 empirical harmonization approach as advocated by the ICHCLR have been published in recent literature, particularly for TSH [27] and insulin [28]. Establishment of internationally agreed protocols and disciplined approaches to clinical laboratory measurement harmonization following Method 2 will be critical to the success of the ICHCLR initiative. 5. Summary and conclusions The laboratory medicine community is being challenged to better satisfy the needs of health care systems and patients by making improvements in clinical laboratory measurement systems that will enable the efficient use of all measurement results generated, regardless of when, where or how a particular laboratory measurement is performed. To ultimately achieve this goal, all values reported for the

Table 1 Achieving harmonization (equivalence) of measurement results by two methodologies. Attribute

Method 1

Method 2

Scheme

Hierarchical standardization per ISO17511:2003 [7]. Top down approach passing ‘trueness’ to lower order measurement procedures and calibrators.

Reference measurement procedures

One or more higher order reference measurement procedures available, preferably fulfilling requirements of ISO 15193:2009 [20] (http:// www.bipm.org/en/committees/jc/jctlm/) Certified purified reference materials and/or commutable secondary reference materials with associated uncertainties assigned to material value, preferably fulfilling requirements of ISO 15194:2009 [21] (http:// www.bipm.org/en/committees/jc/jctlm/). Non-commutable reference materials without proven commutability may hinder achieving equivalence of results. Commercial calibrators and reported results for routine measurement procedures traceable to SI unit via a metrological reference system (hierarchy of higher order reference measurement procedures and reference materials).

Inter-method comparison as described by ICHCLR Harmonization Initiative (www.harmonization.net) [11] Bottom up approach among routine (commercial) measurement procedures, with no SI traceability. None available.

Reference materials

Calibration traceability

5

Sustainability

Inbuilt sustainability via hierarchy of well-characterized and reproducible higher order and lower order reference measurement procedures and reference materials (http://www.bipm.org/en/committees/jc/jctlm/).

Harmonization goals

Equivalence of measurement results among different routine measurement procedures over time and space according to defined analytical and clinical performance goals.

No higher order reference materials available. Panel(s) of commutable human samples may be available, to be assigned consensus values and associated uncertainties (not traceable to SI) via harmonization studies (see www.harmonization.net) Some International Conventional Calibrators may be available (e.g. WHO materials), but usually not proven for commutability and may hinder achieving equivalence of results. Commercial calibrators and reported results of routine measurement procedures not traceable to SI unit. Traceability linked via inter-method comparison studies of available commercial measurement procedures coupled with mathematical recalibration for removal of systematic differences among reported values. Risk for non-sustainability of harmonized calibrations over time as routine methods and commercial calibrator lots change. Panels of patient samples used as “calibrators” in harmonization studies to be renewed over time (consumption and/or stability concerns.) Second & subsequent patient sample panels with values traceable to initial sample panel; presumes well-defined specifications for panel member selection. See ICHCLR Toolbox of Technical Procedures available at: http://harmonization.net/ Resource/Documents/Tool_Box_2013.pdf Equivalence of measurement results among different routine measurement procedures over time and space according to defined analytical and clinical performance goals.

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045

6

N. Greenberg / Clinica Chimica Acta xxx (2014) xxx–xxx

same measurand must be harmonized, even when multiple measurement systems are in use. The traditional metrological approach to standardization of calibrations for measurement procedures, with traceability of reported values to higher order references that constitute realizations of the SI unit, is well established and widely understood and accepted by the technical, regulatory and business communities. However, this idealized model has been successfully implemented in only a very limited way in the field of laboratory medicine due to various technical barriers and limitations inherent in the available measurement technology as well as in key characteristics at the molecular level of many of the measurands of interest to the laboratory medicine field. The use of calibration harmonization strategies that are based on non-traditional metrological schemes such as those proposed by the ICHCLR potentially offer practical solutions to address the reality that for the vast majority of measurands of clinical interest, standardization according to the traditional metrological principles of the SI system of units is simply not available for the foreseeable future. If the ICHCLR calibration approach is to be accepted as an effective solution for the establishment of harmonized calibrations for routine measurement procedures (including commercial kits as well as lab-developed measurement procedures) that are traceable to a single unified arbitrary unit of measurement, even if not an SI unit, it will still be necessary to (1) create objective and well-documented procedures for the determination of the appropriate corrections for each routine measurement procedure, and to (2) develop mechanisms to ensure the reproducibility and sustainability over time of the particular harmonization process being defined for each measurand. To achieve a sustainable measurement harmonization using Method 2 for a particular measurand, details of the full chain of relationships underlying the calibration of each routine (commercial) measurement procedure contributing data to the key method comparison/harmonization studies (including estimates for all significant sources of uncertainty) must be documented, published and maintained under careful change control. Additionally, it will be necessary to establish a mechanism for realization of the prototypical unit of measure underlying Method 2 — harmonized calibrations for measurands with no RMP, so that any laboratory or any manufacturer will have full and timely access, to enable establishment or validation of a new calibration to support development of novel technologies and new routine measurement procedures for these measurands. One approach that has been proposed for realization of the units of measure for those measurement systems harmonized according to Method 2 calls for the establishment of selected reference panels of patient samples [27]. While this approach has been demonstrated at the pilot laboratory scale, it is anticipated that development of a scaled-up version to support ongoing use by end-users including both the routine clinical laboratory and the in vitro diagnostics reagent kit manufacturing industry may be expensive and technically complex to sustain. Nevertheless, if the goal of establishing equivalence of reported values among different measurement procedures for the same measurand is to be achieved for a very large proportion of the measurands of interest to the laboratory medicine community, current and future harmonization projects undertaken according to the ICHCLR Method 2 harmonization model may be well worth the investment. Disclosure The author is convenor of ISO/TC 212/WG2, which is responsible for developing and maintaining International Standards related to the calibration of clinical laboratory measurement procedures. References [1] Floyd E, McShane TM. Development and use of biomarkers in oncology drug development. Toxicol Pathol 2004;32(Suppl. 1):106–15.

[2] Laboratory Standardization Panel of the National Cholesterol Education Program. Current status of blood cholesterol measurement in clinical laboratories in the unitEd States: a report of the laboratory standardization panel of the national cholesterol education program. Clin Chem 1988;34:193–201. [3] Myers GL, Kimberly MM, Waymack PP, Smith SJ, Cooper GR, Sampson EJ. A reference method laboratory network for cholesterol: a model for standardization and improvement of clinical laboratory measurements. Clin Chem 2000;46:1762–72. [4] Morbidity & mortality: 2012 chart book on cardiovascular, lung and blood diseases. National institutes of health. National Heart, Lung and Blood Institute; 2012 Feb 33 [Chart 3–26, Available from: http://www.nhlbi.nih.gov/resources/docs/2012_ ChartBook.pdf [Accessed 19MAY2013.]]. [5] Hoerger TJ, Wittenborn JS, Young W. A cost-benefit analysis of lipid standardization in the United States. Prev Chronic Dis 2011;8:A136 [Available from: http://www. cdc.gov/pcd/issues/2011/nov/10_0253.htm [Accessed 10JUN2013]]. [6] Directive 98/79/EC Of the European Parliament and of the council of 27 October 1998 on in vitro diagnostic medical devices, OJ; Dec 7 1998 [331:1–42]. [7] International Organization for Standardization. ISO 17511:2003. In vitro diagnostic medical devices – measurement of quantities in biological samples – metrological traceability of values assigned to calibrators and control materials. Geneva: International Organization for Standardization; 2003. [8] United States Code of Federal Regulations. 21CFR 809.10 Available from http://www. fda.gov/MedicalDevices/DeviceRegulationandGuidance/Overview/DeviceLabeling/ InVitroDiagnosticDeviceLabelingRequirements/default.htm. [Accessed 10JUN2013]. [9] Gantzer ML, Miller WG. Harmonisation of measurement procedures: how do we get it done? Clin Biochem Rev 2012;33:95–100. [10] Miller WG, Myers GL, Gantzer ML, Kahn SE, Schönbrunner ER, Thienpont LM, et al. Roadmap for harmonization of clinical laboratory measurement procedures. Clin Chem 2011;57:1108–17. [11] American Association for Clinical Chemistry. Clinical Laboratory Test Harmonization – Harmonization.net. Available from: http://www.harmonization.net/Pages/default.aspx. [Accessed 07JUN2013.] [12] Joint Committee for Guides in Metrology. Working Group 2 of the Joint Committee for Guides in Metrology (JCGM/WG 2). JCGM 200:2008, International Vocabulary of Metrology — Basic and general concepts and associated terms (VIM). Available from http://www.iso.org/sites/JCGM/VIM-introduction.htm; 2008. [13] Physical Measurements Laboratory of NIST. The NIST Reference on Constants, Units, and Uncertainty. Available from http://physics.nist.gov/cuu/Units/units.html. [Accessed 16SEP2013]. [14] International Laboratory Accreditation Cooperation. Policy on traceability of measurement results — ILAC P-10. 7 Leeds Street, Rhodes, NSW, Australia 2138: International Laboratory Accreditation Cooperation; 2002. [15] Clinical and Laboratory Standards Institute (CLSI). Characterization and qualification of commutable reference materials for laboratory medicine; approved guideline. CLSI document EP30-A (ISBN 1-56238-726-X). 950 West Valley Road, Suite 2500, Wayne, Pennsylvania 19087 USA: Clinical and Laboratory Standards Institute; 2010. [16] International Organization for Standardization. Guide to the expression of uncertainty in measurement (GUM), ISO/IEC Guide 98-3. Geneva: International Organization for Standardization; 2008. [17] Clinical and Laboratory Standards Institute (CLSI). Expression of measurement uncertainty in laboratory medicine; approved guideline. CLSI document EP29-A. 950 West Valley Road, Suite 2500, Wayne, Pennsylvania 19087 USA: Clinical and Laboratory Standards Institute1-56238-787-1; 2012. [18] Eurachem/CITAC Guide CG4: quantifying uncertainty in analytical measurementIn: Ellison SLR, Williams A, editors. 3rd ed. ISBN 978-0-948926-30-3; 2012 [Available from: http://www.eurachem.org.]. [19] JCTLM: Joint Committee for Traceability in Laboratory Medicine. Available from http://www.bipm.org/en/committees/jc/jctlm/. [Accessed 10JUN2013]. [20] International Organization for Standardization. ISO 15193:2009 In vitro diagnostic medical devices – measurement of quantities in samples of biological origin – requirements for content and presentation of reference measurement procedures. Geneva: International Organization for Standardization; 2009. [21] International Organization for Standardization. ISO 15194:2009 In vitro diagnostic medical devices – measurement of quantities in samples of biological origin – requirements for certified reference materials and the content of supporting documentation. Geneva: International Organization for Standardization; 2009. [22] International Organization for Standardization. ISO 15195:2003 laboratory medicine — requirements for reference measurement laboratories. Geneva: International Organization for Standardization; 2003. [23] WHO International Biological Reference Preparations. Available from http://www. who.int/bloodproducts/catalogue/SafeFeb2013.pdf. [Accessed 21SEP2013]. [24] Donadio S, Morelle W, Pascual A, Romi-Lebrun R, Michalski JC, Ronin C. Both core and terminal glycosylation alter epitope expression in thyrotropin and introduce discordances in hormone measurements. Clin Chem Lab Med 2005;43:519–30. [25] Sturgeon CM, Berger P, Bidart JM, Birken S, Burns C, Norman RJ, et al. Differences in recognition of the 1st WHO international reference reagents for hCG-related isoforms by diagnostic immunoassays for human chorionic gonadotropin. Clin Chem 2009;55:1484–91. [26] Oxford Online Dictionary of American English. Available from http://oxforddictionaries. com/us/definition/american_english/harmonize. [accessed 17SEP2013]. [27] Thienpont LM, Van Uytfanghe K, Beastall G, Faix JD, Ieiri T, Miller WG, et al. Report of the IFCC working group for standardization of thyroid function tests, part 1: thyroidstimulating hormone. Clin Chem 2010;56:902–11. [28] Van Houcke SK, Van Aelst S, Van Uytfanghe K, Thienpont LM. Harmonization of immunoassays to the all-procedure trimmed mean — proof of concept by use of data from the insulin standardization project. Clin Chem Lab Med May 2013;51(5):e103–5. http://dx.doi.org/10.1515/cclm-2012-0661.

Please cite this article as: Greenberg N, Update on current concepts and meanings in laboratory medicine — Standardization, traceability and harmonization, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2013.12.045