Netherlands standard NEN 7777+C1. (en) Environment and food - Performance characteristics of measurement methods

Transcription

1 Dit document mag slechts op een stand-alone PC worden geinstalleerd. Gebruik op een netwerk is alleen. toestaan als een aanvullende licentieovereenkomst voor netwerkgebruik met NEN is afgesloten. This document may only be used on a stand-alone PC. Use in a network is only permitted when a supplementary license agreement for us in a network with NEN has been concluded. Netherlands standard NEN 7777+C1 (en) Environment and food - Performance characteristics of measurement methods Milieu en voedingsmiddelen - Prestatiekenmerken van meetmethoden Supersedes NEN 7777:2011 ICS ; ; ; december 2012

2 Normcommissie "Normcommissie Milieukwaliteit" THIS PUBLICATION IS COPYRIGHT PROTECTED DEZE PUBLICATIE IS AUTEURSRECHTELIJK BESCHERMD Apart from exceptions provided by the law, nothing from this publication may be duplicated and/or published by means of photocopy, microfilm, storage in computer files or otherwise, which also applies to full or partial processing, without the written consent of the Netherlands Standardization Institute. The Netherlands Standardization Institute shall, with the exclusion of any other beneficiary, collect payments owed by third parties for duplication and/or act in and out of law, where this authority is not transferred or falls by right to the Reproduction Rights Foundation. Auteursrecht voorbehouden. Behoudens uitzondering door de wet gesteld mag zonder schriftelijke toestemming van het Nederlands Normalisatie-instituut niets uit deze uitgave worden verveelvoudigd en/of openbaar gemaakt door middel van fotokopie, microfilm, opslag in computerbestanden of anderszins, hetgeen ook van toepassing is op gehele of gedeeltelijke bewerking. Het Nederlands Normalisatie-instituut is met uitsluiting van ieder ander gerechtigd de door derden verschuldigde vergoedingen voor verveelvoudiging te innen en/of daartoe in en buiten rechte op te treden, voor zover deze bevoegdheid niet is overgedragen c.q. rechtens toekomt aan de Stichting Reprorecht. Although the utmost care has been taken with this publication, errors and omissions cannot be entirely excluded. The Netherlands Standardization Institute and/or the members of the committees therefore accept no liability, not even for direct or indirect damage, occurring due to or in relation with the application of publications issued by the Netherlands Standardization Institute. Hoewel bij deze uitgave de uiterste zorg is nagestreefd, kunnen fouten en onvolledigheden niet geheel worden uitgesloten. Het Nederlands Normalisatie-instituut en/of de leden van de commissies aanvaarden derhalve geen enkele aansprakelijkheid, ook niet voor directe of indirecte schade, ontstaan door of verband houdend met toepassing van door het Nederlands Normalisatie-instituut gepubliceerde uitgaven Nederlands Normalisatie-instituut Postbus 5059, 2600 GB Delft Telefoon (015) , Fax (015)

3 Table of Contents Foreword Scope Normative references... 4 Terms and definitions... 4 Symbols and abbreviations Procedure for intralaboratory validation Validation plan Validation test Combination procedure for performance characteristics Reporting Annex A (normative) Types of testing Annex B (informative) The deduction of requirements from the values of performance characteristics Annex C (normative) Critical values for testing variability characteristics (for use see clause 6) Annex D (informative) Procedure for intralaboratory reproducibility if the number of (re)calibrations is less than the number of validation days Annex E (informative) Detection limit, decision limit, reporting limit and limit of detection Annex F (informative) Determination of between-sample variability Annex G (informative) Schematic example for the combined determination of performance characteristics (statistically independent) Bibliography

4 Foreword Measured quantity values are only approximations of the true value. The extent to which measured quantity values may deviate is quantified using performance characteristics. They reflect the performance of the measurement method under various conditions. Performance characteristics are used: by the individual laboratory in method choice, method development, validation and operational quality control; in interlaboratory tests to compare the performance of various laboratories with one another; in the standardization process to indicate the quality and range of a standardized method; in accreditation to demonstrate the quality of the measurement to third parties. The standard is intended to be used in the validation of a measurement method for physical or chemical quantities within one s own laboratory. The values obtained are therefore only valid for the laboratory in question. The requirement is to record the method in a measurement procedure or measurement protocol. The standard is also intended to test whether a measurement method meets requirements for performance characteristics. A factor here is that in current practice the status of requirements is not always clear: does it have to be demonstrated with great reliability that a performance characteristic lies on the right side of an absolute limit value, or is it sufficient if a value found for a performance characteristic does not deviate significantly from an externally given standard value? When using this standard the status of requirements for performance characteristics shall be made explicit and the test shall be adapted to this. The standard gives: definitions of the generally used performance characteristics; procedures for establishing these performance characteristics; procedures for testing the values obtained against reference values such as standard and limit values. Procedures for performance characteristics of a method that are valid for a group of laboratories are given in the various parts of ISO The determination of such performance characteristics requires interlaboratory testing. Revisions with respect to NEN 7777:2003. The scope has been extended to food (see title, Clause 1). The definitions have been revised in accordance with NPR-ISO/IEC GUIDE 99 (see Clause 2). The validation plan has a more prominent place (see Clauses 5 and 6). The representativeness of the performance characteristic is more strongly emphasised than in NEN 7777:2003. This is expressed in the choice of validation samples and validation procedures. Two forms of representativeness are distinguished (see 6.2): representativeness for the collection of routine samples (the laboratory performance for the laboratory s own samples); representativeness for an imposed sample (the laboratory performance for an assigned sample). The concept most difficult matrix is explicitly discussed (see 6.3.2). The minimum number of validation levels has been made more flexible. The assumption is that the measurement range to be validated is incorporated in one (sub-)scope (see 6.3.3) where possible. 2

5 In addition to imposed requirements attention is also paid to deduced requirements (see 6.6.2). The deduction of requirements is explicitly dealt with in informative Annex B. Detection based on reproducibility is a guiding factor. Detection is now only based on repeatability if there is a historic need (see 7.2.1). The reporting limit has been added (see 7.2.2). Repeatability and intralaboratory reproducibility have also been introduced as quantitative terms. Both performance characteristics are used in particular for food testing (see 7.3 and 7.4 respectively). Instead of the term ' trueness' the term 'bias' is used to estimate the quantitative term 'systematic error'. The term trueness is maintained as a qualitative term. This approach corresponds to NPR-ISO/IEC Guide 99. In addition to the existing determination of the bias for an individual sample, the determination of the bias for the whole collection of own routine samples has been added. Because the bias differs from sample to sample, the bias that all the samples have is the same as the mean of the biases of individual samples. Between-sample variability has been introduced as a term for the variability of biases (see 7.5). Determining the between-sample variability is included in informative Annex F and is therefore not specified as a performance requirement in this standard. The performance characteristic 'memory effect' has been added (7.9). The schedule for combined execution of validation experiments has been revised to ensure better independence of the observations (see clause 7). A more extensive schedule, to guarantee statistical independence for reproducibility, is included in an informative Annex (see Annex G). 3

6 Environment and food Performance characteristics of measurement methods 1 Scope This standard describes procedures: to establish performance characteristics of a measurement method for testing food and the environment; to test values of performance characteristics against requirements (standard values or limit values) relating to the intended use. The standard is intended to be used for intralaboratory validation of measurement methods for physical and chemical quantities and requires the method to be laid down in a measurement procedure or measurement protocol. For validation of the measurement method by a group of laboratories, please refer to the various parts of ISO NOTE 1 The quality evaluation of the measurement method is not limited to the validation that precedes it being taken into use. The quality of components of the measurement method or alternative methods is already established in the development of a measurement method as a basis for optimization and acceptance. Also while a measurement method is in operation, critical performance characteristics are used to check regularly that the initial performance is maintained (quality assurance). Compared with this standard similar, abbreviated or simplified procedures are used. NOTE 2 This standard is not normative for multicomponent methods. It may for example be decided to carry out validation of some of the collection of components (critical or representative components) according to this standard. NOTE 3 Certain (international) regulations, for example EU Directives, specify that the performance of measurement methods should be determined in another way. Depending on the objective of the test it may be decided to follow these regulations. In this case the regulations applied should be reported giving brief reasons for this choice. 2 Normative references There are no normative references to other documents in this standard. 3 Terms and definitions For the purpose of this standard the following terms and definitions apply. 3.1 detection limit measured quantity value, obtained by a given measurement procedure, for which the probability of falsely claiming the absence of a component in a material is, given a probability of falsely claiming its presence NOTE 1 Entering the value of 0,05 recommended by IUPAC for both and leads to a detection limit of 3,3 times the standard deviation at this level where there are a large number of repetitions; for the minimum of eight repetitions this factor is 3,8. NOTE 2 This standard uses the operational definition that equates the detection limit with three times the standard deviation at this level. As a result the detection limit is then that value of the measurand at which the coefficient of variation is by convention a third. NOTE 3 For a further discussion of detection limit and related terms such as decision limit, reporting limit and limit of detection, see Annex E. 4

7 3.2 absolute limit value value of a performance characteristic that is intended as an absolute requirement NOTE It should be demonstrated by statistical testing that the performance characteristic complies with this. 3.3 non linearity deviation from the assumed linear relationship between measurand and measurement indication 3.4 test sample sample that is prepared from the laboratory sample according to the measurement procedure NOTE Test samples play a part in validation if the laboratory sample is the measurement object and several test samples have to be obtained from the laboratory sample (see 5.5.2). EXAMPLE 1 The laboratory soil sample is first dried, ground and divided at the start of the measurement procedure. The test sample is the part that is used for the continuation. EXAMPLE 2 From the laboratory soil sample ten sub-samples are taken according to the measurement procedure to determine volatile substances. The sub-samples are combined into the test sample. 3.5 bias measurement bias estimate of a systematic measurement error [NPR-ISO/IEC Guide 99, modified] NOTE 1 Bias is a practical measurement for systematic error (3.30). NOTE 2 The reference value may be a conventional value. NOTE 3 Large number relates to samples if the bias of the (sub-)scope is meant and to measurement results for one sample if the bias of this sample is meant. 3.6 conventional value (of a quantity) quantity value attributed by agreement to a quantity for a given purpose EXAMPLE 1 The consensus value of a group of expert laboratories. EXAMPLE 2 The consensus value of the measurement results of all the participants in interlaboratory testing: often the (robust) mean or the median. 3.7 requirement (for a performance characteristic) requirement because of the intended use or imposed by a relevant body in the form of a value of a performance characteristic NOTE 1 The requirement should be further described, or be interpreted as either an absolute limit value or an estimated limit value, or a standard value (typical value). NOTE 2 For most performance characteristics (e.g. standard deviations, detection limit) a better performance is characterized by a lower value. Requirements then relate to upper limits. In some cases (e.g. recovery) it is primarily the lower limits that are important. If both lower and upper limits are important (e.g. bias), the requirement should be formulated as two absolute limit values, or two estimated limit values or one standard value. 5

8 3.8 memory effect carry-over effect on present measurement result of previous values of the measurand NOTE A memory effect is expressed as a difference between the previous and the present value of the measurand. 3.9 estimated limit value value of a performance characteristic, obtained from a previous validation test or from other types of estimates, intended to be a limit value for the performance characteristic in question NOTE 1 Estimates of performance characteristics obtained on testing are compared directly (that is without statistical testing) with the estimated limit value. NOTE 2 If the limit is exceeded (in the case of an upper limit) or not reached (in the case of a lower limit) the effectiveness of the measurement method is not directly called into question repeatability (of measurement results) measurement repeatability degree of agreement between indications or measurement results obtained by successive measurements of the same object or similar objects under repeatability conditions [taken from NPR-ISO/IEC Guide 99, modified] NOTE 1 Repeatability conditions include: the same measurement method; the same observer; the same measurement instrument; the same place and conditions; repetition over a short time span. NOTE 2 Repeatability is in practice quantified as lack of repeatability. NOTE 3 Repeatability in food testing is also used as a quantitative measurement. The definition is: 95 % confidence interval of the difference between two measurement results obtained under repeatability conditions. To avoid confusion with repeatability as an umbrella term, the term is indicated by repeatability r influence quantity quantity that, in a direct measurement, does not affect the quantity that is actually measured, but affects the relation between the indication and the measurement result EXAMPLE 1 EXAMPLE 2 EXAMPLE 3 EXAMPLE 4 EXAMPLE 5 Content of minerals in the determination of elements in water with ICP techniques. Concentration of a poorly separated component in the chromatographic analysis. Concentration of carbonate in the polarimetric determination of starch in cattle feeds. Content of particles in spectrophotometric analysis. Temperature of the testing environment when performing a leaching test. 6

9 3.12 trueness measurement trueness trueness of measurement closeness of agreement between the average of an infinite number of replicate measured quantity values and a reference quantity value NOTE 1 Trueness is not a quantity and cannot therefore be expressed numerically. NOTE 2 Systematic error (3.30) is the quantitative quantity for the qualitative term trueness. In practice bias (3.5) is used to estimate systematic error calibration curve expression of the relation between indication and corresponding measured quantity value EXAMPLE NOTE 1 The peak area as a function of the concentration in chromatographic analyses. The mathematical relationship is in practice often approximated by a regression function. NOTE 2 The calibration in principle includes the whole measurement method. If (for practical reasons) part of the measurement method cannot be included in the calibration the quantity referred to in the definition is "an auxiliary quantity related to the measurand" (e.g. the "concentration in the extraction liquid" as an auxiliary quantity for the measurand "concentration in soil") [9] laboratory sample amount of material intended for testing in the form and state in which it is delivered to the laboratory NOTE The laboratory sample is obtained from the measurement object. If the instructions are to make a statement on the value of the measurand in the laboratory sample, the laboratory sample is the measurement object measuring range measuring interval set of values of quantities of the same kind that can be measured by a given measuring instrument or measuring system with specified instrumental uncertainty, under defined conditions 3.16 measurand quantity intended to be measured EXAMPLE 1 EXAMPLE 2 EXAMPLE 3 Concentration of nitrate in salad. Rate of deposition of particles in air. Dry matter content of soil. NOTE For the specification of a measurand the indication of quantities such as time, temperature and pressure can be demanded. 7

10 3.17 measurement method generic description of a logical organization of operations used in a measurement EXAMPLE An example of a measurement method is the classic measurement standard. The principle of measurement is developed into a description of the technical execution including equipment and tools, but often still leaves a lot of room for (equivalent) interpretation. See also measurement procedure (3.21) measurement object object, defined in nature, size and state, for which the value of the measurand applies EXAMPLE 1 EXAMPLE 2 EXAMPLE 3 1 kg soil, stored in a light- and air-tight packaging, a maximum of 1 year after sampling. 0.1 m 3 outside air, to be aspirated during 30 min at a height of 3 m in an urban area. Ship s cargo of grain. The laboratory sample is a random sample from the measurement object uncertainty of measurement non-negative parameter characterizing the dispersion of the quantity values being attributed to a measurand, based on the information used NOTE 1 The parameter may for example be a standard deviation (or a certain multiple thereof), or half of an interval that has a certain confidence level. NOTE 2 The quantification of uncertainty of measurement is not part of this standard. See NEN 7779 for this principle of measurement phenomenon serving as a basis of a measurement EXAMPLE 1 EXAMPLE 2 EXAMPLE 3 Distribution over two phases applied to the chromatographic separation of compounds. Radiation absorption applied to the determination of light absorbing substances. X-ray fluorescence applied to the determination of elements measurement procedure detailed description of a measurement according to one or more measurement principles and to a given measurement method, based on a measurement model and including any calculation to obtain a measurement result NOTE 1 Measurement model in the definition means the scientific relationship between all the quantities which are known to play a part in the measurement. NOTE 2 The following terms are also used for measurement procedure: (measurement) instructions; analysis instructions; work instructions; SOP (Standard Operating Procedure). 8

11 3.22 measurement result value obtained by measurement, assigned to a measurand NOTE 1 The measurement result may be defined as the mean of a multiple analysis or may be the result of a measurement followed by one or more operations such as correction for procedure blank. NOTE 2 A full indication of the measurement result includes information on the uncertainty of measurement indication quantity value provided by a measuring instrument or a measuring system EXAMPLE 1 The peak area of a chromatographic system. EXAMPLE 2 The number of counts of a mass spectrometer detector lack of fit error in the calibration model error in the assumed relationship between indication and corresponding measurement result NOTE 1 As the number of measurements to determine the lack of fit increases, smaller errors may be determined (variability of the mean measurement result decreases with the number of measurement results). NOTE 2 There will always be an error in the assumed relationship. Sometimes this error is so small that it cannot be determined experimentally due to the inherent variability of measurement results reporting limit lowest value of the measurand that is reported quantitatively NOTE The reporting limit is a minimum of the detection limit reproducibility (of measurement results) measurement reproducibility closeness of agreement between indications or measurement results obtained by of successive measurements on the same or similar objects under reproducibility conditions [taken from NPR-ISO/IEC Guide 99, modified] NOTE 1 Reproducibility conditions are varying conditions that may include: measurement systems (see note 2); the observer; the measurement instrument; the reference standard; the place; the conditions of use. NOTE 2 Reproducibility in this standard relates to the measurement method; the principle of measurement and the measurement method are excluded as a varying condition. NOTE 3 It is important to make a distinction between intralaboratory reproducibility and interlaboratory reproducibility. The varying conditions in these cases relate to varying conditions within the laboratory or within a group of laboratories. If the group of laboratories is representative of the whole, interlaboratory reproducibility of the method is the same as reproducibility of the method. 9

12 NOTE 4 Intralaboratory reproducibility and reproducibility are umbrella terms. In food testing the term reproducibility, indicated as R, is also used as a quantitative measurement. The corresponding quantitative measurement intralaboratory reproducibility is indicated here by Rw and Rw rel. To avoid confusion with the umbrella term the quantitative term is indicated by intralaboratory reproducibility Rw and intralaboratory reproducibility Rw rel robustness degree of insensitivity of the measurement result for errors in execution, conditions and quality of materials, as these may occur in practice NOTE Errors that represent a(n influence) quantity (3.11) may be quantified separately. Examples of such quantities are: temperature of the testing environment, humidity, light intensity, mains voltage and the chemical content selectivity dependence of the measurement result on a quantity other than the measurand NOTE 1 Quantities referred to are influence quantities (3.11) that only relate here to the composition of the sample. NOTE 2 The dependence of the measurement result not only means the quantity that itself causes an indication (often called interference) but also the quantity that changes the indication of the measurand (often called matrix effect). NOTE 3 The measurement method is the more selective the smaller the effect of influence quantities on the measurement result. NOTE 4 Specificity is a term that describes the same phenomenon in a different way standard value typical value value of a performance characteristic, obtained in a previous validation test that aims to give a pure estimate of the performance characteristic in question NOTE Estimates of performance characteristics obtained in tests should not deviate significantly in an unfavourable direction from the standard value systematic error component of measurement error that in replicate measurements remains constant or varies in a predictable manner NOTE The systematic error is estimated with the quantity bias (3.5) recovery fraction of the measurement component that is recovered upon analysis, after adding a known amount of measurement component to the sample under defined conditions [9] NOTE 1 Recovery determined from addition of the measurement component gives (too) optimistic results if it is difficult or impossible to incorporate the added part of the measurement component in the sample in the same way as in the original part. NOTE 2 In the absence of a sample with a conventional value, recovery is in many cases used as a measurement to help quantify the bias. Recovery does not cover bias due to interference and non linearity. NOTE 3 Sometimes the performance characteristic recovery is also taken to compare the measurement result with the conventional value of a sample. In this standard only the performance characteristic bias is used for this (see also note 2). 10

13 Bestelformulier Stuur naar: NEN Standards Products & Services t.a.v. afdeling Klantenservice Antwoordnummer WB Delft NEN Standards Products & Services Postbus GB Delft Vlinderweg AX Delft Ja, ik bestel ex. NEN 7777+C1:2012 en Milieu en voedingsmiddelen - Prestatiekenmerken van meetmethoden T (015) F (015) Wilt u deze norm in PDF-formaat? Deze bestelt u eenvoudig via Gratis nieuwsbrieven Wilt u op de hoogte blijven van de laatste ontwikkelingen op het gebied van normen, normalisatie en regelgeving? Neem dan een gratis abonnement op een van onze nieuwsbrieven. Gegevens Bedrijf / Instelling T.a.v. O M O V Klantnummer NEN Uw ordernummer BTW nummer Postbus / Adres Postcode Plaats Telefoon Fax Factuuradres (indien dit afwijkt van bovenstaand adres) Postbus / Adres Postcode Plaats Datum Handtekening Retourneren Fax: klantenservice@nen.nl Post: NEN Standards Products & Services, t.a.v. afdeling Klantenservice Antwoordnummer 10214, 2600 WB Delft (geen postzegel nodig). Voorwaarden De prijzen zijn geldig tot 31 december 2018, tenzij anders aangegeven. Alle prijzen zijn excl. btw, verzend- en handelingskosten en onder voorbehoud bij o.m. ISO- en IEC-normen. Bestelt u via de normshop een pdf, dan betaalt u geen handeling en verzendkosten. Meer informatie: telefoon , dagelijks van 8.30 tot uur. Wijzigingen en typefouten in teksten en prijsinformatie voorbehouden. U kunt onze algemene voorwaarden terugvinden op: LEREN, WERKEN EN GROEIEN MET NEN preview