Interlaboratory certification study

This flexible approach to certification has also led us to take a more holistic view of RM production. Interlaboratory certification studies frequently highlight significant measurement problems, sometimes in a large number of participating laboratories. The tendency has been to make participants aware of this situation through the certification report but to take little other action. In future we aim to provide better value for money to both... 

It has frequently been suggested that the organisers of interlaboratory certification studies should only accept reliable laboratories as participants and one obvious selection criterion would be to use only accredited laboratories. However, an evaluation of data obtained from several studies organised by LGC over the past 7 years indicates that accredited laboratories are not necessarily more reliable than nonaccredited laboratories. For example, Fig. 1 shows that in an interlaboratory study to determine magnesium in water, accredited laboratories (marked with an asterisk) were just as likely to produce results with a large deviations as nonaccredited laboratories. 

UV). Consequently the reliability of the signal relies on the prior chromatographic separation. Matrix matching of the calibration solution or standard addition procedure can help to overcome matrix influences. An example of the too often neglected matrix effect is given in Table 2.1. These results were obtained in an interlaboratory certification study of trace elements in soils. It demonstrates the importance of matrix matching for the calibration [25]. Except for Co and Cu, for which another source of error remained, the matrix matching of the calibrant solutions allowed the laboratory to come closer to the certified values. 

Value of g the required precision value a can be found in the certificate of the CRM itself a can be the uncertainty of the certified value when the same method as the certification method is used by the analyst. It can also come from the individual set of measurement values of one of the methods used in the interlaboratory certification study. In such a case, all individual data and the methods must be available from the certificate or the certification report. Such information is given in some CRM reports of BCR and is illustrated in Annex 3.1. The a can also be stated in a written standard as a minimal or target precision value to be obtained. Finally, a can also simply come from the laboratory itself which applied another method previously or with another instrument or from another laboratory experienced with the method, or it may be requested by a customer. Care must be taken in extrapolating a simply from another element or substance which is analysed in parallel using a multi-elemental or multiresidue methods, as these may not be comparable at all because of matrix effects etc. 

When having all individual data available from the certification report, the analyst may wish to have a closer insight into his performance in terms of comparison to the results achieved by a similar method used in the certification. In the case of materials certified for organic trace substances e.g. dioxins in fly ashes, where methods inevitably remain affected by small systematic errors, the uncertainty calculated from the set of data of the interlaboratory certification study may appear small compared to the spread of the accepted sets of data. The half-width of the 95% confidence interval of the set of data is used as uncertainty. The user can adapt the way of determining the value of Sq above. He may wish to replace the 95% confidence interval by a tolerance interval... 

The distribution of a limited number of microbes within a solid matrix is essentially inhomogeneous, i.e. for a certain level of intake, differences due to the distribution of the microbes will appear. The more microbes that have been introduced into the matrix, the less these differences will be important and the lower the sample intake will be where differences appear. The theoretical distribution of the microbes in the sample is described by a Poisson distribution [29,30]. The RIVM group, which developed these materials, used a modified Cochran s dispersion test to evaluate the variation within a single capsule and between capsules. In fact overdispersion (more inhomogeneity than theoretically expected) has been noticed for nearly all RMs and CRMs produced [31]. The certification trial revealed that it had no influence on the outcome of the interlaboratory certification study as the between-laboratory reproducibility largely covered this overdispersion factor. 

Attempts to correlate analytical performance with other seemingly indicative laboratory characteristics, such as participation in proficiency testing schemes, regular use of certified RMs, number of years of experience and number of samples analysed per year were all equally unsuccessful. Therefore, in the absence of any simple and obvious means of identifying and preselecting only reliable laboratories as participants in certification studies, an investigation was undertaken of the validity of adopting the consensus mean (after outlier elimination) from an interlaboratory study as a certified value. 

Material certification study - an interlaboratory study that assigns a reference value ( true value ) to a quantity (concentration or property) in the test material, usually with a stated uncertainty/ ... 

External quality assurance systems (EQAS) refer to systems for objectively assessing and demonstrating the reliability of laboratory results by means of interlaboratoiy comparison. Interlaboratory comparison can be performed as collaborative, prcficiency, or certification studies [7]. The ISO 5725 series of documents provide practical descriptions of methods for estimating precision and trueness by means of collaborative interlaboratory experiments [5]. 

A certification study is an interlaboratory study in which a group of selected laboratories analyze a candidate reference material by a method judged most likely to provide the least biased estimates of concentration and the smallest associated uncertainty. The purpose is to provide a reference value of the analyte concentration in the material. Certification studies are organized by, e.g., BCR and NIST [7]. 

Principles and Characteristics According to the lUPAC definition, an interlaboratory study is one in which several laboratories measure a quantity in one or more identical portions of homogeneous materials under documented conditions, the results of which are compiled into a single report. Three types of interlaboratory studies are distinguished, namely method-performance, laboratory-performance or material-certification studies. The aim of method-performance or collaborative studies is to assess the performance characteristics of a specific method. In laboratory-performance or proficiency studies a homogeneous test material is analysed of which the true concentrations are known or have been assigned in some way. The participants apply whatever... 

Fig. 3.4 Results from the interlaboratory study used for certification of benzo[b]fluoranthene in BCR CRM 088 Sewage Sludge...

The enormous difference in certified values between methods and between analytes illustrates well how much care is needed in matrix/method matching. Further evidence of the importance of matrix matching is provided by an interlaboratory study on trace elements in soil reported by Maier et al. (1983) and the certification of a sewage sludge described by Maaskant et al. (1998). 

Maier, E A., Quevauviller, Ph, and Griepink, (1993), Interlaboratory studies as a tool for many purposes proficiency testing, learning exercises, quality control and certification of matrix materials. Analytica Chimica Acta, 283 (1), 590-99. 

Improvement schemes can be defined as a succession of individual interlaboratory studies in which several laboratories analyse the same test samples for the same characteristics (usually the content of an analyte), following a similar protocol to validate each individual step of their own analytical method (Quevauviller, 1999a). They enable laboratories to develop and validate all steps of new or existing analytical procedure(s) in adequately organised successive exercises which may be considered as preliminary studies for laboratory or method performance studies or certification of RMs (Griepink and Stoeppler, 1992 Quevauviller, 1998b). Such programmes are particularly valuable in the case of speciation studies since the analytical procedures include several complex and critical steps. 

Statistics should follow the technical scrutiny, not the other way round. A statistical analysis of data of an interlaboratory study cannot explain deviating results nor can alone give information on the accuracy of the results. Statistics only treat a population of data and provide information on the statistical characteristics of this population. The results of the statistical treatment may give rise to discussions on particular data not belonging to the rest of the population, but outlying data can sometimes be closer to the true value than the bulk of the population (Griepink et al., 1993). If no systematic errors affect the population of data, various statistical tests may be applied to the results, which can be treated either as individual data or as means of laboratory means. When different methods are applied, the statistical treatment is usually based on the mean values of replicate determinations. Examples of statistical tests used for certification purposes are described elsewhere (Horwitz, 1991). Together with the technical evaluation of the results, the statistical evaluation forms the basis for the conclusions to be drawn and the possible actions to be taken. 

Following a feasibility study, it was found that trimethyllead (TML) would be of sufficient stability for the purpose of interlaboratory studies and possibly certifications of RMs whereas dialkyllead species and triethyllead were not stable enough. Interlaboratory studies were performed in 1992-1994 on TML in artificial rainwater and urban dust and enabled a significant improvement of the state-of-the-art to be achieved (Quevauviller et al., 1995b). These exercises were followed by a certification campaign aiming to certify TML in candidate RMs of artificial rainwater and urban dust. The between-laboratory agreement obtained was very satisfactory and both materials were proposed for certification. 

The motivation and dedication of the laboratories participating in the various interlaboratory studies and certifications organised by BCR represent one of the keys to the success of the projects with a clear impact on the improvement of quality control of speciation measurements. All project coordinators and participants in the projects are gratefully acknowledged and are listed in the RSC book on speciation (Quevauviller, 1998b). 

Recently, a group of experts working under the auspices of the Commission of the European Communities, Community Bureau of Reference (BCR) sought to improve comparability between sequential extraction results obtained by different laboratories, and proposed a simple, three-stage sequential extraction procedure for sediment (Table 10.3) (Ure et d., 1993a, b). The protocol was then refined through two sets of interlaboratory trials (Quevauviller et d., 1994) and studies were conducted to assess whether adequate homogeneity and stability could be achieved to allow certification of a reference material for metals extractable by the procedure (Fiedler et d., 1994). 

In view of these issues, we have over several years undertaken a substantial research programme to develop definitive methods appropriate for in-house certification of matrix RMs, particularly for analytes at trace levels. These definitive measurement methods, most of which use isotope dilution mass spectrometry (IDMS), been the subject of extensive validation, including CCQM key comparisons and pilot studies involving other national measurement institutes. Hence, we are now able to augment interlaboratory data with data obtained at LGC using these very accurate measurements. We have also... 

---