Interlaboratory trials

For trade issues, proficiency testing could also respond to the need for improving measurement quality. The EC should stimulate the establishment of links between various proficiency testing schemes in order to make the best use of them. 

The problem of too high costs of CRMs was also debated. It was stressed that nobody complains about the high costs of e.g. suprapure acids, which are mandatory for trace analysis. The perception of high costs hence seems to be related to a problem of culture and education. Unless the education background is established, the awareness will remain at a very low level and the market will not develop. Information should appear in all academic training teachers need also to be trained  

Many laboratories do not use reference materials because they are not forced to do so which imply that the users of the data should primarily be briefed on the importance of CRMs as a tool for the quality of data. 

A systematic collection of information from different research groups developing in-house RMs for their own needs would be of considerable value for many analytical sectors. This collection could be achieved by networking institutes. Participants stressed that a lot of information is available on attempts to certify CRMs, which could be of considerable help for the preparation of LRMs. 

Such types of structure could also offer facilities for laboratories to test their procedures on-site in the case where CRMs cannot be produced, e.g. reference sites  

---

Butler OT, Howe AM (1999) Development of an international standard for the determination of metals and metalloids in workplace air using ICP-AES evaluation of sample dissolution procedures through an interlaboratory trial, f Environ Monit 1 23-32. 

Reproducibility in the context of Directive 96/46/EC is defined as a validation of the repeatability of recovery, from representative matrices at representative levels, by at least one laboratory, which is independent of the laboratory which initially validated the study. This independent laboratory may be within the same company, but may not be involved in the development of the method. This concept of independent laboratory validation (ILV) substitutes the conduct of interlaboratory trials (e.g., according to ISO 5725) because the resources are not available taking into consideration the high number of a.i., matrix types and concentration levels which must be validated in the registration procedure. 

Within CEN TC 275, a distinct protocol for the conduct of inter-laboratory method validation trials is not selected or required, but it is recommended to apply ISO 5725 or the ISO/International Union for Pure and Applied Chemistry (lUPAC) harmonized protocol. " In the past, results of several national and international interlaboratory trials had been accepted. Nevertheless, the following fundamental aspects of the harmonized protocol for the design of method-performance studies should be fulfilled ... 

Kimber, I. et al., The murine local lymph node assay Results of an interlaboratory trial, Toxicol. Lett., 55, 203, 1991. 

CIPM), which meets annually and oversees the work of the International Bureau of Weights and Measures (BIPM). The BIPM, based at Sevres just outside Paris, has the responsibility for international standards and is a center for international research and cooperation in metrology. The CIPM has created a number of advisory specialist committees (consultative committees) that are each chaired by a member of CIPM. The committee of interest to chemists is the Consultative Committee on the Amount of Substance (CCQM). It oversees the Avogadro project and coordinates a series of international interlaboratory trials called Key Comparisons (BIPM 2003), which... 

For data that are to be compared with expected limits of a distribution (e.g., in an interlaboratory trial), the z score is calculated using either the mean and sample standard deviation of the data,... 

In interlaboratory trials that are not informed by an assigned value and target standard deviation, a robust z score is used to account for all the available data... 

The Cochran critical values for 99% and 95% for 9 degrees of freedom are 0.4775 and 0.5727, respectively. The calculated value for laboratory A is 0.514, and so the laboratory fails at 95% but not at 99% (the actual probability is 98.5%). In this interlaboratory trial the laboratory would be flagged, but its data still included in calculations of group means and standard deviation if no other problems were encountered. 

Standard methods of analysis published by bodies such as the American Society for Testing and Materials (ASTM), de Normalisation (CEN), or ISO are rigorously tested and validated in method performance, or validation, studies. These interlaboratory trials can establish reproducibility and method bias and also give some confidence that the method can be used in different environments. Laboratories are chosen with an expectation that they can competently follow the proposed method, which will have already been extensively validated before an interlaboratory trial is contemplated. To this end, a pilot trial is sometimes undertaken to ensure the method description can be followed and to give an initial estimate of the precision of the method. 

When ASTM, followed by ISO and others, started conducting systematic interlaboratory trials to obtain precision data for test methods, the true state of affairs became apparent15. For many standards the variability was worse than realised and in some cases was so bad as to question whether the tests were worth doing at all or whether specifications based on them could be considered valid. The general advance of the quality movement prompted these investigations and have ensured that reproducibility has continued to occupy one of the top spots for attention in recent years. 

Interlaboratory trials with the organiser making detailed assessments of the laboratories is clearly particularly suited to helping individual laboratories and will at least qualitatively indicate the parameters requiring attention. This approach is, however, very expensive in total effort. The Intercal approach does not identify the causes of variability immediately but certainly alleviates the effect and, because trials are on-going, allows improvement to be monitored. Systematic quantification of the effect of individual parameters is probably the most cost effective approach and is the most useful for aiding standards committees to improve the specification of methods, but is of less direct help to individual laboratories. 

Any shortcoming in a standard can only be put right after analysis has pinpointed the problems. Hence, standards committees cannot act quickly if an interlaboratory trial reveals excessive variability. It is highly unlikely that faults in standards account for the majority of variance, although clearly it is important that any that do exist are identified and action taken. 

The most critical part of the method selection is determining the relative merits of seemingly diverse HPLC separations. Unfortunately, there is no standard against which all HPLC methods for vitamins are presently compared. However, there are indicators of assay reliability for which the analyst should look. These include measures of precision, accuracy, and reproducibility recoveries from spiked food samples linearity of calibration limits of detection measures of peak purity comparisons with existing recognized methods and results of collaborative or interlaboratory trials (9,10). 

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). 

A modified BCR sequential extraction was recommended based on this work, and its performance compared with the original BCR procedure via an interlaboratory trial (Rauret et al, 1999). Improved reproducibility was obtained. A new sediment reference material CRM 701 has been certified for metals extractable by the modified BCR procedure (Pueyo et al., 2001). The modified extraction protocol is given in the Appendix. It is important to emphasise that extractions should be performed exacdy as described if results comparable with other users are to be obtained. 

A common example where ANOVA can be applied is in interlaboratory trials or method comparison. For example, one may wish to compare the results from four laboratories, or perhaps to evaluate three different methods performed in the same laboratory. With inter-laboratory data, there is clearly variation between the laboratories (between sample/treatment means) and within the laboratory samples (treatment means). ANOVA is used in practice to separate the between-laboratories variation (the treatment variation) from the random within-sample variation. Using ANOVA in this way is known as one-way (or one factor) ANOVA. 

A relevant juristical statement about the precision of a method can only be made after defining the performance characteristics obtained from a round robin or interlaboratory trial study, as for instance described in ISO 5725 (ISO 1994). This study is used to determine the statistical key data about the precision of a method. The international standard ISO 5725 has been adopted by many countries. 

Gylseth B, Churg A, Davis JM, et al. 1985. Analysis of asbestos fibers and asbestos bodies in tissue samples from human lung An international interlaboratory trial. Scand J Work Environ Health 11 107-110. 

This often has distinctly different requirements. Since residue testing is a factor in world trade, it is vital that there is comparability of results between national laboratories undertaking drug residue monitoring. Thus, an interlaboratory trial of the method is very desirable. Furthermore, it is most unlikely that all regulatory laboratories will be identically equipped, and a regulato-... 

Klimisch HI, Pauluhn J, Hollander HW. 1988. Inhalation hazard test. Interlaboratory trial with OECD method 403. Arch Toxicol 61(4) 318-320. 

A three step extraction procedure was designed based on acetic acid extraction (step 1), hydroxylammonium chloride extraction (step 2) and hydrogen peroxide/ammonium acetate extraction (step 3). This scheme (described elsewhere [22]) was tested in two interlaboratory trials on Cd, Cr, Cu, Ni, Pb and Zn on freshwater sediment reference materials [23,24], the results of which are summarised in Table 9.6. 

Welz B, Verlinden M. 1986. IUPAC interlaboratory trial - selenium determination in human body fluids using hydride-generation atomic absorption spectrometry. Acta Pharmacol Toxicol 59 577-580. 

Some estimates of the precision of various typical modem procedures are given in Table 3. General clinical chemists may well be surprised at some of the values when compared with objective experience in other areas. Thus it is well known that the estimation of serum bilimbin has a coefficient of variation (C.V.) of around 15% in most laboratories, and that with creatinine a value of 8-10% is fairly normal (W3). In a recent blind (but not double-blind) interlaboratory trial of a standardized method for 17-ketogenic steroids and one for 17-ketosteroids, Gray et al. (GIO) obtained C.V. s varying between 4% and 14% for both methods. Six of the ten laboratories cooperating in the trial had special steroid experience and were asked to obtain duplicate estimates of any thirty routine specimens of urine. These results are for two well-established and relatively simple procedures. 

Van Dyck K, Robberecht H, Van Cauwenbergh R, Deelstra H, Arnaud J, Willemyns L, Benijts E, Centeno JA, Taylor H, Soares ME, Bastos ml, Eerreira MA, D Haese PC, Lamberts LV, Hoenig M, Knapp G, Lugowski SJ, Moens L, Riondato J, Van Grieken R, Claes M, Verheyen R, Clement L and Uytterhoeven M (2000) Spectrometric determination of silicon in food and biologcal samples an interlaboratory trial. J Anal Atom Spectrom 15 735-741. 

These collaborating studies have been extremely wide ranging and have included (a) co-operation in interlaboratory sun/eys of analytical proficiency (b) development of reference methods (c) collaborative interlaboratory trials of reference methods (d) cooperative assessment of reference materials. All of these lUPAC collaborative studies have had a beneficial influence on work carried out by other groups and organisations concerned with trace element analysis. 

In a subsequent lUPAC sponsored interlaboratory trial, selenium was determined in a range of lyophilized human materials using hydride generation atomic absorption spectrometry (Welz et al., 1987). The results obtained by the 13 laboratories showed excellent agreement with the values established in previous interlaboratory trials by experienced laboratories using a range of independent techniques. 

Brown, S.S., Healy, M.J.R. and Kearns, M. (1981a) Report on the interlaboratory trial of the reference method for the determination of total calcium in serum. Part I, J. Clin. Chem. Clin. Biochem. 19, 395. 

Ihnat, M., Thomassen, Y., Wolynetz, M.S. and Veillon, C. (1986a) Trace element data reliability in Clinical Chemistry — interlaboratory trials and reference materials, Acta Pharm. Toxicol. 59, suppi VII, 566. 

Ihnat, M., Wolynetz, M.S., thomassen, Y. and Verlinden (1986b) Interlaboratory trial on the determination of total selenium in lyophilized human blood serum. Pure Appl. Chem. 58, 1063. 

---