Accreditation and Quality Assurance (v.14, #4)
What is our best measurand when measuring “something” in “something”? by Paul De Bièvre (177-178).
Comparison of measurement uncertainties in direct plasma low-density lipoprotein cholesterol method of measurement and indirect estimation according to Friedewald equation by Xavier Fuentes-Arderiu; Socorro Braga-Fernández; Laura Freire-Campo; José Vicente García-Lario; María Inmaculada García-Martín; José Luís Jorde-Andrés; Encarna Largo-Caballerizo; Javier Lugo-Arocena; Cristina Pardo-Laseca; Santiago Villanueva-Curto; Santiago Juvé-Cuxart (179-183).
Concentration of low-density lipoprotein (LDL) cholesterol in plasma is appropriate to detect coronary heart disease risk. Many laboratories use the Friedewald equation for estimating this quantity instead of direct measurement. Nine clinical laboratories having an Hitachi selected 10 adult patients having results of cholesterol, high-density lipoprotein (HDL) cholesterol, LDL-cholesterol, and triglyceride concentrations in serum or plasma. Additionally, LDL-cholesterol concentrations were estimated using the Friedewald equation. Combined and expanded uncertainties were calculated for all results. The mean relative combined uncertainty corresponding to direct measurement of LDL-cholesterol concentrations is 6.9% and that corresponding to estimation is 19.4%. The expanded uncertainty (k = 2) of a result equal to 3.40 mmol/L (the discriminant value universally used to detect cardiovascular risk) obtained by direct measurement and by estimation using the Friedewald equation, are 0.22 and 0.55 mmol/L, respectively. The uncertainty of the direct measurement is very much lower than the estimation with the Friedewald equation.
Keywords: LDL-cholesterol; Friedewald equation; Measurement uncertainty; Coronary heart disease
Non-parametric estimation of reference intervals in small non-Gaussian sample sets by Johan Bjerner; Elvar Theodorsson; Eivind Hovig; Anders Kallner (185-192).
This study aimed at validating common bootstrap algorithms for reference interval calculation.We simulated 1500 random sets of 50–120 results originating from eight different statistical distributions. In total, 97.5 percentile reference limits were estimated from bootstrapping 5000 replicates, with confidence limits obtained by: (a) normal, (b) from standard error, (c) bootstrap percentile (as in RefVal) (d) BCa, (e) basic, or (f) student methods. Reference interval estimates obtained with ordinary bootstrapping and confidence intervals by percentile method were accurate for distributions close to normality and devoid of outliers, but not for log-normal distributions with outliers. Outlier removal and transformation to normality improved reference interval estimation, and the basic method was superior in such cases. In conclusions, if the neighborhood of the relevant percentile contains non-normally distributed results, bootstrapping fails. The distribution of bootstrap estimates should be plotted, and a non-normal distribution should warrant transformation or outlier removal.
Keywords: Reference intervals; Bootstrap; Re-sampling; Algorithm; Non-parametric; Percentile; Confidence intervals; Gaussian; Distribution
Utilisation of customer feedback in a university hospital laboratory by Paula Irmeli Oja; Timo Tapani Kouri; Arto Juhani Pakarinen (193-197).
The article describes the customer feedback system at a university hospital laboratory and the analysis of the feedback data from clinical customers in 2001–2006. The most common subject matters of the feedback were suspicion of the validity of laboratory results, delay in service and lacking test results, covering 82% of the 115 reports. The investigations of the cases revealed errors or defects in laboratory services in 81 cases. The most common errors or defects were erroneous test results (35 cases; 43%), delayed test results (24 cases; 30%) and lacking test results (15 cases; 18%). The most common underlying causes for a laboratory error or defect were unintended errors and non-compliance with operating instructions. Seventy-six percent of the feedback reports led to corrective actions. It is important to react to every instance of customer feedback and to find out possible errors or defects in the laboratory process. Uncovering the underlying causes makes adequate corrective and preventive actions possible.
Keywords: Complaints; Customer feedback; Laboratory errors; Quality improvement
Integration of metrological principles and performance evaluation in a proficiency testing scheme in support of the Council Directive 98/83/EC by Andreja Drolc; Magda Cotman (199-205).
Proficiency testing as a means of external quality assessment plays the role of independent evidence of laboratories’ performance. To enable laboratories to fulfil the requirements stated in legislation, methodology for evaluation of laboratories’ performance in proficiency testing schemes should incorporate principles of measurement results which are fit for intended use and incorporate evaluation of laboratories’ performances based on independent reference value. A proficiency testing scheme was designed to support Drinking Water Directive (98/83/EC) specifically. The methodology for performance evaluation, which takes into account a “fitness for purpose”-based standard deviation for proficiency assessment, is proposed and discussed in terms of requirements of the Drinking Water Directive. A ζ′-score, modified by application of target uncertainty was developed in a way that fulfils requirements defined in the legislation. As an illustration, results are reported for nitrate concentration in water. The approach presented can also be applied to other fields of measurements.
Keywords: Proficiency testing; Directive 98/83/EC; Fitness for purpose; Target measurement uncertainty; Measurement traceability
Implementation and assessment of the quality management system in research in a laboratory of the French Food Safety Agency: application to PhD student work by Jérôme Henri; Pascal Sanders; Hélène Lelievre; Nathalie Cadieu (207-217).
The approach of quality in research has rarely been described from a practical point of view. It has been decided to develop a system of management of quality in research in our laboratory to guarantee the results of the research. It is based on the requirements of standard ISO 17025 for the execution of tests associated to the research, on good laboratory practices for the description of studies and on standard ISO 10006 for all the requirements associated to the management of a project. The tools used and the role of all the participants are described here to help researchers to implement a system in their own laboratory. We are aware of the difficulties that implementing such a system pose in academia, especially with meeting the requirements of ISO 17025, but it is still achievable and even essential because research is mediatised and must therefore provide guarantees on its practices.
Keywords: Quality in research; ISO 17025; ISO 10006; GLP; Project management plan
Response on “Regulations in the field of residue and doping analysis should ensure a well-defined risk of a false positive declaration” by N. M. Faber by Peter Van Eenoo; F. T. Delbeke (219-221).
Dr. Faber provided support for a valuable methodology for the comparison of mass spectra. However, since many other factors are involved in the evaluation process before declaring a sample “positive”, i.e. reporting an adverse analytical finding, his conclusions on this subject are incorrect. In this paper a number of arguments are listed showing that Dr. Faber did not take all factors under consideration when making his claims. These factors involve parameters directly related to mass spectrometric criteria, but also to other equally important factors including sample preparation specificity, retention time (since in doping control mass spectrometry is used as part of a hyphenated technique), etc. These arguments clearly illustrate that the risks for a false positive declaration can not be calculated according to Dr. Faber’s proposed methodology.
Keywords: Residue analysis; Doping analysis; Mass spectrometry; Diagnostic ion; Abundance ratio; False positive
Towards more fair and effective doping tests based on chromatography with mass spectrometric detection by Nicolaas (Klaas) M. Faber (223-226).
Van Eenoo and Delbeke in Accred Qual Assur (2009) have criticized Faber (in Accred Qual Assur, 2009) for not taking “all factors under consideration when making his claims”. Here, it is detailed that their criticism is based on a misunderstanding of examples that were merely intended to be illustrative. Motivated by this criticism, further discussion is provided that may help in the pursuit of more fair and effective doping tests, here exemplified by chromatography with mass spectrometric detection. Surely, any doping test can only be improved or even optimized if the risks of false positives and false negatives are well defined. This requirement is consistent with a basic principle concerning mathematical approximations (Parlett in “The symmetric eigenvalue problem”, Prentice-Hall, Englewood Cliffs, 1980): apart from just being good, they should be known to be good.
Keywords: Doping analysis; Mass spectrometry; False positive; False negative
Metrology in Africa: role of the CAFMET by Abderafi Charki; C. Ferrier; G. Bonnier; M. Berrada; E. A. Hassani; L. Osseni (227-230).
This article presents the CAFMET (African Committee of Metrology), a non-profit organization that aims to meet the needs of African countries in the field of metrology. Its objectives are to spread a metrology culture in Africa via various domains in metrology such as, analysis, tests, and calibration. The difficulties of the development of metrology in Africa and the different activities, events and projects of the CAFMET are all presented.
Congress, conferences, workshops and courses (231-232).