Accreditation and Quality Assurance (v.18, #4)

Exploiting interlaboratory comparisons by Hendrik Emons (267-268).

Methodology in internal quality control of chemical analysis by Michael Thompson; Bertil Magnusson (271-278).
Internal quality control (IQC) is an essential feature of routine analysis, serving to ensure that the uncertainty of results found during the validation of a procedure is maintained over long periods of time. The primary method of IQC is to analyse a surrogate material alongside the test materials in every run of analysis and thus address run-to-run precision (a subset of VIM3-defined ‘intermediate conditions’). This ‘control material’ must be as similar as practicable in composition to the routine test materials, although there are always some differences. Results from the control material (control values) are plotted on a control chart, and out-of-control results have to be investigated and problems rectified. Considerable care is needed in obtaining correct values of the parameters for determining statistical control limits, and these can be adequately estimated only during routine use of the analytical procedure. In contrast, target control limits have to be set on a fitness-for-purpose basis and are necessarily wider that statistical control limits. An additional type of internal quality control can be executed by the analysis of duplicate test portions of some of the actual test samples. This provides a realistic dispersion, but addresses only repeatability precision. A further complication of duplication is that the precision of results typically varies with concentration of the analyte.
Keywords: Internal quality control; Statistical control; Target limits; Surrogate control materials; Duplication; Repeatability precision; Run-to-run precision; Intermediate precision

The trail from initial evidence examination to a DNA profile reported to match a suspect is long and complex. The different nature and great variability in the biological and DNA evidence to be recovered and analyzed, add to this complexity. Internal quality controls play an important role in maintaining a high-quality performance in daily forensic biology and DNA profiling practice. In many cases are empirical rather than analytical approaches adopted. Obviously, despite the fact of being necessary, the internal quality controls performed still need to be kept rational at a limited, yet acceptable level. Quality control from a forensic biology and DNA profiling horizon has a wider context and does not only concern obvious fit-for-purpose verifications of analytical processes, chemicals, or reagents in daily routine practice. It also includes control on computerized laboratory management and expert systems, laboratory environmental DNA monitoring, and the use of elimination DNA databases. In addition, a structured recording and handling of non-conformances and “near failures” is essential. Proper management of the non-conformances supports continuous quality improvements by learning from the errors occurring in daily practice. High transparency of non-conformances is important not only for internal improvements, but also for the criminal justice system as well as to maintain public confidence and trust. Together the quality controls used aim at maintaining evidence and DNA sample integrity and to accomplish correct results and interpretations by verifying that methods used data transfers and interpretations made are correct and performed according to validated and accredited conditions.
Keywords: Internal quality control; Forensic science; DNA analysis; ISO/IEC 17025

Control charts to evaluate long-term performance in proficiency tests by Angela Sorbo; Andrea Colabucci; Laura Ciaralli (291-298).
As proficiency tests (PTs) provider, the European Union Reference Laboratory for Chemical Elements in Food of Animal Origin (EURL-CEFAO) yearly organizes exercises on the determination of chemical elements within the area of its mandate. The scheme is addressed to the National Reference Laboratories (NRLs) of the European Union Member States. Each PT is planned giving priority to the matrices and elements for which maximum levels are set in Commission Regulation No. 1881/2006 and following amendments as well as to those mainly considered in National Residues Control Plans. Samples to be distributed are prepared in the Laboratory’s facilities using materials purchased on the market and often adjusting the concentration levels of the analytes around their maximum levels. In the 2006–2011 period, PTs were planned as a long-term programme pursuing the general objective of providing the NRLs with a scheme that can give them the opportunity to check and improve the performance of their analytical methods and to verify the effectiveness of any corrective actions through the repetition of the same matrix. Control charts give participants the possibility of a long-term follow-up of their performance promoting the improvement of Quality Control. To this aim, Shewhart and Cusum control charts for z-scores are prepared according to ISO 13528:2005 and updated after each PT for each laboratory. z-scores are assigned using values of σ pEURL-CEFAO that correspond to the level of performance appropriate to the demand. Cusum charts are also evaluated by EURL-CEFAO in order to identify problems that may cause a bias in the measurement method.
Keywords: European Union Reference Laboratory; Chemical elements; Proficiency test; Shewhart control charts; Cusum control charts; Long-term performance

The laboratories verify data from control samples through the application of typical control charts by using commercial spreadsheets. The criteria to decide whether a measurement procedure is in statistical control involves the interpretation of the control chart after each analytical run although the decision is not always practical. So, to detect long term trends in the process, a Cusum chart graphic representation can be implemented. The use of a V-mask verification tool will show the point from which the deviations are significant. Since the construction of a V-mask is complicated in practice, in this article is proposed the definition of a visual representation and also how to construct a tabular procedure with a commercial spreadsheet form in which the V-mask serves as control limits for the Cusum chart.
Keywords: Spreadsheet; Cusum control charts; Control limits; Type I and II errors; V-mask

Control charts in multi-element analysis of human urine by A. Sorbo; A. Semeraro; A. A. Pastorelli; M. Patriarca (307-312).
A plan for internal quality control (IQC) was designed to monitor the quality of analytical results obtained from the simultaneous determination by ICP-MS of nine trace elements (As, Cd, Mo, Ni, Pb, U, V, W and Zr) in 1350 samples of human urine collected as part of an epidemiological study. Analytical performances, assessed during the method validation, were as follows: limit of quantification ranging from 0.002 μg/L (U) to 0.9 μg/L (Ni); recovery rates varying between 82 % (As) and 110 % (W); intermediate precision as relative standard deviation ranging from 2.5 % (Mo) to 14.2 % (V). Only for Zr, a significantly higher relative standard deviation was obtained. Initially, two commercially certified reference materials (CRMs) based on freeze-dried human urine and containing the elements of interest at suitable levels of concentration were used as control materials. However, due to considerations of reliability, convenience and cost, a control material based on fresh human urine was prepared in-house, evaluated by comparison with the CRMs and implemented for the IQC. The IQC strategy was as follows: at least one control material was analysed in each analytical session and one more control material was analysed after every ten urine samples. The central lines of the control charts were based on the assigned concentration values; the warning and action limits were set on a statistical basis, as two-times and three-times the value of the uncertainty of certified values or the observed standard deviation. At the end of the study, the data collected for the in-house control material were re-evaluated and confirmed its suitability for the purpose. The use of an in-house control material had important advantages: it was more representative of the test samples; its preparation did not involve additional steps, therefore, reducing the risk of contamination; the cost of implementing the IQC was also reduced. Some issues concerning the control charts for As and Mo were addressed and will be discussed.
Keywords: Multi-element analysis; Human urine; In-house control material; Internal quality control; Shewhart control charts

Internal quality control as a tool for planning a robustness study regarding a multiresidue method for pesticides found in olive oil by Patrizia Stefanelli; Tiziana Generali; Silvana Girolimetti; Danilo Attard Barbini (313-322).
The implementation of the internal quality assurance program allows for demonstration of the performance characteristics of a method, as well as the avoidance of erroneous results. There is increasing concern in testing laboratories to ensure that the analytical process remains stable, giving reliable results under statistical control. Robustness is tested by introducing variations in experimental conditions and examining the effects on the results. The European document SANCO/12495/2011, “Method validation and quality control procedures for pesticide residue analysis in food and feed”, defines robustness as a parameter that can be derived from ongoing method verification. Internal quality control tools, such as the recovery control chart, were employed in this study to enhance the stability of the recovery rates and to investigate the experimental conditions that have a major influence on the quantification of recovery rates. The method investigated in this study permitted us to investigate eighteen pesticides in olive oil by using a gas chromatography–mass spectrometry technique. This method has been accredited to ISO/IEC 17025:2005 standards and was applied for 1 year in routine conditions for pesticide residues at the Italian National Reference Laboratory. The recovery control chart has showed that all recovery rates for 1 year were close to the maximum limit (120 %) of recovery performance criteria. Consequently, a “positive” bias has affected all data over longer periods of time. A robustness test was planned in order to investigate the grounds that most influenced the variability of the results. The robustness test involved the following three parameters: solvent used to dissolve the final extract, internal standards, and type of olive oil.
Keywords: Robustness study; Pesticide; Multiresidue method; Olive oil; Internal quality control

Preparation of in-house reference material of benzylpenicillin in milk and results of a Brazilian proficiency testing scheme by Bernardete Ferraz Spisso; Mychelle Alves Monteiro; Mararlene Ulberg Pereira; Rosana Gomes Ferreira; Rafaela Pinto da Costa; Betânia Souza Carlos; Shilton Thompson Cruz Negris; Marcus Henrique Campino de la Cruz; Armi Wanderley da Nóbrega (323-331).
Two milk test materials containing benzylpenicillin and a benzylpenicillin-free material were prepared and used for the operation of a Brazilian proficiency testing (PT) scheme according to the requirements of ISO/IEC 17043. The PT scheme was designed to include laboratories in charge of confirmatory analysis as well as screening analysis. Sets of sample vials, benzylpenicillin-free and spiked items containing 0.2 g of lyophilized milk, were distributed to each participant. Of 20 participants who reported screening data, 18 accomplished satisfying results, providing evidence of the capability of Brazilian laboratories to produce reliable qualitative information. The assigned value (robust average calculated from 6 results) and its uncertainty were (8.28 ± 0.52) μg/L. Participants’ performance was evaluated using z-scores. A small number of participants were able to report quantitative benzylpenicillin results, and consequently care should be taken in the interpretation of the laboratory’s performance, considering the statistical nature of z-scores. Long-term stability testing on remaining PT items showed that a stable benzylpenicillin quality control material was produced and can be used for internal quality control or validation purposes.
Keywords: Reference material; Proficiency testing; Internal quality control; Veterinary drug residues; Milk; Benzylpenicillin

Laboratories dealing with the analysis of food of animal origin have to face the shortage of both appropriate certified reference materials and pertinent proficiency tests (PTs). Therefore, internal quality control (IQC) materials have to be prepared in-house. The combination matrix/analytes/concentration, as well as the physical state, is a key issue. The European Union Reference Laboratory for Chemical Elements in Food of Animal Origin (EURL-CEFAO), within its activity as PT provider, has set a procedure to prepare test items of liquid milk spiked with arsenic, cadmium and lead at adequate values of concentration. This work describes in detail the steps of this procedure. Homogeneity and stability of the PT test items thus prepared were evaluated by the EURL-CEFAO using a sensitive and accurate method accredited according to ISO/IEC 17025. The analyte concentrations in spiked milk were found stable up to 18 months which is of particular interest for cadmium due to the low value of spiking (roughly 5 μg/kg). The outcome proves that the procedure used to prepare liquid milk spiked with chemical elements is suitable to produce in-house reference materials stable enough for analytical purposes. Furthermore, this procedure is easy to perform and can be applied by routine laboratories to produce samples for their IQC.
Keywords: Proficiency test; Internal quality control; Certified reference materials; Liquid milk; European Union Reference Laboratory; Chemical elements

A candidate material for the use as primary standard for silver determination was characterized with respect to total purity. Except the radioactive elements and He, all possible impurities were considered. Based on glow discharge mass spectrometry, inductively coupled plasma-mass spectrometry and carrier gas hot extraction measurements, the demonstrated total purity and its standard uncertainty is w(Ag) = (99.999 52 ± 0.000 11) %. The purity value and its uncertainty is dominated by the contributions from the measurements of the nonmetallic impurities, namely S, N, C and O.
Keywords: Silver; Purity; Total purity; Primary standard; Element determination

Evaluation of measurement uncertainty in EA–IRMS: for determination of δ 13C value and C-4 plant sugar content in adulterated honey by Hui Chen; Chun-lin Fan; Zhi-bin Wang; Qiao-ying Chang; Wei Wang; Xiao-ying Li; Guo-fang Pang (351-358).
In this study, the method for determining the stable carbon isotope ratio value was validated. Measurement uncertainty of stable carbon isotope ratio value of whole honey and its extracted protein derived from repeatability, reference gas, reference standards and calibration curve was calculated by applying the “bottom-up” approach according to Eurachem/CITAC guide. The expanded uncertainties for all results ranged from 0.14 to 0.19 ‰, with most of them between 0.15 and 0.16 ‰ (the coverage factor k = 2, the level of confidence p is approximately 95 %). The percentage contribution of each source to the relative combined uncertainty was calculated. The data indicated that calibration curves have more contribution to the relative combined uncertainty than repeatability and reference standards. On the other hand, the measurement uncertainty of C-4 sugar content in honey was estimated. Based on these results, 58 honey samples, such as acacia, chaste, Northeast China black bee, flowers and jujube honey, have been gathered to determine the C-4 plant sugar content adulteration in honey by elemental analyzer with an isotope ratio mass spectrometer. It can be found that all honey samples were not adulterated by C-4 plant sugar.
Keywords: Uncertainty; Honey; Stable carbon isotope ratio; Adulteration; C-4 plant sugar

Modelling the calibration of the industrial platinum-resistance thermometers according to GUM by Inseok Yang; Kee Sool Gam; Wukchul Joung; Yong-Gyoo Kim (359-362).
According to the Guide to the Expression of Uncertainty in Measurement (GUM, JCGM 100: 2008), the calibration process and its uncertainty evaluation should be expressed in terms of mathematical function(s) of input quantities. However, in practice, expressing measurement or calibration in a way that is fully compliant with GUM might be unrealistic and require a clear definition of the calibration process itself. Depending on the applied calibration process, different modelling equations with various complexities can be written. In this paper, four different approaches are given to model the calibration process of industrial platinum-resistance thermometers.
Keywords: GUM; Modelling of measurement; Calibration; IPRT; Uncertainty