Lorena Zardo

External Quality Assessment Schemes: need for recognised requirements

Programs for Accreditation of clinical laboratories consider participation in External Quality Assessment Schemes (EQAS) a key element in the evaluation of testing procedures and improving them. One of the main functions of EQAS is to assess whether laboratories perform tests competently. It is therefore of utmost importance for laboratories to participate in EQAS that are in line with formally recognised requirements. Specific proposals have been made on how to design and execute EQAS by International Working Groups, but there seems to be no consensus on the best strategies to use and quality specifications to set out. The Clinical Pathology Accreditation (CPA) Program for EQA Scheme Accreditation (CPA-EQA) is the only program in Europe to provide a formal recognition of the quality of EQAS activities. The present paper reports on the experience of the Centre of Biomedical Research which is following an accreditation process for their own schemes in line with the CPA-EQA program and a proposal to set requirements that Italian schemes must follow to be recognised as valid and effective.

Risk management in laboratory medicine: quality assurance programs and professional competence

Abstract To guarantee excellent performance and service, the process of identifying and treating error risks must be integrated into the total testing process. Quality Assurance Programs (QAPs) represent an important tool that allows us to identify errors and pinpoint any need for further systematic investigations, and to rectify procedures to improve the inputs and processes by which the service is delivered. The models used by the laboratory to assure quality and manage the risk of errors have been modified in line with an approach in which the identification of quality goals and the redefinition of professionals duties and responsibilities are indispensable. Error risk is currently high in some areas of laboratory activity, and QAP is needed now more than ever. The present paper provides some descriptive examples of an approach that can be followed to manage an External Quality Assessment Scheme (EQAS) and quality indicators (QIs), the main tools used by laboratories to assure the quality of their service, for the prevention of error risk. In particular, we describe the correct approach to choose EQAS, to use information from the EQAS report, to design a QI model, and to analyze any QI data. The examples highlight that any well-designed quality system can be ineffective if it is not managed by highly competent professionals with a deep sense of responsibility. Clin Chem Lab Med 2007;45:756–65.

Reference intervals: are interlaboratory differences appropriate?

In laboratory medicine, analytical results and their relative reference intervals (RI) represent the terms of a binomial that makes a report effective. If gross errors are neglected, an analytical result is subject to two kinds of error: inaccuracy and imprecision. A hierarchy of models has to be applied to set analytical specifications, starting from the evaluation of the effect of analytical performance on clinical outcomes in specific clinical setting (1). This now applies only to few analytes. Instead, the evaluation of the effect of analytical performance on clinical decision and, in particular, data based on components of biological variation is applicable to a wide range of analytes (2). The reference intervals may be related to the results, which means that they are compromised by the same degree of relative inaccuracy. It has been proven that only with the adoption of an appropriate reference range we can obtain a clinically effective report. The recommendation that the single laboratory should calculate the “proper” RI springs from the need to make the relative RI “appropriate”. To verify the appropriateness of RI we evaluated data from the External Quality Assessment scheme (EQA) managed by the Biomedical Research Centre (160 laboratories, mainly in the Veneto area). To identify homogeneous groups of methods, the Centre requires from participant laboratories not only methodological information, but also the reference ranges used (in fact the results are sent from laboratories to the Centre as a patient report). Only a few analytes, are presented and discussed as examples. Despite its importance as risk factor for atherosclerosis and myocardial infarction, there are significant differences in the reference intervals used for cholesterol by the laboratories that use the same analytical methods (Table 1). About a third of the laboratories use the so-called “desirable value” (5.18 mmol/l) as the upper limit, but others use an upper limit as high as 7.25 mmol/l. These differences would be less serious if they were related to a different relative accuracy. However, the results from laboratories participating in the EQA were very similar. For instance, for the control sample 6B, the mean value (consensus value) was 6.79 mmol/l and standard deviation (SD) was 0.28 mmol/l. If this had been a sample from a real patient, the report would have been “normal” according to some laboratories and “abnormally high” according to others. Clearly, many laboratories erroneously use the populationbased reference intervals. Although to a lesser extent, some differences were found for glucose. In some laboratories the lower limit is so high (Table 2) that they should classify the control sample 5A (mean ± SD = 3.33 ± 0.17 mmol/l) as “hypoglycemia.” The differences in reference intervals used for sodium are large (Table 3) ranging from 130–140 mmol/l to 130–160 mmol/l. The most widely used range is 135–145 mmol/l (17% of the laboratories), but the second most widely used is 50% larger (135–150 mmol/l). Also in this case, the low interlaboratory inaccuracy is not consistent with different reference values, Tab. 1 Values for cholesterol obtained on control sample 6B: mean ± SD = 6.79 ± 0.28 mmol/l.

An Italian external quality assessment (EQA) program on urinary sediment.

BACKGROUND EQA programs on urinary sediment are rare. We describe an EQA Italian program which started in 2001 and involves today more than 300 laboratories. METHODS The program, which started with a questionnaire about the methodological aspects on urinary sediment, includes today four surveys per year. These ask the participants the identification and clinical associations of urinary sediment particles shown by colour images (surveys 1 and 3) and the diagnosis of clinical cases presented by both images and a short clinical history (surveys 2 and 4). The results of each survey are then scored and commented. RESULTS Questionnaire (participants = 287): most methodological aspects were not dealt with properly. IDENTIFICATION cells, lipids, casts and some contaminants were poorly known. However, when 27 particles were presented for the second time and 16 particles for the third time, the correct identification rate for most of them increased significantly. Clinical associations (No presented = 16): a correct answer was indicated by > or = 84% of participants for all particles but one. Clinical cases (No presented=4): lowest correct identification for urine contamination from genital secretion (77.3%), highest for ureteric stone (94.4%). CONCLUSIONS Our program shows that EQA programs are both useful and needed.

Appropriateness of cholesterol and triglycerides reporting checked by External Quality Assessment programs.

BACKGROUND The recommendations of the Second Joint Task Force of European and Other Societies on Coronary Prevention and the third Adult Treatment Panel report (ATPIII) released by the National Cholesterol Education Program are based on accumulating evidence concerning the contribution of lipoproteins and other risk factors in the development of coronary heart disease (CHD). The laboratories play an important role in the successful adoption of these guidelines. METHODS In External Quality Assessment (EQA) programs managed by the Center of Biomedical Research, results and respective reference intervals (RI) are sent as laboratorys medical form. We assessed how well the 200 participants to EQA scheme 2002 for clinical biochemistry reported total cholesterol (TC) and triglycerides (TGs) results according to either European or National Cholesterol Education Program (NCEP) guidelines. RESULTS Only 18% of laboratories reported total cholesterol concentrations correctly in terms of desirable, borderline-high, and high risk for the CHD development, 12% reported a single desirable value (180, 190, or 200 mg/dl), and 70% reported the RI (85 laboratories in the whole interval, 34 are the only upper reference limit and 15 are the desirable value in addition to RI). The upper reference limit was 200 mg/dl in 65% of cases, but 32% of laboratories presented higher limits, reaching values as high as 250-260 mg/dl. Only the 3.7% of laboratories reported triglyceride concentrations in terms of risk-oriented ranges for the CHD development, 6.8% the single desirable value, and 89.5% the RI. CONCLUSION Our study demonstrates that the current practice of reporting results for cholesterol and triglycerides does not follow the guidelines, and appropriate changes are required to be made.

Reply to W.G. Wood. Questionable results--who directs the EQAS organisers? Clin Chem Lab Med 2004;42:1073.

Prof. Wood raised the important question of how to handle ‘‘obvious mistakes’’ and encouraged the national EQAS organisers to report how they proceed in the case of blunders, so that in future guidelines a clear answer to this question can be given (1). There are two diametrically opposed approaches used by organisers of External Quality Assessment (EQA) schemes: (a) to strictly handle the results as they are sent or (b) to correct for an obvious mistake before further evaluation. We agree with the second approach: when we see an ‘‘obvious mistake’’ ascribable to preand post-analytical phases we ‘‘correct’’ the result before entering it, following a documented procedure in the framework of our quality management system that includes communication to the participant laboratory of the correction itself. The aim of EQA organisers must be to include as many correct results as possible in a survey and to ensure that the results from participants are correctly represented in the report. It is in the interest of all participants that as many errors as possible are corrected in order to improve the validity of the EQA report and to maintain the correctness of measures, such as consensus means or medians and the precision of different analytical methods. Even if the utilisation of a valid statistical procedure can individuate most obvious errors as outliers, we believe that errors due to the analytical phase could be separated from errors due to preor post-analytical phases (2, 3). The Centre of Biomedical Research (CRB) has managed EQAS in Italy since 1985 and its role is to identify laboratories with problems, to notify laboratories if their performance is unacceptable and to offer help and advice. Its programmes are the only ones to be accredited by the Clinical Pathology Accreditation Program (CPA-EQA) in Italy (4), they have a voluntary