Alternative Sample-Homogeneity Test for Quantitative and Qualitative Proficiency Testing Schemes.

Proficiency Testing (PT) External Quality Assessment (EQA) schemes are designed to ascertain the ability of individual laboratories to perform satisfactorily with respect to their peer laboratories or to limits imposed by external sources. Observed deviation of a laboratory result for a PT sample must be entirely attributed to the laboratory and not to the PT provider. To minimize the probability that deviations could be attributed to the PT provider, sample homogeneity should be assured. It is generally required that for quantitative parameters, the standard deviation among PT units should be calculated on the basis of duplicate measurements of at least 10 samples chosen at random, and the standard deviation among PT units should not exceed 0.3 times the standard deviation used to evaluate laboratories. Because this approach has important drawbacks, an alternative procedure is proposed by applying the theory of acceptance sampling to the assessment of sample heterogeneity for both quantitative and qualitative data and deriving acceptance limits on the basis of minimizing the probability of falsely evaluating laboratories. For obtaining acceptance limits for quantitative parameters, a distinction is made between laboratory evaluation using fixed limits on the one hand and laboratory evaluation using limits that are based on the variability of the reported results on the other hand. Sequential tests are proposed to evaluate sample heterogeneity by means of a comparison with the χ2 distribution. For qualitative parameters, acceptance-sampling plans are proposed that are based on minimizing the joint probability of rejecting batches that have a satisfactory amount of defective samples and accepting batches unnecessarily. The approach for quantitative parameters is applied on samples for a PT scheme of ethanol quantification and for qualitative parameters on the presence of monoblasts in a blood smear. It was found that five samples could already be enough to prove that the batch was homogeneous for quantitative parameters, although more than 20 samples were needed to prove homogeneity for qualitative parameters. This study describes a direct relation among the objective of an PT round, the criteria for evaluating the results, and the sample heterogeneity. When samples are effectively homogeneous, less measurements are needed than current practices require. A drawback of the proposed approach is that the number of samples to be tested is not known beforehand, and good knowledge of the analytical variability is crucial. The formulas to be applied are relatively simple. Despite the drawbacks, the proposed approach is generally applicable for both quantitative and qualitative data.

HbA1c: EQA in Germany, Belgium and the Netherlands using fresh whole blood samples with target values assigned with the IFCC reference system.

BACKGROUND: External quality assessment/proficiency test (EQA/PT) organizers play an important role in monitoring the performance of HbA1c measurements. With increasing quality of the assays, HbA1c is increasingly used for diagnosis of diabetes and the demands on EQA/PT organizers themselves are rising constantly. EQA organizers in Germany (INSTAND), Belgium (WIV/IPV), and the Netherlands (SKML) organized a program with commutable samples and target values assigned with the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) reference system. The aim of this project was to confirm the logistic feasibility of organizing synchronically in the three countries, an accuracy-based EQA program with fresh whole blood, to investigate the performance of HbA1c assays within and across countries and manufacturers, and to review the EQA acceptance limits. METHODS: Throughout 2015, ten fresh whole blood samples were supplied to the participants. Aggregated results were evaluated according to the IFCC model for quality targets at four levels: overall, per country, per manufacturer, and per country per manufacturer. RESULTS: Robust results in summer and winter demonstrated the feasibility of organizing an EQA with fresh whole blood samples in three countries. The overall performances, as well as the performance for each country were very similar: results fell within the IFCC criteria. Although substantial differences between results from different manufacturers were present, the performances of laboratories using tests of the same manufacturer were strikingly similar in the three countries, suggesting that the quality of HbA1c assays is for the most part manufacturer- related. The improved design of the EQA program also suggested that acceptance limits for performance can be reduced to approximately 8%.

Comparison of different approaches to evaluate External Quality Assessment Data.

In EQA programs, Z-scores are used to evaluate laboratory performance. They should indicate poorly performing laboratories, regardless of the presence of outliers. For this, two different types of approaches exist. The first type are "outlier-based" approaches, which first exclude outlying values, calculate the average and standard deviation on the remaining data and obtain Z-scores for all values (e.g., Grubbs and Dixon). The second type includes the "robust" approaches (e.g., Tukey and Qn or the algorithm recommended by ISO). The different approaches were assessed by randomly generated samples from the Normal and Student t distributions. Part of the sample data were contaminated with outliers. The number of false and true outliers was recorded and subsequently, Positive and Negative Predictive Values were derived. Also, the sampling mean and variability were calculated for location and scale estimators. The various approaches performed similarly for sample sizes above 10 and when outliers were at good distance from the centre. For smaller sample sizes and closer outliers, however, the approaches performed quite differently. Tukey's method was characterised by a high true and a high false outlier rate, while the ISO and Qn approaches demonstrated weak performance. Grubbs test yielded overall the best results.