Investigation of the minimum required months for imprecision analysis

Abstract

Objective To discover and certify the minimum required months of IQC (Internal Quality Control) data which were used to quantify the imprecision To identify the impact of test items, different concentrantions and years on the defective rate. Methods IQC data involving 20 analysis items (including Albumin, Creatine kinase, Total bilirubin, Alanine aminotransferase, Aspartate aminotransferase, Lactate dehydrogenase,γ-Glutamyl transferase, Alkaline phosphatase, Calcium, Chlorine, Creatinine, Glucose, Potassium, Sodium, Phosphorus, Triglyceride, Total cholesterol, Total protein, Uric acid, Urea) and six measurement systems (including Roche, Architect, Hitachi, Dade/Behring, Beckman, Olympus) were collected from hospitals in Shanghaibetween 2009 and 2011. A total of 3 534 groups, referred to one year laboratory′s IQC data of one concentration range, were analysed to find the minimum required months in each group when the cumulative months were compared with the population by using T test. The correlation coefficient of hospital′ grades, measurement levels and test items were evaluated by u test, and the percentage of groups of P>0.05 were collected. The cumulative IQC data′ coefficient variation (CV) of six months and eleven months were compared with the total CV, respectively. Imprecision higher than the professional specification was regarded as unqualified. Difference of unqualified rate among test items, concentrations and years were expolored. Results Rates of hospital′ grades, measurement levels, items are 94.2%，100%, 100% respectively, 88% of groups′ imprecision became stable in ten months. 25% groups reach the criteria that the relative bias is ≤5% when calculated the cumulative IQC data′ CV of six months, while 88% groups do when calculated the cumulative IQC data′ CV of ten months. The unqualified rate is 20% and the most unqualified item is Ca, the unqualified groups of low level are larger. With increase of year, the unqualified rate showed a downtrend. Conclusions Ten months′ IQC data is more reliable to quantify imprecision. Unqualified rates are different according to test items, measure levels and years, the establishment of professional specification should consider them. （Chin J Lab Med,2014,37:386-390） Key words: Laboratories, hospital; Quality control; Clinical chemistry tests; Uncertainty