Objective To investigate the comparability between laboratories and the performance of precision in clinical laboratories for blood gas and acid-base analysis by external quality assessment (EQA) and internal quality control (IQC).Methods Fifteen vials of EQA materials were distributed to the laboratories by global Express Mail Services (EMS).The activities were carried out three times,five-level samples were determined every time.After the measurement results and coefficients of variation of internal quality control data of April were reported,the collected data were divided into peer groups based on laboratory instruments.The medians of each group were taken as the target value to analyse the pass rate and coefficient of variation (CV) of each group after the removal of outliers.Internal quality control data was collected by Web-based the EQA software acquisition system which collect the CVs of the control data of nationwide blood gas and acid-base analysis project in April 2012 internal quality control,the CVs of IQC data were compared according to analyzers and testing items after the removal of outliers.All the data were analyzed using EXCEL 2007 and SPSS.Results A total of 570 laboratories participate in the EQA scheme.The CV of Inter-laboratory of PO2 was largest in all the same instrument groups; For the same item,Siemens group displayed larger CVs than other instrument groups.A total of 525 laboratories returned internal quality control data in accordance with the provisions,there are no significant difference in CVs among the same instrument group and among the same test of inter-instrument.Conclusions The comparability of results for blood gas and acid-base analysis are mostly good between laboratories in China,besides,some of which need to improve.The laboratories should pay more attention to IQC so as to secure the reliability of results.

Objective to evaluate the analytical performance of 7 open automatic biochemistry analysis systems in terms of precision,linearity,anti-interference ability and trueness on determination of cholesterol.Methods Performance verification test.There were 7 open analysis measurement systems composed of 7 kits as well as calibrators from Biosina,Baiding,Fosun,Dongou,Kehua,Maker and Wako company respectively,and Hitachi 7170 automatic analyzer were chosen as test systems.The repeatability CV and inter-lab CV were assessed according to Clinical and Laboratory Standards Institute (CLSI) protocol EP5-A2.The linearity range was evaluated on the basis of CLSI EP6-A,the series concentrations of cholesterol were 0,2.07,4.14,6.21,8.28,10.35,12.93,20.69 and 25.86 mmol/L cholesterol.Hemoglobin,ascorbic acid (vitamin C) and intralipid were applied as interfere materials in interference testing according to CLSI EP7-A2.The trueness was evaluated on the basis of China national lipids standardization program,the concentrations of 10 samples ranged from 2.88 to 5.42 mmol/L measured by reference methods.Results When the low level sample (2.71 mmol/L) measured,the repeatability CV were 0.54％,0.79％,0.56％,0.51％,0.56％,0.48％ and 0.49％ respectively,intra-lab CV were 1.00％,1.06％,1.28％,0.89％,1.08％,1.13％ and 1.05％ respectively.When the high level sample (5.12 mmol/L) measured,the respective repeatability CV were 0.40％,0.41％,0.51％,0.48％,0.47％,0.45％ and 0.47％,the respective intra-lab CV were 0.82％,0.69％,1.27％,0.70％,0.70％,1.08％ and 0.69％.The upper limits of linearity range of A,B,D,F was 12.93 mmol/L and for C,E,G was 20.69 mmol/L.There is no significant interference on 7 systems with chyle concentration of 1.6％ or hemoglobin concentration of 4 g/L.Given the interference bias ≤ 4％,the interference concentrations of ascorbic acid were 228,215,225,2840,2840,217 and 2840 μmol/L respectively.In trueness verification experiment,the bias of 7 systems all met the target value (≤ 3％).Conclusion The analytical performance of 7 systems in terms of precision,linearity,trueness and anti-interference all met the requirements of clinical specifications.The performance of anti-interference and measurement trueness of several systems could be improved.

Objective To verify the trueness and assess the traceability of results from a routine chemistry system procedure for measurement of urea and ereatinine in serun.Methods Series of fresh frozen patieot sera,whose values of urea or creatinine were assigned by isotope dilution gas chromatography mass spectrometry (ID-GC/MS) or isotope dilution liquid chromatography tandem mass spectrometry (ID-LC/MS/MS),were chosen to be analyzed by a routine chemistry system.The measurement results of urea and creatinine by the routine chemistry system were used for linear regression analysis against the assigned values bv the ID-MS method to calculate the percentage deviation and assess the expected bias.Results For urea and creatinine,the linear regression equations between the routine chemistry system and ID-MS methods were Y =0.9890X + 0.0192 (R2 =0.9990) and Y =0.9815X-6.4794 (R2 =0.9989),and the average percentage bias were-0.41％ (P ＞0.05) and-4.20％ (P ＜ 0.05),respectively.The expected percentage bias at three medical decision levels were-0.46％,-0.83％ and-0.96％ for urea and -15.90％,-5.87％ and-2.95％ for creatinine.Conclusions The results of urea analyzed by the routine chemistry system were consistent with the ID-MS method,which suggested that the results of the routine system procedure could be traced to ID-GC/MS method.For creatinine,the bias between the results of routine procedures and the assigned values met the minimum acceptance criteria' derived from biologic deviations,which would be better if its specificity improved.

Objective To investigate the current situation of precision of internal quality control (IQC) in total cholesterol,triglyceride,HDL-cholesterol,and LDL-cholesterol and provide improvement measurements.Methods Web-based External Quality Assessment (EQA) system was used to collect IQCdata of lipid tests from 581 EQA participant laboratories nationwide.The data include the coefficient of variation (CVs) of IQC data under control in April 201 1 and long-term cumulative data.Excel 2007 was applied for data processing after excluding the invalid data.Acceptable rates of CVs of two-lot internal quality controls in 4 lipid testing were calculated according to 6 criteria,that were 1/4TEa,1/3TEa,allowable imprecision of National Cholesterol Education Program (NCEP) and the specifications based on biological variation including the optimal,appropriate and minimal allowable imprecision.Results Four hundred and thirty-five,434,405 and 360 laboratories reported the data of level 1 IQC for total cholesterol (TC),triglyceride (TG),HDL-C,LDL-C respectively,while 214,214,192 and 171 reported the data of level 2 IQC respectively.Acceptable rates of TC,TG,HDL-C,LDL-C based on NECP criteria were 69.2％ (304/435),85.3％ (370/434),61.3％ (48/405) and 69.0％ (248/360) for level 1 respectively while 81.3％(174/214),91.6％ (196/214),75.5％ (145/192) and 81.3％ (139/171) for level 2 respectively.In the group which met the NECP criteria,the proportion of using matching detection system was much higher than the group which did not meet the criteria.Conclusions It is an effective way for clinical laboratories to improve test quality by monitoring the current and cumulative CVs of internal quality control and comparing them against proper evaluation criteria to evaluate if the analysis system can meet quality requirements.

Objective To investigate and analyze the reasons of fa0ilure in external quality assessment(EQA) for routine chemistry and provide the basis for the corrective and preventive actions.Methods Based on the network system of NCCL EQA the reasons of failure in 2013 national routine chemistry external quality assessment program were investigated,among which the reasons were classified and analyzed with seven sources of problems which were clerical errors,methodological problems,equipment problems,technical problems,EQA materials problems,EQA Evaluation problems and unable to explain after investigation.Results The return rate of this root cause investigation for each analyte ranged from 33.3％ to 80.0％.The major reason for unacceptable analyte included clerical errors (6.5％) (decimal point position error:70.1％;unit error:20.8％;instrument or method coding error:8.1％),methodological problems (45.1％)(calibration:54.2％;reagent:38.0％;EQA material:7.8％),equipment problems (28.5％) (no regular maintenance:98.0％;pipeline error:2.0％),technical problems (8.2％) (do not follow SOP:80.4％;EQA material redissolve error:10.6％;placing order error:9.0％) and unable to explain (11.7％) (system error:68.2％;random error:31.8％).There were no EQA materials problems or EQA Evaluation problems in this survey.Analysis systems' grouping statistics were implemented for seven analytes including sodium,chlorine,phosphorus,direct bilirubin,total iron binding capacity,copper,and zinc.Unsatisfied EQA proportions of mating system were lower than nonmatching ones for the majority of analytes.Conclutions Further work on EQA should be undertaken by clinical laboratories.Laboratories should use reagents with high quality as well as improve the operation technology and sense of responsibility.Only in this way,can the accuracy and reliability of testing results be guaranteed.

Objective To evaluate the analytical performance of lipid testing with six sigma(a) in order to promote the quality improvement.Methods Sixty-four Laboratories participated in trueness verification program and Internet-Based Interlaboratory Comparison of Internal Quality Control Data program lipid testing in 2012 were included in this study.The CVs and Bias of TC,TG,HDL-C,and LDL-C for each laboratory were evaluated from the reporting data.The sigma metrics based on the Tea from National Cholesterol Education Program according to the formula:σ =(TEa-Bias)/CV for each analyte.Quality goal index (QGI) was also calculated to investigate the reason for dissatisfied performance as follows:QGI =Bias/ (1.5CV).Results The rates for σ≥6 of TC,TG,HDL-C,and LDL-C were 21.9％,34.4％,9.38％,and 18.8％,respectively.The corresponding rates for 3 ≤ σ ＜ 6 were 31.3％,29.7％,29.7％,and 34.4％,while the rates for σ ＜ 3 were 46.9％,35.9％,60.9％,and 46.9％,severally.For analytes which σ were less than 3,91.3％-100％ of the laboratories should improve the trueness first,0.00％-8.75％ of the laboratories should improve the precision,and 0.00％-2.08％ should improve both.Conclusions Six sigma is an effective tool for quality control in clinical laboratory,which can help improve the quality level for laboratory testing.The analysis of performance of the lipid testing shows that further effort is acquired to enhance the supervision of the trueness verification.

Objective To describe and compare the roles of 3 external quality assessment programs in assessment of analytical quality of serum creatinine and urea.Methods Research in quality management methods.Sixty-five laboratories those enrolled in the Natonal Center for Clinical Laboratories′programs of routine chemistry external quality assessment ( EQA ) , trueness verification ( TV ) for small molecular metabolites and external comparison of internal quality control ( IQC) simultaneously in 2013 were selected , the performances of those laboratories of serum creatinine (crea) and urea in terms of total errors(TE), bias and CV were obtained by using the above 3 programs, and these performance were assessed against the criterion listed in the analytical quality specifications for routine analysis in clinical biochemistry ( WS/T 403-2012).The failure ratio of 65 laboratories on each performance was calculated , the sensitivity of 3 external quality assessment programs in detection of analytical quality deficiency among clinical laboratories were compared.Results Only 1 laboratory failed in the 1st routine chemistry EQA in terms of TE of creatinine , failure ratio is 1.5%(1/65).Three laboratories failed in the 2nd EQA and caused a failure ratio of 4.6%(3/65).For serum urea, 3 laboratories failed in the 1st routine chemistry EQA with a failure ratio of 4.6%(3/65).Two laboratories failed in the 2nd EQA with a failure ratio of 3.1%(2/65).The failure ratios of creatinine determination in two samples in TV were 41.5%(24/65) and 21.5%(14/65) respectively, and the failure ratio of urea determination were 53.8%( 36/65 ) and 32.3%( 21/65 ) respectively.In the program of external comparison of IQC , the CVs of creatinine and urea determination ranged from 0.7% to 6.2%and from 1.0%to 7.2%respectively, their respective failure ratio range were 15.4%(10/65) and 40.0%(26/65).The failure ratio in routine EQA were much less than those in the other two programs , the laboratories failed in routine EQA program were all failed in trueness verification or /and the comparison of IQC programs, but not vice versa.Conclusions By participating in the programs of routine EQA , TV and comparison of IQC laboratories could assess the performances of inaccuracy , bias and imprecision.Laboratories should participate in different external quality assurance programs to detect their quality issues and get improved.

Objective To study the applicability of a new analytical specification defined in WS/T 403-2012 in the external quality assessment schemes and the external comparision of internal quality control .Methods It was a quality management method study.Total error allowable criterions listed in WS/T 403-2012 and GB/T 20470-2006 were selected to assess the results of 23 analytes in the 1st challenge of 2013 routine chemistry external quality assessment.The acceptable rate of 23 analytes were calculated with the two specifications.Criterions of imprecision derived from the two standards were applied to assess the coefficient of variation with internal quality control data.Results With the specification based on WS/T 403-2012, the ratio of laboratories that all five samples were passed in the 1st challenge for 23 analytes ranged from 55.5%to 94.7%.The ratio of laboratories with 80%or more samples passed in 2013 EQA ranged from 73.9%to 98.5%.While ratios of two kinds described above evaluated based on GB /T 20470-2006 ranged from 63.0%to 99.2%, and from 90.0% to 99.7%, respectively.The acceptable rate of CV according to the two criterions ranged from 55.5% to 94.7% and 63.0% to 99.2%, respectively.Conclusions As evaluation criterions of external quality assessment allowable total error and internal quality control imprecision in routine chemistry , the specification in WS/T 403-2012 can be used to assess the analytical performance of clinical laboratory more objectively and comprehensively.It can help laboratories to identify the latent problems for further quality improvement.

External quality assessments play important roles in quality improvement in clinical laboratories, but most laboratories focused on the unsatisfied data only.With appropriate quality controls, laboratories can detect not only the error sourses of unsatisfied data but the potential error sourse of satisfied data as well.

Objective To evaluate the commutability of certified reference material IRMM ERM-DA 471/IFCC for cystatin C measurement among 8 methods.Methods 46 individual samples used for the commutability study were residual serum samples collected from clinical laboratories.The individual samples interspersedwith pooled sera and the certified reference material IRMM ERM-DA 471/IFCC and the whole set of samples was divided into 2 subsets for measurements in 2 days.The measurements were performed by8 different methods o.The samples were measured in duplicate order.Calibration was performed every day. Passing-Bablok regression was performed to compare the slopes and intercepts of mean values of the serum samples derived from different methods. Pearson correlation cofficient was also calculated. Deming regeression and 95%confidence intervals were calculated to evaluate the statistics commutability of ERM-DA 471/IFCC.The minimal specification of bias derived from biological variations was calculated to evaluate the clinical commutability.Results The within-laboratory CVs of pool sera ranged from 0.5% to 4.0%.The Passing-Bablok slope ranged from 0.765 to 1.311 and intercepts ranged from -0.04 to 0.241.The determination coefficient of Pearson regression ranged from 0.988 to 0.999.Deming regeression and 95%confidence intervals demonstrated commutability of ERM-DA 471/IFCC in 4/28 (14.3%) methods pairs. The minimal specfication bias ( 5.12%) demonstrated commutability of ERM-DA 471/IFCC in15/28 (53.6%) methods pairs.Conclusions The ERM-DA 471/IFCC domonstrated poor commutability between some methods pairs. The commutability of ERM-DA 471/IFCC should be evaluated before used as calibrators.