ABSTRACT
Objective:To establish a candidate reference method for the determination of angiotensin Ⅱ in human plasma by isotope dilution liquid chromatography tandem mass spectrometry (ID-LC-MS/MS) and to evaluate its performance.Methods:Using [ 13C 6- 15N]-angiotensin Ⅱ as the internal standard, the plasma was accurately weighed by gravimetric method and mixed with a certain amount of internal standard. At the same time, enzyme inhibitor was added. After zinc sulfate solution protein precipitation and reversed-phase solid-phase extraction plate treatment, it was analyzed by liquid chromatography tandem mass spectrometry. The multi reaction ion monitoring mode(MRM)was selected by mass spectrometry to detect specific ion fragments of angiotensin Ⅱ and internal standard. The linearity, sensitivity, precision, recovery rate and uncertainty of the performance of the established method were evaluated according to ISO15193. Results:Angiotensin Ⅱ had good linearity in the range of 10-1 000 pg/g ( r=0.999 5), the lower limit of quantification was 7.68 pg/g, the analytical recoveries were 97.14% to 102.85%, intra-batch imprecision≤3.21%, inter-batch imprecision≤2.96%, and total imprecision≤3.67%. Conclusion:A method for the determination of plasma angiotensin Ⅱ was established by ID-LC-MS/MS. The method is accurate and reliable, and is expected to be a reference method for the determination of plasma angiotensin Ⅱ.
ABSTRACT
Objective:To establish a candidate reference method for serum progesterone using isotope dilution liquid chromatography tandem mass spectrometry (ID-LC/MS/MS) in our laboratory, validate the analytic performance of five clinical routine detection systems to explore the comparability of serum progesterone detection by different detection systems.Methods:A candidate reference method for serum progesterone using ID-LC/MS/MS method was established. The sample was pretreated by liquid-liquid extraction method, and the reversed phase liquid phase separation in positive ion mass spectrometry mode was used to detect progesterone in human serum, and the detection time of a single sample was controlled within 5 minutes by gradient elution. In order to improve the accuracy of the method, the bracketing calibration method (BCM) was used to establish the standard curve. The sensitivity, accuracy, precision and specificity of BCM and classical calibration curve method were evaluated according to CLSI C62-A, EP15-A2, EP6-A2 and EP9-A3, and the analytical performance and comparability of five clinical routine progesterone detection systems were evaluated,compared with ID-LC/MS/MS method, the bias at medical decision level 2 and 25 ng/ml was evaluated to see if they were <1/2TEa (12.5%).Results:The limit of detection (LOD) of ID-LC/MS/MS was 0.005 ng/ml. The recoveries of BCM method and classical calibration curve method are 97.95%-101.58% and 96.88%-110.70%, respectively. The measurement results of BCM method for certified reference materials are within its declared uncertainty range. The intra-and inter-assay coefficient of variation ( CV) of BCM method was less than 3.0%, which was better than that of classical calibration curve method ( CV: 2.48%-9.33%). The precision and linear range of the five clinical routine detection systems can meet the detection requirements. The measurement bias of detection system 1, 3 and 5 at 25 ng/ml of medical decision level was less than 1/2TEa, and the measurement bias at 2 ng/ml of medical decision level was more than 1/2TEa. The measurement bias of detection system 2 and 4 at two medical decision levels was less than 1/2TEa. Conclusion:The candidate reference method for serum progesterone ID-LC/MS/MS established in our laboratory meets the requirements of the reference method. BCM has better detection performance than classical calibration curve method. The precision and linearity of the five progesterone clinical detection systems are satisfactory. The five clinical detection systems could meet the clinical requirements at the medical determination level of 25 ng/ml, however, only two of the five clinical detection systems meet the clinical requirements at the medical determination level of 2 ng/ml.
ABSTRACT
Objective:To establish a candidate reference procedure for the enumeration of cell particles in urine and applied to the multi-center performance evaluation of an automated urine formed elements analyzer.Methods:According to the standardized mannual microscopic examination of fresh non-centrifuged urine samples and the recommended reference method for enumeration of cell particles in urine published by ISLH, we established a candidate reference procedure for the enumeration of cell particles in urine. From four class A tertiary hospitals′ clinical laboratories, three rigorous trained technicians per hospital tested the same specimen respectively using the reference procedure. Each specimen was repeatedly counted 5 times, obtaining the quantitative results of cell particles were obtained in urine. Four hospitals used the established candidate reference measurement procedure and the automated urine formed elements analyzer to detect 40 to 60 urine specimens from September 2020 to January 2021, and evaluate the established reference method, meanwhile evaluate the accuracy and consistency of the each count from automated urinalysis analyzer.Results:Using the candidate reference measurement procedures, the coefficient of variation of results derived from three trained technicians per hospital was less than 6.98% (red blood cells), 6.99% (white blood cells), 13.94% (epithelial cells) and met the quality requirements. The performance evaluation results of automated urine formed elements analyzer showed that the accuracy of red blood cells, white blood cells and epithelial cells met the requirements (bias≤4.98%) and was well consistent with the reference measurement procedure ( R2≥0.989). Conclusions:A candidate reference measurement procedure for the enumeration of urine cell particles was successfully established with satisfactory precision and accuracy. This procedure was applied to multicenter performance evaluation of an automated urine formed elements analyzer with good accuracy and consistency.
ABSTRACT
INTRODUCCIÓN: el aumento de la incidencia de las micosis ha generado la necesidad de desarrollar técnicas in vitro para el estudio de la susceptibilidad a los antifùngicos; El documento CLSI M27-A2 es el método de referencia para estudios de sensibilidad en levaduras. No obstante, este no subsana las necesidades de rutina de los laboratorios, principalmente por ser laboriosos; en consecuencia, métodos alternativos surgen ante la necesidad de contar con técnicas más sencillas, una de ellos es el ATB FUNGUS 3 que permite determinar la sensibilidad de Candida frente a diferentes antifùngicos. OBJETIVO: validar el método comercial ATB FUNGUS 3, frente al método de referencia M27-A2, con el fin de conocer su valor diagnóstico. MATERIAL Y MÉTODOS: se determinó la eficacia del método a través de parámetros de test diagnóstico; además, se evaluó la sensibilidad de 50 cepas de Candida albicans frente a Fluconazol (FLZ) e Itraconazol (ITZ) mediante el método comercial y el de referencia. RESULTADOS: se encontró que el ATB - FUNGUS 3 presenta una especificidad para FLZ de 100%, sensibilidad de 91%, valor predictivo positivo (VPP) de 56%, valor predictivo negativo (VPN) de 100%, con una eficacia diagnóstica de 92%, calculados para un intervalo de confianza (IC) de 95%; para ITZ la especificidad y sensibilidad fue de 88 % y 90% respectivamente, con un VPP de 64%, un VPN de 97%, eficacia diagnóstica de 90%, IC 95%. Para las pruebas de concordancia, el índice Kappa para FLZ e ITZ fue de 0,67 y 0,68 respectivamente. La prueba de Likelihood ratio para el FLZ fue (LR+) de 11,25 mientras que el (LR-) fue 0; para el ITZ (LR+) de 9,19 y el (LR-) fue 0,14. Reproducibilidad de 90 % (FLZ) y 85% (ITZ). CONCLUSIONES: el ATB FUNGUS 3, es una técnica rápida, de fácil realización y reproducible; pero el desempeño global de la técnica, sugiere que aún no es confiable para el diagnóstico en laboratorios, debido a los valores bajos obtenidos en los VPP, que indican que se podría derivar en errores al momento de determinar una cepa como sensible o resistente, punto importante al momento de decidir la conducta terapéutica.
INTRODUCTION: the higher incidence of mycoses has generated the need to develop in vitro techniques for susceptibility study to antifungal agents. CLSI M27-A2 is a reference method for yeast susceptibility studies. However, this method does not meet the needs of routine laboratories because it is difficult to follow all the processes. Consequently, alternative methods arise due to the need for simpler techniques. Then, one of them is ATB FUNGUS 3 which allows determining Candida's sensitivity to different antifungal agents. OBJECTIVE: validate the commercial method ATB FUNGUS 3 compared with the reference method M27-A2 in order to know its diagnostic value. MATERIAL AND METHODS: efficacy was determined by diagnostic test parameters. Moreover, sensitivity of 50 strains of Candida albicans at Fluconazole (FLZ) and Itraconazole (ITZ) was evaluated by the commercial and reference methods. RESULTS: ATB - FUNGUS 3 presents a specificity for FLZ of 100%, sensitivity of 91%, positive predictive value (PPV) 56%, negative predictive value (NPV) 100% with a diagnostic efficacy of 92%, calculated for a 95% confidence interval (CI). For ITZ, the specificity and sensitivity were 88% and 90% respectively, with a PPV 64%, a NPV 97% with a diagnostic efficacy of 90%, 95% CI. For the concordance tests, the Kappa index for FLZ and ITZ was 0.67 and 0.68 respectively. The Likelihood ratio test for FLZ was (LR +) of 11.25 while the (LR-) was 0; for ITZ (LR +) of 9.19 and the (LR-) was 0.14. Reproducibility of 90% (FLZ) and 85% (ITZ). CONCLUSIONS: the ATB FUNGUS 3 is a fast, easy and reproducible technique. However, the overall performance of the technique suggests that this method hasn't been reliable for diagnostic laboratory yet, because PPVS obtained low values. These PPVS indicate that it could lead to errors when determining a strain as sensitive or resistant. This is an important point when deciding the therapeutic conduct.
Subject(s)
Candida albicans , Antifungal Agents , Confidence Intervals , DiagnosisABSTRACT
Chronic kidney disease caused by diabetes and hypertension has become a global public health problem. Urinary albumin is one of the key indicators for the diagnosis and grading of chronic kidney disease, and it has been widely used in practice. Accurate and comparable results of urinary albumin detection served as an important basis for clinical application. This paper discussed the status of urinary albumin detection and the progress of standardization, in order to help the laboratory correctly interpret the test results and promote the quality of diagnostic products.
ABSTRACT
Objective To select the calibrator for the conventional measurement system of serum a-Amylase (Amy).Methods The Amy levels of forty frozen serum samples were detected by the IFCC reference method (reference system),the conventional system A which used the Bioassay routine reagent and Randox calibrator,and was calibrated by the Roche PNPG7 method,and the conventional system B which used the Bioassay routine reagent and Randox calibrator,and was calibrated by the Rondox liquid stable PNPG7 method,respectively,and the acceptability of the two conventional systems was evaluated.Results The regression equations of the measurement values between the IFCC reference method and the conventional systems A and B were Y =0.964X +0.376 and Y =1.095X + 3.131,respectively.Among them,X and Y represented the results of the IFCC reference method and the conventional system,respectively.Compared with the IFCC reference method,the results of the conventional system A was reliable.Condusion With the guidance of the IFCC reference method,the domestic biochemical reagents matched with the suitable calibrators may provide the acceptable results.
ABSTRACT
Important objectives of external quality assessment (EQA) is to detect analytical errors and urge laboratories to take corresponding corrective actions.The paper described knowledge required to interpret EQA results and present a structured approach on how to handle unacceptable EQA results.The interpretation of EQA results depends on five key points:the control material,the target value,the number of replicates,the acceptance limits and between lot variations in reagents.When there are unacceptable EQA results,these factors may be the sources of errors.The ideal EQA sample has two important properties:having no matrix effects;having a target value established with a reference method.If either of these two criteria is not entirely fulfilled,results not related to the performance of the laboratory may arise.To help and guide the laboratories in handling an unacceptable EQA result,National Center for Clinical Laboratories has developed a preliminary investigation on the sources of errors and corrective actions for nonconforming EQA results in fifteen EQA schemes.Then a flow chart with additional comments was developed based on the investigation and the document of QMS24 to help laboratories improve quality by use of EQA results.
ABSTRACT
Objective To summarize and review the comparative results of the isotope dilution mass spectrometry in Cholesterol Reference Method Laboratory Network (CRMLN)project of the Centers for Disease Control and Prevention (CDC)of the United States in order to provide quality controls for determination of blood lipid.Methods The isotope dilution liquid chromatography tandem mass spectrometry (ID-LC/MS/MS)methods for determination of serum total cholesterol (TC)and triglycerides (TG)were developed to participate in the comparison of CRMLN.The survey was conducted every three months before 2016 and every half year from 2016.Four kinds of reference materials with two parallel tubes for each material and each tube in duplicate were determined in every survey.At least two certified reference materials used as quality control samples were analyzed simultaneously in each determination.Results In the 15 comparisons the CV of the TC determination method in our laboratory was 0.43％ while the CV of all the participated laboratories in CRMLN was 0.42％.The bias from the overall mean value in our laboratory was 0.22％ while the bias from the CDC target values was 0.58％.The CV of the TG determination method in 15 tests of our laboratory was 0.62％ and the bias was-0.98％ from the overall mean value and-0.80％ from the target values of CDC.Among the 60 results for comparison,98％ (59/60)of CV in the TC determination met with CDC requirement for precision (CV≤ 1％),and 70％ of bias (42/60) of the results met with CDC requirement for accuracy (bias ≤ 1％).For the 60 results in comparison of TG determination,92％ (55/60)of bias of the results met with the accuracy requirement of CDC (bias ≤2.55％).Conclusion In CRMLN comparison the results of TC and TG determined by ID-LC/MS/MS method were consistent with the values which were certified by CDC and determined by other network laboratories.The comparative surveys may play an important role in the standardization of lipid determination,and should be expected to provide experiences and technical supports in the comparative plan for reference measurement in our country.
ABSTRACT
Objective To explore achieving the consistent method of blood lipid examination by comparing the results of 5 dif-ferent blood lipid detection system commonly used in the use of refernce method to assign freach blood serum before and af-ter calibration.Methods Used the indoor quality control total variation (CV%)to evaluate the 5 blood lipid examination system of the imprecision.Referenced the United States Clinical and Laboratory Standardization Institution (CLSI)9A2 EP program,compared with 54 fresh blood serum in 5 commonly used examination system of Total Cholesterol (TC)and Tri-glyceride (TG),and then estimated the bias between the different detection systems and mean value.8 of the samples were determined by the reference method and estimate the bias of different system.The fresh frozen serum samples assigned by reference method were used to evaluate the above examination system,then compare and estimate the bias again with the same 54 fresh serum samples.Compared the variation of 54 samples in different detection system before and after calibra-tion.Results The TG imprecision of 5 examination system were between 3.76%~23.65%,the TC imprecision between 2.19%~23.43%,that mean the results were good,the r value of TG were between 0.996 7~0.999 6 and the TC were 0.956 2~0.996 7.But there were obvious differences between the results of the systems,and the biggest difference were 14.72%~34.21% in TG and 3.11%~14.57% in TC.After use the serum assignment by reference method,the variation of the systems has been significantly decreased.Conclusion Using the reference method to assign the fresh serum of different blood lipid detection system can effectively improve the consistency of the results.
ABSTRACT
Objective To develop an uncertainty study model of the enzymological reference method to systematically research the uncertainty influencing factors of 6 enzymological reference methods ALT,AST,LDH,AMY,GGT and CK for determining the key factors influencing the reference method and better guiding the establishment and operation of the enzymological reference methods.Methods The uncertainty of the six enzymological reference methods was evaluated by the theoretical evaluation com-bined with the experiment design according to series of standards including JJF1059-1999 Evaluation and Expression of Uncertainty in Measurement,JJF1135-2005 Evaluation of Uncertainty in Chemical Analytical Measurement,CANA-GL06 Guidance on Evalua-ting the Uncertainty in Chemical Analysis and CNAS-CL33 Guidance on the Application of Testing and Calibration Laboratories Competence Accreditation Criteria in the Field of Clinical Enzymology Reference Measurement.Results The established enzymo-logical uncertainty study model was C=ΔAΔt · 1ε · 1L ·V 1 +V 2 +V SV ·fT ·f pH ·fλ·fkit ·fresult .Conclusion The uncertainty e-S valuation model is set up and successfully applied in the uncertainty evaluation of the six enzymological reference methods.
ABSTRACT
Objective To develop a candidate reference method for the determination of serum creatinine and to evaluate the ac-curacy of conventional detection systems though method comparison to achieve traceability .Methods The candidate reference method was established according to the sarcosine oxidase and the accuracy and reliability of the method was verified through par-ticipation in international reference laboratories EQA activities (IFCC-RELA) .20 fresh single human serum samples with different concentration and calibrator were simultaneously measured by using conventional detection system and candidate reference method . Results The calibration curve for serum creatinine was linear in the concentration range from 50-2 000 μmol/L with a correlation coefficient of 0 .999 9 under the optimum experimental conditions (the linear equation was Y=0 .000 884 2X-0 .000 325 3) and the imprecision was less than 1 .0% .The proposed method has been applied to the determination of RELA samples with satisfactory re-sults .The measured results with conventional detection systems were consistent with candidate reference method ,and the slope of the regression equation was 1 .005 6 .Conclusion The candidate reference method of serum creatinine is successfully established and which can be used for traceability and standardization .It may provide an effective way for conventional detection system traceable to the reference method or reference material .
ABSTRACT
Objective To discuss the feasibility of enzymatic reference methods in Routine Chemistry external quality assessment (EQA)inlaboratorymedicine.Methods Samplesofthe1stEQAin2012byNationalCenterforClinicalLaboratories(NCCL)and patients′sera were measured by reference methods and 5 clinical analytic systems for the catalytic activity of CK ,LDH ,ALP ,ALT , AST ,GGT and AMY ,then the results of 5 clinical systems were compared with the reference methods′or target value of NCCL by calculating the bias ,and evaluated them according to the criteria of EQA by NCCL .Results The results of EQA samples measured by reference methods was within ± 10% compared with NCCL target value .Compared with the results of reference method ,the through put was 100 .0% for wet clinical chemistry systems measuring both EQA samples and patients′serum ,and the dry clinical chemistry systems was 77 .1 for EQA samples and 97 .1% for patients′serum according to the criterion of EQA ,and the through put was 72 .9% and 63 .6% of wet clinical chemistry systems according to the standard of enzymatic trueness of NCCL .Conclusion Reference method could be applied to EQA ,and will be a great help for the trueness of clinical testing .
ABSTRACT
Objective To establish a candidate reference method for the determination of serum lithium based on ion chromatog-raphy and evaluate its analytical performance.Methods A simple sample treatment procedure,which can be remove the proteins and/or organics in human serum,has been developed for the determination of serum lithium.Method precision was evaluated with different concentration of fresh human serum and EQA sample RELA-A/B.Method accuracy was investigated with the recovery ex-periments in fresh human serum and RELA-A/B sample.Results The linear equation was Y =0.817 1X -0.001 3 with a correla-tion coefficient of 0.999 95 under the optimum experimental conditions,the detection limit (3S/N)for lithium was 6 μg/L and the imprecision was less than 1.0%.The results of the recovery experiments indicated that the recoveries were reasonable for the deter-mination of serum lithium,in a range of 99%-101%.Conclusion The candidate reference method of lithium was successfully es-tablished and which can be used for traceability and standardization.It may provide an effective way for routine testing of lithium traceable to the reference method/reference material.
ABSTRACT
Background: There are several methods for counting platelets, of which the international fl ow reference method (IRM) is considered to be the gold standard. We compared the platelet count given by this method to the count given by automated analyzers using other methods, such as optical fl uorescence and impedance. Aims: The aim of this study is to compare the platelet counts obtained by Sysmex XE 2100 by Impedance (Sysmex-I), optical fl orescence (Sysmex-O) and reported (Sysmex-R) based on the switching algorithm and LH-750 by Impedance (LH-750) with the IRM in thrombocytopenic blood samples. To calculate the sensitivity, specifi city, positive predictive value (PPV) and negative predictive value (NPV) of various technologies at the clinically relevant transfusion thresholds of 10 × 109/l and 20 × 109/l. Materials and Methods: A total of 118 blood samples with platelet count of <50 × 109/l were selected for the study. Platelet counts of all samples were analyzed by all methods using the Sysmex analyzer, LH-750 and IRM in parallel within 6 h of collection. Statistical Analysis Used: Pearson correlation, bland Altman analysis, sensitivity and specifi city, PPV and NPV. Results and Conclusions: Sysmex-R had the least Bias and 95% limits of agreement (95%LA) range and thus correlated best with IRM values. LH-750 had a higher Bias compared to Sysmex-O and Sysmex-R, but a strikingly similar 95% LA ensures similar results in all three methods. In fact, in the oncology subset, it had the narrowest 95% LA, which made it the best performer in this subgroup. Of the three Sysmex results, Sysmex-I had the highest bias, widest 95% LA and highest potential risk of over transfusion. Hence, Sysmex-R and LH-750 were found to be reliable tools for estimation of platelet count in thrombocytopenic patients.
ABSTRACT
Characteristics and measurement principles of reference methods in clinical biochemistry were described.Implementation of reference systems is one of the most effective approaches to improve the accuracy and comparability of clinical laboratory test results.Reference methods are the key components of reference systems.Reference methods should have measurement uncertainties that meet the requirements of the intended use,and thus should be based on reliable measurement principles.For the well-defined biochemistry analytes,reference methods have been almost all based on instrumental analysis.Isotope dilution mass spectrometry (ID/MS) is considered most reliable and has been the major analytical principle of the reference methods.ID/MS analysis is accurate but expensive.Use of other validated instrumental analyses as reference measurement principles would be justified.
ABSTRACT
Objective To investigate the efficacy of using a calibrator with values assigned with the reference method for improving the comparability of serum gamma-glutamyltransferase (GGT) measurements.Methods The IFCC reference method for GGT was established and the performance was verified by testing a certified reference material (CRM).A calibrator was prepared and its value for GGT was assigned with the reference method.Forty serum samples were measured on different (including HITACHI 7600,7060,7170,7180 and BECKMAN LX20,OLYMPUS AU 400) chemistry analyzers with Zhongsheng GGT reagent kits calibrated with the calibrator.The samples were also measured on the same analyzers using a theoretical factor.Biases of results obtained with the calibration and with the theoretical factor based calculation were compared.Results The reference method resuhs on the CRM agreed the certified value within the stated uncertainty.Serum results calculated from the theoretical factor showed various biases and inter-analyzer variations.When the analyzers were calibrated with the calibrator,the number of results with biases less than 10% became significantly higher and those with biases more than 20% significantly lower.The variation of the results on 5 serum samples was reduced from 11.0%～14.0% to less than 5% by using the calibrator.Conclusion Accuracy and comparability of GGT measurements with of ZhongSheng GGT kits can be improved by using a calibrator that has a value assigned with the reference method.
ABSTRACT
BACKGROUND: Hospital infection surveillance is a dynamic process for gathering, managing, analyzing, reporting and re-evaluating the data. Recently there has been an increased awareness of the importance of hospital infection surveillance and management program in Korea. The most ideal way among the hospital infection surveillance systems is known to be the "reference method". In this method all hospital patient records and charts are reviewed and the infected patient are investigated in daily basis. However it requires enormous efforts to apply this method in hospitals with limited personnel resources. Although the number of the hospital having full-time hospital infection control nurses has been increased considerably in Korea the effective hospital control programs have not been established yet in most hospitals owing to the lack of full-time hospital infection control nurses. Nevertheless it became indispensable to develop an alternative hospital infection surveillance program that is readily available. This study was carried out to investigate epidemiologic characteristics, and assess the efficiency and validity of ward liaison surveillance method for nosocomial infection surveillance in a general hospital without full-time infection control nurses. METHOD: During the period of the study, from March 1 to March 31, 2000, cases of hospital infection collected by two different methods, reference method and ward liaison nurse surveillance, were compared. The validity of ward liaison surveillance data was examined using the data collected by the reference method as gold standard. RESULT: In the data collected by the reference method, 94 cases of hospital infection were identified whereas 83 cases by the ward liaison nurses. The incidence rate of hospital infection was 9.5% during one month; the incidence rates were higher in males (12.6%) than female (6.7%) and in age group of 50s. The incidence rates by ward were 38.8% in intensive care unit, 45.5% in neurosurgery, 18.6% in neurology ward, 12.8% in internal medicine, 10.6% in orthopedic ward, and 8.6% in general surgery. Sites of hospital infection in the order of decreasing frequency were urinary tract (24.8/1000 discharge patients), lung (22.2), wound (18.2), and other respiratory systems (15.2). The type of microorganisms isolated were 16: three gram-positive bacteria, eleven gram-negatives and two fungi. Staphylococcus was the most frequently isolated organism, 21 strains, among which 17 strains were methicillin-resistant Staphylococcus aureus (only one strain was sensitive to methicillin) and three strains were methicillin resistant Staphylococcus epidermidis. Seventeen strains of Pseudomonas aeruginosa were isolated from pneumonia, urinary tract, and wound. Escherichia. coli, Serratia marcencecs, Acinetobacter baumannii, Klebsiella pneumoniae, Enterococcus faecalis, E. faecium, Enterobacter cloacae, Streptococcus pneumoniae, and Candida albicans were also isolated. There were twenty-two specimens that revealed no growth of any organisms. In the ward liaison nurse surveillance method, the number of false positive hospital infection was eleven cases and the false negative was 22 cases. The validity evaluated by four different measurements were sensitivity 76,7%, specificity 98.7%, positive predicted value 86.7%, negative predicted value 97.5%. Thus the ward liaison nurse surveillance method was shown to be a valid method with high efficiency. The false positive and false negative cases were mainly occurred by the deficient knowledge in the definition of hospital infection, and deficient skills of investigating the patient's symptoms and clinical course; the liaison nurses had not checked all the surgical site resulting in low sensitivity in surgical site infection. CONCLUSION: According to the results, the epidemiologic characteristic of hospital infection in this particular community hospital studied was not much different from other study results; the incidence rate of hospital infection for one month was 9.5%. On the other hand the ward liaison nurse surveillance method was shown to be satisfactory in detecting hospital infection. This could be a useful method for hospitals without full-time infection control nurses. Furthermore, the validity of this method could be improved by accumulation of the knowledge and skills on hospital infection surveillance through a well planned on-the-job training program for the nurses.