摘要目的 利用患者数据均值和百分位数建立临床生化日常室内质量控制程序.方法 收集2016年1月至2017年3月在浙江省人民医院就诊的门诊和住院所有患者生化项目检测结果,统计出每日患者数据的算术平均数(Xˉ)和百分位数(P2.5、P5、P10、P25、P50、P75、P90、P95和P97.5),计算出这些统计点15个月累积变异系数(CV统计).将CV统计与不同项目质量标准比较,筛选出CV统计小于或接近质量标准的生化项目和浓度控制点;使用PASS11.0软件计算选择Xˉ为控制点项目的最低样本量,同时比较患者数据控制点和同期质控品CV.最后在实验室信息系统(laboratory information system,LIS)基础上设计质控程序,并使用临床患者数据进行验证.结果 门诊患者中,AFU、APOA、APOB、CA、CL、HDL、K、MG、NA、NEFA、TP和URIC可以采用作为质量控制点,所需统计最低样本量分别为23、23、30、8、10、24、34、8、8、20、13和22;ALP和TBIL采用P25作为控制点;AST、GLU、GPDA和PHOS采用P50作为控制点;ALB、CHE、CREA和DBIL采用P75作为控制点.住院患者中,AFU、ALB、APOA、APOB、CA、CL、HDL、K、Lpa、MG、NA、NEFA、TP和URIC可以采用Xˉ作为质量控制点,所需统计最低样本量分别为73、19、34、18、10、30、36、21、87、12、17、51、26和52;ALP、ALT、AST、CREA、DBIL、LDH、TBIL和TG采用P25作为控制点;PHOS采用P50作为控制点;GPDA采用P75作为控制点;CHE采用P90作为控制点.采用百分位数作为控制点的所需最低样本量为200.筛选出的控制点CV多数大于同期质控品CV.在设计患者数据质控程序中绘制L-J图和Z分数图帮助判断是否出现系统性偏移.结论 最终在LIS系统中建立了基于患者数据的质控程序.采用合适的患者数据来进行质量质控,是一种经济实用的方式,能弥补现有质控品质控的不足.

abstractsObjective Quality control procedure based on the patient data in clinical chemistry was set up in laboratory information system (LIS). Methods Clinical chemistry tests results of outpatients and inpatients were collected from January 2016 to March 2017 in Zhejiang Provincial People's Hospital. Statistical results of daily patient data, including Xˉ, P2.5, P5, P10, P25, P50, P75, P90, P95 and P97.5 were calculated. Secondly, cumulative coefficients of variation (CV) of these statistical datawere calculated and compared to different criterions. Optimal analytes and related control concentrations were chosen. The minimum number of patient sample which use Xˉ as control point was calculated by PASS 11.0 software. Finally, the quality control procedure was set up base on the LIS and was verified by patient data. Results In outpatients, Xˉwas chosen as control point in AFU, APOA, APOB, CA, CL, HDL, K, MG, NA, NEFA, TP and URIC and the minimum number of sample needed were 23, 23, 30, 8, 10, 24, 34, 8, 8, 20, 13 and 22. P25 was chosen in ALP and TBIL. P50 was chosen in AST, GLU, GPDA and PHOS.P75 was chosen in ALB, CHE, CREA and DBIL. In inpatients, Xˉ was chosen as control point in AFU, ALB, APOA, APOB, CA, CL, HDL, K, Lpa, MG, NA, NEFA, TP and URIC and the minimum number of sample needed were 73, 19, 34, 18, 10, 30, 36, 21, 87, 12, 17, 51, 26 and 52;P25 was chosen in ALP, ALT, AST, CREA, DBIL, LDH, TBIL and TG. P50 in PHOS, P75 in GPDA, and P90 in CHE. 200 samples were needed in the tests which used percentiles as control points. Most CVs of these control points were higher than the commercial quality control used every day. Finally, a quality control procedure based on patient data were set up in LIS. L-J and Z score charts were used to find out systematic bias. Conclusion Patient data used in internal quality control was an economical and practical way, which can make up for the deficiency of traditional method.