静脉用药调配中心调配环境微生物动态监测数据的统计分析

目的为静脉用药调配中心( PIVAS)调配环境微生物动态监测提供控制限度制订方法和合适的限度值。方法使用复旦大学附属中山医院东院区静脉用药调配中心 2018年 4月至 2021年 3月调配环境微生物监测数据建立数据库，利用参数模型、阈值法及公差限制法等统计分析方法分别构建控制限度。通过应用效果的覆盖率误差及警报频率，评价限度制订的合理性以确定最佳方法。结果除参数模型外，阈值法与公差限制法成功构建了 PIVAS调配环境微生物动态监测的控制限度。阈值法: A级操作台沉降菌警戒限为 0.67、纠偏限为 <1；A级操作台浮游菌警戒限为 0.5、纠偏限为 <1；C级调配间沉降菌警戒限为2.75、纠偏限为 3.51；C级调配间浮游菌警戒限为 5.18、纠偏限为 6.29；在应用中警报 13次(频率为 5.47%)覆盖率误差为1.88%。公差限制法: A级操作台沉降菌警戒限为 <0.95、纠偏限为 <0.99；A级操作台浮游菌警戒限为 <0.95、限为 <0.99；C级调配间沉降菌警戒限为 47.5、纠偏限为 49.5；C级调配间浮游菌警戒限为 95、纠偏限为 99；在应用中警报 1次(频率为 纠偏，0.47%)，覆盖率误差为 2.58%。结论阈值法确定的控制限度可行性较强、灵敏度高，可作为 PIVAS调配环境动态微生物控制限度的制订方法。

A statistical analysis of microbiological dynamic monitoring data in the PIVAS deployment center

Objective To provide control limit formulation methods and appropriate limit values for microbiological dynamic moni-toring in the deployment center for Pharmacy Intravenous Admixture Services (PIVAS).Methods A database was established usingmicrobiological monitoring data, which were obtained from the deployment center for PIVAS in the East Campus of Zhongshan Hospitalaffiliated to Fudan University from April 2018 to March 2021. The control limits were constructed separately using statistical analysismethods such as parametric model, threshold method and tolerance limit method. The reasonableness of the limit formulation was evalu-ated to determine the best method by applying the coverage error calculation of the effect and frequency of alarms.Results In addition to parametric model, threshold method and tolerance limit method were successfully used to establish the microbiological dynamic con-trol limits for the PIVAS deployment center. The control limits determined by threshold method were as follows: the alert limit for set-tling microbe at the A-class operation table was 0.67 and action limit was less than 1; the alert limit for airborne microbe at the A-class operation table was 0.5 and the action limit was less than 1; the alert limit for settling microbe at the C-class deployment room was 2.75 and the action limit was 3.51; the alert limit for airborne microbe at the C-class deployment room was 5.18 and the action limit was 6.29. There were 13 alarms in the application (frequency being 5.47%), and the coverage error was 1.88%. The control limits deter-mined by tolerance limit method were as follows:the alert limit for settling microbe at the A-class operation tabble was less than 0.95 and action limit was less than 0.99; the alert limit for airborne microbe at the A-class operation table was less than 0.95 and the action limit was less than 0.99; the alert limit for settling microbe at the C-class deployment room was 47.5 and the action limit was 49.5; the alert limit for airborne microbe at the C-class deployment room was 95 and the action limit was 99. There was 1 alarm in the application(frequency being 0.47%), and the coverage error was 2.58%.Conclusion The control limit determined by the threshold method hasgreater feasibility and higher sensitivity, which can be used as a method to formulate the dynamic microbial control limit in PIVAS de-ployment center.