運用Sigma Metrics優化生化檢驗項目品管流程

Utilizing Sigma Metrics to Optimize the Quality Control Process of Clinical Chemistry Tests

在臨床實驗室中sigma metrics是用來評估實驗室品質管理的表現，它代表偵測百萬次事件中發生錯誤的機率。Sigma metrics是定量項目整體表現的量化，藉由bias/ imprecision的計算及品管的監控，作為檢驗品質管理的方法。本研究計算於Hitachi機台32項之生化項目兩個品管濃度sigma表現，平均sigma在6以上有23項，sigma 5-6有5項，而sigma 4-5有4項。接著分析這些項目每小時的檢體量，並且利用bracketed QC方法及機台特性，重新檢討更適合的生化品管頻率。原本生化項目品管的操作頻率是白班每小時操作一次，小夜班和大夜班則是四小時操作一次，經由調整適當的品管頻率後，每日生化品管點數由原本636點下降為364點。但並非所有項目調降品管點數，其中hsCRP Level-2由每日7次提高到10次，Cystatin C則是由1次提高到2次。藉由優化生化檢驗項目品管流程，每年可節省試劑、品管液和人力成本約52萬1千元。綜上，本研究利用找出表現不好項目sigma metrics並將檢驗項目分級，在不影響檢驗品質下客製化品管頻率，並把心力用來處理sigma表現較差的檢驗項目上。

Sigma metrics symbolizing the defects per million are used for evaluating the performance of quality management in the clinical laboratory. The equation of a sigma metric is listed as following and all terms are represented in percentages: Sigma metric = (total allowable error%-Bias%)/CV%. In this study, we calculated the sigma metrics of 32 clinical chemistry items with two level controls on Hitachi. The performance of 23 items (71.9%) was above six sigma, 5 items (15.6%) were between five and six sigma and the last 4 items (12.5%) were between four and five sigma. Furthermore, we analyzed the sample size of those 32 items per hour and used the bracketed QC to review quality control frequency. In our laboratory, the previous quality control frequency of most clinical chemistry tests with good sigma performance was too high, which was once per hour in the daytime and once every four hours in the night. After adjusting the quality control frequency according to the sigma performance of each item, the number of total quality control of clinical chemistry items by Hitachi H008 analyzer declined from 636 to 364 per day. Meanwhile, not all items decreased frequency but 2 items increased frequency, including hsCRP and Cystatin C. The frequency of hsCRP increased from 7 times to 10 times per day, and that of Cystain C increased from one time to twice per day. By optimizing the quality control process, our laboratory would save the cost of reagent, quality control and the labors about 521,000 NT dollars per year. Above all, our study was to find out those items with poor sigma and customized the QC frequency, so as to make more efforts to deal with those items.