Abstract Background Reagent lot variation is often considered a major source of concern for infectious disease serology testing. Regular review of quality control testing data for specific assays confirms that, when using traditional statistical methods, such as using small datasets to establish mean/SD and then applying Westgard rules, reagent lot variation is the most common cause of variation observed. To investigate every incident is time consuming, costly and likely unnecessary in the majority of cases. But how do we create an alternative to monitoring QC data that accepts normal reagent lot variation yet highlight unacceptable variation? Applying the results from two assays using the same multimarker QC sample to traditional and an alternative QC monitoring method such as QConnect highlights when we should be concerned with lot variation and when the variation is normal. Methods QC testing data from 14 lots of the multimarker QC “Optitrol Blue” over a three year period tested on two assays—Abbott ARCHITECT Anti-HCV (n = 54 102) and Abbott ARCHITECT HIV AgAb Combo (n = 62 682) were reviewed for lot variation using traditional and alternative QC monitoring methods. Results Numerous examples of lot variation were identified from the two assays over the three-year period. Applying traditional QC methods resulted in frequent recalculation of mean/SD to accommodate the lot variation, effectively making the lot variation part of the routine, negating its importance, and making the clinical significance of reagent lot variation difficult to ascertain. Using an alternative QC monitoring method such as QConnect, which incorporated normal lot in calculations to create a constant acceptance range, removed the uncertainty and highlighted instances where variation was outside ‘normal’, triggering further investigation by using meaningful QC ranges. A clear example was observed where two similar looking Levey-Jennings charts (one for ARCHITECT HCV and one for ARCHITECT HIV) demonstrated reagent lot variation that was detected as an outlier whereas Traditional methods identified all reagent lots as being outliers. Conclusion In infectious disease serology testing, reagent lot variation is frequent, normal, and rarely a cause for concern. However, it is difficult to determine how much variation is acceptable when using traditional QC methods, because most lot changes are flagged as “rejections”. QC processes that incorporate normal reagent lot variation, derived from historical data, allows for rapid identification of unacceptable variation, saving money and time (particularly when all reagent lot variation identified as outliers need unnecessary investigation), while reducing risk of poor results.
Read full abstract