Sigma Metrics misconceptions and limitations.

Abstract

Objectives: This paper further explores the Sigma Metric (SM) and its application in clinical chemistry. It discusses the SM, assay stability, and control failure relationship.

Content: : SM is not a valid measure of assay stability or the likelihood of failure. When an out-of-control event occurs for an assay with a higher SM value, the same QC rule will have greater power to detect error than assays with a lower SM value. Thus, it is easier to prevent errors from happening for higher SM assays. This rationale encourages using more frequent QC events and more QC samples for a QC scheme of a low SM assay or simply more QC cost for low SM assays. A laboratory can have a high-precision instrument that frequently fails and a low-precision instrument that hardly ever fails. Parvin's patient risk model presumes the bracketed continuous mode (BCM) testing workflow. If overlooked when designing QC schemes, this leads to the common misconception of the SM that one can save the cost of QC since assays with high SM require less frequent QC to ensure patient risk. There is no evidence that an assay's precision is correlated with its failure rate. Schmidt et al., in a series of papers, showed that an assay with a higher P_f or shift in probability will have a higher expected number of unacceptable results. Incorporating P_f into the QC design process presents significant challenges despite the proactive quality control (PQC) methodology.

Summary: Unfortunately, TEa Six Sigma, as widely practiced in Clinical Chemistry, is not based on classical Six Sigma mathematical statistics. Classical Six Sigma would facilitate comparing results across activities where the principles of Six Sigma are employed.