Continuous Improvement in Continuous Quality Control

“Quality is never an accident. It is always the result of intelligent effort.” So observed John Ruskin, Victorian writer and critic of art, architecture, and society (1), unknowingly foreshadowing the emergence of quality science in the 20th century. Statistical QC came from Bell Telephone Laboratories, where Shewhart developed the statistical control chart in 1924, and Dodge and Romig pioneered statistical methods for acceptance sampling rather than 100% inspection of products. These beneficial techniques were not widespread until the manufacturing demands of World War II necessitated better control of product quality. With Deming's involvement, quality science was further developed in postwar Japanese manufacturing (2). Levey and Jennings introduced clinical laboratory statistical QC in 1950 (3). The Levey–Jennings chart remains a staple of laboratory QC. Development of stabilized control materials enhanced QC practices (4), and the testing of controls at fixed intervals became and remains the QC mainstay. Control-based QC has many strengths, including sensitivity to small changes in bias and ready availability of controls. Test intervals may be adjusted to account for assay stability, regulations or a laboratory's experience. Decision rules may be selected to maximize both sensitivity and specificity of QC procedures (5, 6). Weaknesses of control-based QC include the cost of materials and labor, scope limited to the analytical component, and matrix-effect masking of clinically relevant drift. Furthermore, a risk inherent in the intermittency of control-based QC is that analytical instability goes undetected until the next QC episode, while the laboratory releases clinically significant erroneous results. Ideally, QC would be continuous. In a limited sense, continuous QC exists for sample QC through delta checks, anion gap calculations, indices of hemolysis, icterus, lipemia, etc., and for analyzer performance through function checks, temperature sensors, automated reaction pattern analysis, etc. In this issue of Clinical Chemistry , Ng et …