Assessing quality on the Sigma scale from proficiency testing and external quality assessment surveys.

Abstract

There is a need to assess the quality being achieved for laboratory examinations that are being utilized to support evidence-based clinical guidelines. Application of Six Sigma concepts and metrics can provide an objective assessment of the current analytical quality of different examination procedures. A "Sigma Proficiency Assessment Chart" can be constructed for data obtained from proficiency testing and external quality assessment surveys to evaluate the observed imprecision and bias of method subgroups and determine quality on the Sigma scale. Data for hemoglobin A1c (HbA1c) from a 2014 survey by the College of American Pathologists (CAP) demonstrates that approximately two-thirds of the examination subgroups provide only two-Sigma quality when evaluated against the CAP requirement of an allowable total error of 6.0%. The weighted averages were 1.46 Sigma for a survey sample with an assigned value of 6.49% Hb (average bias 2.31%, CV 2.87%), 1.45 Sigma at 6.97% Hb (average bias 2.29%, CV 2.81%), and 1.75 at 9.65% Hb (average bias 1.55%, CV 2.71%). Maximum biases for examination subgroups were 5.7%, 5.8%, and 4.1%, respectively. Assessment of quality on the Sigma scale provides evidence of the analytical performance that is being achieved relative to requirements for intended use and should be useful for identifying and prioritizing improvements that are needed in the analytical quality of laboratory examinations. In spite of global and national standardization programs, bias is still a critical limitation of current HbA1c examination procedures.