A narrative review about regulatory acceptability standards for clinical assays

Abstract

Objective: To review acceptance standards, particularly those used by regulatory agencies to approve products. A hierarchy of standards is discussed ranging from regulatory (FDA), quasi-regulatory (CLSI, ISO 15189, ISO 15197), to academic standards (Milan conference). Background: Many clinical chemistry assays produce results that can be compared to reference. This allows regulatory bodies such as FDA to have acceptance protocols that require evaluation results between the candidate and reference assay to meet certain acceptability limits. Methods: This paper analyzes the problems arising from acceptability standards including: (I) a generic problem with the standards; (II) protocols used to evaluate the standards; (III) how the data are analyzed; (IV) how often results are observed that potentially can cause serious patient harm, and (V) why people do not pay more attention to dangerous results. Conclusions: Suggested recommendations include: (I) specifications should better reflect the harm when the magnitude of error increases; (II) results should be provided with and without pre- and post-analytical error; (III) more focus is needed on tools that prevent large errors, especially if pre- and post-analytical errors are detected. These include improved user training, Failure Mode Effects Analysis (FMEA), fault trees and Failure Reporting And Corrective Action System (FRACAS); (IV) for assays on the market, the MAUDE database should be examined for adverse events.