Establishing and maintaining performance claims. A manufacturer's viewpoint

Performance claims are the vehicle by which a manufacturer communicates the analytic capabilities of its methods to laboratory users and to regulatory agencies. Claims describe the expected performance of an analytic system. To be useful, claims must be meaningful, achievable, and verifiable. To ensure consistent interpretation, they must be stated in clear, unambiguous terms. There are no specific guidelines for stating claims. Consequently, comparing the performance of one system to another is difficult. Claims fall into three categories: clinical utility, analytical performance, and boundary conditions. This article focuses on the three basic analytical performance characteristics, precision, accuracy, and specificity, and describes how they are established, validated, maintained, and monitored by manufacturers and how they can be verified by end users. Precision claims describe the inherent variability of the system. They differ in the components of variance included, ranging from the short-term variables within a single run to the long-term variables experienced over time, such as lot changes, calibrations, instrument drift, and environmental fluctuations. Precision may be stated either as typical user experience or as the minimum performance expected. Accuracy claims describe the degree to which the results will approximate the true values. They are more difficult to establish and verify, given the lack of objective standards for many analytes. For this reason, few manufacturers make true accuracy claims, relying instead on comparisons with other marketed methods. A notable exception is cholesterol, for which a model reference system has been established to help laboratories ensure accuracy. External proficiency testing is handicapped by matrix effects from imperfect survey fluids. Specificity claims describe the method's freedom from interference and cross-reactivity. They are based on screening tests by the manufacturer, augmented by feedback from users. There are no agreed-on criteria for what constitutes clinically significant interference, and no consistent approach for disclosing interference information. The National Committee for Clinical Laboratory Standards, Villanova, Pa, is beginning to develop guidelines to promote greater consistency in performance claim statements.