CLSI EP14 A3 : 2014

Abstract
Committee Membership
Foreword
1 Introduction
2 Commutability Determination Process
3 Conclusion
4 Supplement Information
References
Appendix A - Description of Mathematical Model Used
             of Evaluating Commutability of Processed Samples
             Using Deming Regression
Appendix B - Outlier Evaluation for a Measurement
             Procedure Comparison Using Deming Regression
Appendix C - Examples of Completed Analyses
The Quality Management System Approach
Related CLSI Reference Materials

Gives guidance for evaluating the commutability of processed samples by determining if they behave differently than unprocessed patient samples when two quantitative measurement procedures are compared.

This guideline provides protocols that can evaluate commutability in any nonpatient processed samples when tested using quantitative measurement procedures. Such processed samples may be used for proficiency testing/external quality assessment (PT/EQA), measuring interval verification sample sets, or QC samples.The guideline is intended to be used by developers of commercial diagnostic tests as well as laboratory-developed tests, manufacturers of measuring interval sample sets and QC samples, and PT or EQA providers. This guideline may also be useful to all clinical laboratory professionals wishing to investigate a processed sample’s commutability. EP14 is intended to assist in the education of clinical laboratorians, regulators, and diagnostic manufacturers about the commutability of processed materials, and how a sample’s matrix can affect some measurand values and their interpretation (referred to as matrix effects). For example, professionals may not be warned of a matrix effect caused by the interaction of processed PT/EQA material and the measurement procedure, and therefore the data may suggest to them that erroneous patient results are being generated, when in fact the results may be acceptable. Examples of a matrix effect due to the interaction of a processed QC and certain reagent lot(s) exist in the literature.1 Therefore, these types of effects should not be a surprise to experienced laboratory staff and should not lead to erroneous conclusions about the same effect occurring in patient samples. This guideline should assist all interested parties in not only evaluating the presence or absence of a matrix effect, but also increasing awareness that there may be different levels of risk to the quality of patient care that are dependent on the intended use of a processed matrix.This guideline can also be used by laboratorians performing quantitative tests for a wide variety of measurands across various disciplines to understand the commutability of processed samples. This guideline does not apply to qualitative tests. Finally, an added benefit to following the protocol is that manufacturers and PT/EQA providers should be able to provide some documentation to government or accrediting agencies on processed samples commutability to help avoid false conclusions about the adequacy of patient testing.It should be noted that although the protocol in this document is intended to help distinguish between effects caused by measurement procedure malfunctions and those caused by use of artificial or human-based processed samples, it does not describe approaches that specifically establish the exact mechanism or reason for any observed noncommutability. This guideline does not apply to qualitative tests that supply only “yes/no” or “positive/negative” results.Also, it should be noted that this document is not intended to be used to evaluate sample type differences, such as serum vs plasma.