Use of participant EQA results to assess sample homogeneity and stability for qualitative analytes

International requirements for PT and EQA state that providers must demonstrate that samples are homogeneous and stable. However, testing for homogeneity and stability can be expensive, use samples that could otherwise serve as quality-control materials, and can also fail to detect significant inhomogeneity and instability. In some situations it may be preferable to use the results from participants to identify problems with sample lots, if historic results follow predictable distributions and a statistical decision rule can be produced. An unusually high proportion of incorrect results may indicate that samples were inhomogeneous or unstable. Conditions under which this can be demonstrated are discussed, and the efficacy of the proposal is demonstrated with two examples. This procedure is especially effective when there are a large number of participants and/or a historic small proportion of incorrect results. Providers who adopt this proposal will need to retain samples for testing and assume the risk of distributing bad samples.Presented at the Eurachem PT Workshop September 2005, Portorož, Slovenia