A new robust statistical model for interpretation of differences in serial test results from an individual

Abstract

The proposed statistical approach may be more appropriate in assessing difference between serial measurements when individual data are not normally distributed.

Defining analytical performance specifications 15 years after the Stockholm conference

“…Thereafter Carobene describes the heterogeneity of available biological variation data [7]. In general, any time an estimate of intra-individual CV is > 33.3%, it is likely an indication that the distribution of the individual variances is not Gaussian and therefore the parametric statistical handling is not appropriate [8,9]. Despite the questions regarding the reliability of information, this author confirms the importance of a biological variation database that should, however, include only products of appropriately powered studies.…”

Harmonization initiatives in the generation, reporting and application of biological variation data

“…differently sized RCV for a decrease and an increase in the measurand concentration: This approach is also superior to classical RCV calculation in that it is not possible to achieve paradoxical decreases greater than 100% [24]. For the best performance, this approach assumes log-normal distributed data or estimates of CV I <12% [25,26]. A rearrangement of the RCV equation can make z the unknown so that, for any change in serial results detected, using the estimates of CV I and CV A , the probability that the seen change is significant can be calculated [2].…”