Background: According to the ISO 15189 standard, section 5.5.1.4, “The laboratory shall determine measurement uncertainty (MU) for each measurement procedure in the examination phase used to report measured quantity values on patients’ samples. The laboratory shall define the performance requirements for the measurement uncertainty of each measurement procedure and regularly review estimates of measurement uncertainty.” Section 5.5.1.2 of the standard also specifies that “The laboratory shall obtain information from the manufacturer/method developer for confirming the performance characteristics of the procedure without modification.” From this point of view, Bayesian predictive inference is an especially well-adapted approach for IQC management, as several reports have made clear.

Aims: Taking factor-II assay as an example, the present study aimed to explain the usefulness of Bayesian inference applied to IQC results analysis, as a means both of calculation and of assessing the acceptability of estimated MU.

Methods: The Bayesian approach provides a predictive distribution of the future observable. This sequentially updated distribution can be used to determine a region where the future observable will (most likely) be, as long as no special causes are present. Using the predictive distribution, a 100(1-a) % Highest Predictive Distribution (HPrD) region, defined by the lower and upper PCC limits (figs 1 and 2), estimates MU with 95% probability if the chosen a value is 5 % (Tsiamyrtzis et al., Blood Coagulation and Fibrinolysis 2015; 26:590-6).

Results: The manufacturer´s maximum acceptable coefficient of variation defines the maximum acceptable MU in terms of lower and upper range limits (fig. 1). In truly efficient laboratories, MU is significantly lower (fig. 2).

Conclusions: Bayesian inference is an appropriate means of estimating measurement uncertainty for methods not subject to modification.

[Figure 1: Bayesian estimation of uncertainty of measurement with the range Lower PCC limit – Upper PCC limit with respect to a minimal performance]

[Figure 2: Bayesian estimation of uncertainty of measurement with the range Lower PCC limit – Upper PCC limit with respect to a optimal performance]

« Back to ISTH 2020 Congress

ISTH Congress Abstracts - https://abstracts.isth.org/abstract/estimation-and-analysis-of-measurement-uncertainty-acceptability-the-interest-of-bayesian-inference/