An example of a bootstrap technique application for judging a variability of method performance estimates and its consequences for quality control planning.

摘要

BACKGROUND: Flexible fitting of quality control procedures to method performance is an inseparable element of a cost-effective quality management strategy. However, opinions on method performance, based on a restricted number of quality control results used to calculate bias and imprecision, can vary due to the random nature of the data-gathering process. METHODS: We present an example of the bootstrap technique application in evaluating method performance estimates. RESULTS: Our analysis reveals substantial variability in method performance assessment results that should not be omitted. We suggest that the best way of using information concerning method performance is to consider the information in a probabilistic manner. Method performance is a feature presented with an element of probability. The effectiveness of a quality control procedure should be assessed with regard to this probability. CONCLUSIONS: Presented example of a bootstrap technique application increases our awareness of the uncertainty that accompanies method performance judgment and quality control planning.