Robust estimation of between and within laboratory standard deviation with measurement results below the detection limit

In method validation studies, the question arises how to deal with < LOD results in order to ensure reliable precision estimates. It is proposed to elaborate strategies for dealing with < LOD results on the basis of the Q/Hampel method, a particularly robust statistical method. Two different s...

Full description

Saved in:
Bibliographic Details
Published inJournal für Verbraucherschutz und Lebensmittelsicherheit Vol. 10; no. 4; pp. 385 - 391
Main Author Uhlig, Steffen
Format Journal Article
LanguageGerman
Published Basel Springer Basel 01.12.2015
Subjects
Agriculture
Announcements and Reports
Biomedical and Life Sciences
Biotechnology
Chemistry/Food Science
Food Science
Life Sciences
Plant Genetics and Genomics
LOD
Repeatability
Method validation
Censoring
Reproducibility
LOD results
Q/Hampel
Robust statistical methods
Q method
Online AccessGet full text

Cover

More Information
Summary:In method validation studies, the question arises how to deal with < LOD results in order to ensure reliable precision estimates. It is proposed to elaborate strategies for dealing with < LOD results on the basis of the Q/Hampel method, a particularly robust statistical method. Two different strategies are presented and assessed: censoring < LOD results and setting them equal to LOD/2 . The two strategies are then examined in the light of the reliability of the precision estimates obtained on the basis of a comprehensive simulation study. The first method exhibits better precision, while the second method displays better trueness. However, the two methods are only satisfactory as long as the percentage of < LOD results is less than 25 %. For higher percentages of < LOD results, more sophisticated methods should be applied.
ISSN:1661-5751
1661-5867
DOI:10.1007/s00003-015-0967-3