@article{ref1,
title="Scientific assessment of proficiency testing of illicit drugs in blood as organized by the GTFCh in 2003 - Evaluation according to ISO 5725 and THC-test near the limit of detection",
journal="Blutalkohol",
year="2005",
author="Schmitt, G. and Herbold, M and Aderjan, R. and Mattern, Rainer",
volume="42",
number="1",
pages="20-31",
abstract="Forensic inter-laboratory proficiency tests (ring tests) can either be evaluated using the criteria given by the &quot;Harmonized Protocol&quot; [14] or according to the ISO guideline 5725. For a given drug, the latter requires two replicate determinations of each sample, however, limited by the sample volume or other analytical needs like consecutive determination of additional analytes. As a result of data obtained from the inter-laboratory testing BTMF 3/03, it was for the first time possible to confirm an acceptable use of the target standard deviation according to the Horwitz-function for the analytes THC-COOH, morphine and benzoylecgonine. As expected, inter-laboratory testing shows that the repeatability is considerably better than the reproducibility. For THC, the evaluation criterion brought to an issue according to ISO 5725 was 1.9-fold higher than the Horwitz standard deviation. Using the ISO procedure and its standard deviation resulted in a higher success ratio of 93% (HORWITZ standard deviation: 73%). Whether or not variable success criteria according to the participant's performance (ISO 5725) or the - up to now - fixed ones (Harmonized Protocol) are preferable can only be determined if the more costly ring tests are accepted and can be organized in future. Regarding inter-laboratory comparison and analytical performance, proficiency testing using a spike value of 0.5 ng THC / mL serum confirmed that this concentration is equal to the limit of detection as it is defined by the German DIN 32625 standard. However, this statistically calculated limit only accounts for 50% detection certainty, which is far better represented by the term &quot;Erfassungsgrenze&quot; (according to DIN 32645). With a (beta)-error of e.g. 1% accepted, this term signifies a 99% certain acquisition/registration of identifying mass spectroscopic information. For tetrahydrocannabinol such an analytical threshold amounts to a concentration level of 1 ng/mL serum.<p />",
language="de",
issn="0006-5250",
doi="",
url="http://dx.doi.org/"
}