This analytical evaluation of oral fluid screening devices and preceding selection procedures was carried out as an integral part of the DRUID project (Work package 3, Task 2). The duration of the evaluation was from October 2007 to December 2009. The study was carried out by the Faculty of medicine and health sciences, Department of clinical chemistry, microbiology and immunology, Ghent University (UGent) in Belgium, the Alcohol and Drugs Analytics Unit, National Institute for Health and Welfare (THL) in Finland and the SWOV Institute for Road Safety Research in the Netherlands. The Department of Transport, Technical University of Denmark (DTU) was responsible for leading the task due to its connection to the road side survey (Work package 2, Task 2.2a1) for which DTU was Work package leader. THL was responsible for finalising the deliverable. Eight on-site tests were evaluated: BIOSENS Dynamic (Biosensor Applications Sweden AB), Cozart DDS (Cozart Bioscience Ltd.), DrugWipe 5+ (Securetec Detections-Systeme AG), Dräger DrugTest 5000 (Dräger Safety), OraLab6 (Varian), OrAlert (Innovacon), Oratect III (Branan Medical Corporation) and Rapid STAT (Mavand Solutions GmBH). Rapid STAT was tested in all three countries and DrugTest 5000 in Belgium and the Netherlands. All other devices were tested in only one country. Tested substance classes were amphetamine(s), methamphetamine, MDMA, cannabis, cocaine, opiates, benzodiazepines and PCP. A checklist for clinical signs of impairment (CSI) was also evaluated in order to see if visible signs of impairment can be used as preceding selection criteria for performing an on-site test. The checklist was based on several existing checklists, e.g. one developed for the German police and previously used in the European IMMORTAL (Impaired Motorists, Methods Of Roadside Testing and Assessment for Licensing) project. Study populations consisted of randomly selected drivers from the roadside survey for DRUID (Work package 2, Task 2.2a1), drivers suspected of driving under the influence of drugs, patients of treatment centres and rehabilitation clinics and customers of coffeeshops. Oral fluid was collected as the reference sample. For some cases, in the Netherlands, whole blood samples were also collected. The performance of the tests was assessed based on sensitivity, specificity, accuracy, positive predictive value and negative predictive value for the individual substance tests of the device. These were assessed based on both DRUID and manufacturer cut-offs. Sensitivity, specificity and accuracy performance values of 80% or more were set as a desirable target value. The analytical evaluation of the amphetamine test showed sensitivity varying from 0% to 87 %. Specificity values were from 91% to 100% and accuracy values from 84% to 98%. For cannabis tests, sensitivities ranged from 11% to 59%. Specificities were between 90% and 100% and accuracies from 41% to 82%. Cocaine tests scored sensitivities of between 13% and 50%, specificities of 99% to 100% and accuracies from 86% to 100%. Sensitivities of opiate tests ranged from 69% to 90%. Specificities were between 81% and 100% and accuracies between 75% and 99%. Benzodiazepine tests had sensitivities from 48% to 67%. Specificities were from 94% to 100% and accuracies from 77% to 100%. Not enough positive cases were gathered to successfully evaluate any of the methamphetamine, MDMA or PCP tests for the devices in which these were included. None of the tests reached the target value of 80% for sensitivity, specificity and accuracy for all the separate tests they comprised. An overall evaluation, wherein any positive drug screening result was viewed as valid providing that the confirmation sample contained one of the DRUID substances analysed, was performed as a measure of the usefulness of the devices in police controls. Three of the devices performed at >80% for sensitivity, specificity and accuracy in the overall evaluation. Prevalence of drugs in the study population needs to be considered when assessing the evaluation results. In addition, the type and prevalence of drugs within the population for which the device is intended to be used needs to be taken into account when considering the suitability of the device based on the results presented in this report. Some device failures were noted in the study. For one of the tests, 15 individual tests (12%) failed. For other tests, 5 or less tests failed. In the Netherlands the evaluation of Oratect III was stopped because the devices frequently failed to collect oral fluid in a sufficiently short time.. All countries took their own approach to the evaluation of the checklist for clinical signs of impairment. The results of the evaluations were not very promising. The checklist scored a low sensitivity value (Dutch study), low correlation of symptoms and actual presence of drugs (Belgian study) or there were difficulties in correlating the symptoms to actual drug use due to the insufficient data collection (Finnish study). (Author/publisher) This document is available at https://www.bast.de/Druid/EN/Home/home_node.html
Abstract