In this issue of Clinical Chemistry , Desrosiers et al. (1) describe an evaluation of an on-site test for detecting cannabinoids in oral fluid (OF).3 These investigators analyzed 66 samples from study participants who smoked a controlled dose of cannabis. The device has a cutoff of 5 μg/L for Δ9-tetrahydrocannabinol (THC), and comparison of the on-site results with the results for a gas chromatography–tandem mass spectrometry method with a 2-μg/L THC cutoff yielded a diagnostic sensitivity of 91% for the on-site test. Owing to the design of the study, most of the samples contained THC, so the estimation of the specificity was less reliable. Most other published evaluations, however, have shown that the specificity of the Draeger DrugTest 5000 is quite good. In our own evaluation of the same device (5-μg/L THC cutoff) in a methadone clinic, our use of the cutoffs mandated in the Belgian legislation produced a sensitivity, specificity, and efficiency of 81%, 95%, and 92%, respectively (2). Desrosiers et al. refer to the Driving under the Influence of Drugs, Alcohol, and Medicines (DRUID) requirements, for which the sensitivity and specificity of an on-site test should be >80%. The Draeger DrugTest 5000 meets these requirements for THC. These requirements are less demanding than the requirements in the first ROSITA (Roadside Testing Assessment) project in 2000, for which a 90% sensitivity, a 90% specificity, and a 95% efficiency were required, but these requirements were changed because no device satisfied the criteria. They are more stringent than the sensitivity of 75% required by traffic police in their operational evaluation of on-site OF tests in DRUID cases (3). Thus, it seems that the long quest for a sensitive on-site test for detecting cannabinoids in OF is finally over. Possible applications are workplace drug testing (WDT) and …