Background : Driving under the influence of drugs (DUID) is a worldwide problem with potentially major judiciary and life-threatening consequences. Up to now, only classical drugs of abuse (DOA) are tested for DUID detection. A challenging issue for drafting up-dated international drug policies is to take into account the recent and expanding new psychoactive substances (NPS) market. NPS consist in various narcotic or psychotropic drugs, most of them having a “legal” status, that replicate chemical structures and/or pharmacological effects of classical DOA. Although it is obvious that NPS can lead to impaired driving, the prevalence of NPS use in a DUID context is unknown since the applied roadside screening tests are not yet adapted for these compounds. Methods : Between January and December 2016, a total of 391 oral fluid specimens were obtained from used roadside immunochemical test devices for DOA (Drugwipe-5S® device). These specimens were analyzed using liquid chromatography coupled with tandem mass spectrometry and high resolution mass spectrometry. Results : NPS (mainly cathinone derivatives) were detected in 33 out of the 391 oral fluid samples. This NPS positivity rate of 8.4% in oral fluid of drivers who were submitted to a roadside drug testing in 2016 in France and in Belgium is comparable to the available blood data (NPS positivity rate of 7%) observed in 2015 in similar populations. Conclusion : Our results demonstrate the reality of driving after NPS use in French and Belgian drivers who were submitted to a roadside DOA test. As there is a lack of on-site detection methods to screen for NPS, the detection of NPS in a rapid and cost-effective DUID detection strategy is currently impossible. The expanding use of NPS, notably by drivers as reported here, and the inability of currently used drug detection tests, should be urgently addressed by road safety and law enforcement authorities.