Accurate on‐site devices to screen for drug intake are critical for establishing whether an individual is driving under the influence of drugs (DUID); however, on‐site oral fluid (OF) cocaine device performance is variable. We evaluated the performance of a newly developed benzoylecgonine (BE) test‐strip for the Draeger® DrugTest 5000 device (20 µg/L cut‐off) with equivalent cross reactivity for cocaine and BE. Ten cocaine users provided OF, collected with the Draeger cassette and Oral‐Eze® and StatSure Saliva SamplerTM devices, up to 69 h following 25 mg intravenous cocaine administration. All screening results were confirmed by a validated two‐dimensional‐gas chromatography‐mass spectrometry (2D‐GC‐MS) method for cocaine and/or BE. Cocaine test‐strip median Tlast for screening only results was 6.5 h, and 6.5 h with Oral‐Eze® and 4 h for StatSure OF confirmation for cocaine and/or BE at 1, 8, and 10 µg/L; sensitivity, specificity, and efficiency ranged from 85.5 to 100% and 83.3 to 100% for cocaine only confirmation at 8 and 10 µg/L. For the BE test‐strip, median Tlast was 12.5 h for screening only and confirmation for cocaine and/or BE at all three cut‐offs; sensitivity, specificity, and efficiency ranged from 85.5 to 97.5% and 78.4 to 97.4% with cocaine and/or BE confirmation at 8 and 10 µg/L cut‐offs, respectively. The Draeger cocaine test‐strip with cocaine only confirmation offers a useful option for monitoring the acute intoxication phase of DUID; additionally the BE test‐strip with cocaine and/or BE confirmation increases the length of detection of cocaine intake for workplace drug testing, drug court, parole, pain management, drug treatment programs and both the acute cocaine intoxication and cocaine crash/fatigue phase of DUID. Copyright © 2016 John Wiley & Sons, Ltd.