Laboratory evaluation of two passive alcohol sensors.

Abstract

Passive alcohol sensors are screening devices designed to sample nonintrusively the exhaled breath of a person to determine whether or not it contains alcohol and if so approximately how much. Two production passive alcohol sensors (NPAS passive alcohol sensor and the Life-Loc PBA 2000) were evaluated in a laboratory environment to establish appropriate threshold measurements that indicate probable alcohol impairment. The laboratory evaluation was conducted using both instrument types with 12 drinking subjects. Both sensors were able to identify alcohol in exhaled breath with sufficient accuracy to identify persons with high blood alcohol concentrations (BACs). The accuracy of both sensors was related to the distance from the subject's mouth: the further away they were from a subject's mouth the greater the chance that high BACs would not be detected. Under ideal conditions, it was estimated that the Life-Loc sensor could be expected to correctly detect 80% of drivers at 0.10% BAC (99% at 0.15% BAC) as being impaired while incorrectly identifying only about one in eight drivers with BACs of 0.02% (false positives). In comparison, the NPAS sensor could be expected to correctly detect about 75% of drivers at 0.10% BAC (97% at 0.15% BAC) while incorrectly identifying one in five drivers at 0.02% BAC.