Background. Driving under the influence of intoxicating drugs other th
an alcohol may be an important cause of traffic injuries. We used a ra
pid urine test to identify reckless drivers who were under the influen
ce of cocaine or marijuana. Methods. We conducted a consecutive-sample
study in Memphis, Tennessee, in the summer of 1993. Subjects arrested
for reckless driving who were not apparently impaired by alcohol (did
not have an odor of alcohol, tested negative on breath analysis, or b
oth) were tested for cocaine and marijuana at the scene of arrest. The
results of the drug tests were compared with clinical evaluations of
intoxication made at the scene by a police officer. Results. A total o
f 175 subjects were stopped for reckless driving, and 150 (86 percent)
submitted urine samples for drug testing at the scene of arrest. Eigh
ty-eight of the 150 (59 percent) tested positive: 20 (13 percent) for
cocaine, 50 (33 percent) for marijuana, and 18 (12 percent) for both d
rugs. Ninety-four of the 150 tested drivers were clinically considered
to be intoxicated, and 80 of them (85 percent) tested positive for co
caine or marijuana. The intoxicated drivers had a broad range of affec
ts and appearances. Nearly half the drivers intoxicated with cocaine p
erformed normally on standard sobriety tests. Conclusions. Over half o
f the reckless drivers who were not intoxicated with alcohol were foun
d to be intoxicated with other drugs. Toxicologic testing at the scene
is a practical means of identifying drivers under the influence of dr
ugs and is a useful adjunct to standard behavioral sobriety testing.