The purpose of this article is to examine in an emergency room (ER) po
pulation the concordance of self-reports of no alcohol consumption pri
or to injury with breath-analyzer readings in two groups: (1) those pa
tients from whom reports were obtained after they were breath analyzed
compared to (2) patients from whom reports were obtained prior to obt
aining the breath-analyzer reading. Data were collected on a probabili
ty sample of patients attending three health maintenance organization
ERs. Among those sampled were 159 patients admitted for initial treatm
ent of an injury, who were breath analyzed within 6 hours of the event
and reported no drinking following the event that lead to injury. Of
these, 119 were breath analyzed prior to the interview, and none who r
eported not drinking were positive on the breath analyzer, while of th
e 37 breath analyzed after the interview, only one was positive who ha
d reported not drinking. Obtaining the breath-analyzer reading followi
ng the interview was not found to affect the rate of refusal to provid
e a breath-analyzer reading; however, it was found to adversely affect
obtaining the breath-analyzer reading for other reasons. The data sug
gest that the concordance of negative self-reports of consumption with
breath-analyzer readings remains high in ER populations regardless of
when the breath-analyzer reading is obtained; however, it appears bes
t to obtain the reading prior to interviewing the patient for reasons
explained below.