This study evaluates the military performance of soldiers who were identified by the Army's urine screening program as having used drugs at the time they reported for active duty. The study used prospectively defined cohorts of large size (over 1,600 urinalysis positive and over 2,400 urinalysis negative entering males) and used longitudinal outcome measures drawn from existing medical and personnel files. The major findings are that the urinalysis positive individuals were significantly more likely not to complete their tour of duty, were more likely to be hospitalized, and had more bed-days per hospitalization than their urinalysis negative fellow soldiers. In terms of cohorts entering the Army, however, minimal improvement in outcome was found by considering an entirely drug-free cohort (obtained by omitting the urinalysis positive individuals) compared to a random cohort which contained positives at the then current prevalence of 2 per cent.