We studied how emergency qualitative drug screens were ordered and used during the evaluation of 405 consecutive adult patients presenting to a metropolitan emergency department with suspected drug intoxication. Physicians completed a two-part questionnaire outlining diagnosis, diagnostic certainty, suspected drug(s), and management plans immediately following initial evaluation and again at the time of ED disposition. Screen results were reviewed prior to ED disposition in 361 (89%) cases. We found that qualitative drug screens were associated with substantial changes in diagnostic certainty, as measured by changes in probability estimates and changes in absolute mean log-likelihood ratios. In 196 cases for which all data were recorded prospectively, drug screens excluded a diagnosis of drug intoxication or specifically suspected drugs in 81 cases (41%), but identified previously unsuspected drugs in only 21 cases (11%, P less than .001). Management changes followed qualitative drug screen results in 16 of 361 cases overall (4.4%), but seven of these were also associated with diagnostic quantitative serum drug levels. Potentially critical interventions were begun in two cases following positive screens but delayed in another due to a falsely negative screen. Guidelines for more efficient test utilization are proposed.