Study objective: We evaluate a computer-based intervention for screening and health promotion in the emergency department and determine its effect on patient recall of health advice. Methods: This controlled clinical trial, with alternating assignment of patients to a computer intervention (prevention group) or usual care, was conducted in a university hospital ED. The study group consisted of 542 adult patients with nonurgent conditions. The study intervention was a self-administered computer survey generating individualized health information. Outcome measures were (1) patient willingness to take a computerized health risk assessment, (2) disclosure of behavioral risk factors, (3) requests for health information, and (4) remembered health advice. Results: Eighty-nine percent (470/542) of eligible patients participated. Ninety percent were black. Eighty-five percent (210/248) of patients in the prevention group disclosed 1 or more major behavioral risk factors including current smoking (79/248; 32%), untreated hypertension (28/248; 13%), problem drinking (46/248; 19%), use of street drugs (33/248; 13%), major depression (87/248; 35%), unsafe sexual behavior (84/248; 33%), and several other injury-prone behaviors. Ninety-five percent of patients in the prevention group requested health information. On follow-up at 1 week, 62% (133/216) of the prevention group patients compared with 27% (48/180) of the control subjects remembered receiving advice on what they could do to improve their health (relative risk 2.3, 95% confidence interval 1.77 to 3.01). Conclusion: Using a self-administered computer-based health risk assessment, the majority of patients in our urban ED disclosed important health risks and requested information. They were more likely than a control group to remember receiving advice on what they could do to improve their health. Computer methodology may enable physicians to use patient waiting time for health promotion and to target at-risk patients for specific interventions.