Objectives Previous studies, conducted mainly in ICUs, have shown low compliance with hand-washing recommendations, with failure rates approaching 60%. Hand washing in the emergency department has not been studied. We examined the frequency and duration of hand washing in one emergency department and the effects of three variables: level of training, type of patient contact (clean, dirty, or gloved), and years of staff clinical experience. Design Observational. Setting ED of a 1100-bed tertiary referral, central city, private teaching hospital. Participants Emergency nurses, faculty, and resident physicians. Participants were informed that their activities were being monitored but were unaware of the exact nature of the study. Interventions An observer recorded the number of patient contacts and activities for each participant during 3-hour observation periods. Activities were categorized as either clean or dirty according to a scale devised by Fulkerson. The use of gloves was noted and hand-washing technique and duration were recorded. A hand-washing break in technique was defined as failure to wash hands after a patient contact and before proceeding to another patient or activity. Results Eleven faculty, 11 resident physicians, and 13 emergency nurses were observed. Of 409 total contacts, 272 were clean, 46 were dirty, and 91 were gloved. Hand washing occurred after 32.3% of total contacts (SD, 2.31%). Nurses washed after 58.2% of 146 contacts (SD, 4.1%), residents after 18.6% of 129 contacts (SD, 3.4%), and faculty after 17.2% of 134 contacts (SD, 3.3%). Nurses had a significantly higher hand washing frequency than either faculty (p < 0.0001) or resident physicians (p < 0.0001).(ABSTRACT TRUNCATED AT 250 WORDS)