This pilot study was carried out to determine whether converting from a two-tier to a three-tier in-hospital trauma triage system improves the efficiency of emergency department (ED) care and minimizes inappropriate triage. Patients at an urban, Level 1 trauma centre were triaged using either a two-tier (months 1-3; n = 197) or three-tier (months 4-6; n = 240) trauma response system. Patients were assessed for triage type, age, sex, injury severity score, Glasgow coma score, post-ED disposition, total ED time, survival, complication rate, probability of survival and unexpected death. Comparisons were made by ANOVA table analysis; significance was assumed for p<0.05. Two-tier (n = 197) and three-tier patients (n = 240) were matched with respect to mean age, sex, mean injury severity score, mean Glasgow coma score, post-ED disposition, survival and probability of survival. Two-tier patients were triaged to give 20% alerts [criteria = physiological derangement (PD) and/or injury mechanism (MOI)] and 80% consults; three-tier patients were triaged as 20% category I (criteria = PD), 18% category II (criteria = MOI) and 62% consults. Total ED time decreased from two-tier category I times (2.09+/-1.64 vs. 1.95+/-1.75 h; p = 0.72). Category II patients (3.28+/-1.98 h; p = 0.009) spent less time in the ED than did two-tier consults (4.36+/-2.65 h). The mean ED three-tier consult time significantly decreased as well (3.95+/-2.42 h, p = 0.008 vs. two-tier consult). Complications per patient were unchanged from two-tier to three-tier triage (0.17+/-0.52 vs. 0.12+/-0.48; p = 0.15). Under-triage (5%) and over-triage (7.5%) were minimal under three-tier triage. It is concluded that using a three-tier triage system results in an increase in the early involvement of the trauma service while decreasing emergency department time and minimizing over-triage. (C) 1997 Elsevier Science Ltd.