BACKGROUND: Different criteria are employed to activate trauma teams. Because of a growing concern about overtriage, the objective of this study was to investigate the performance of our trauma team's activation protocol. METHODS: Injured patients with trauma team activation (TTA), admission to an intensive care unit or surgical intermediate care unit with a trauma diagnosis, or trauma-related death in the emergency department were investigated retrospectively from 1 January 2004 to 31 December 2005. Different TTA criteria were analysed with respect to sensitivity, positive predictive value (PPV) and overtriage (1 - PPV). RESULTS: Eight hundred and nine patients were included, 185 (23%) of whom had an Injury Severity Score (ISS) of more than 15. The performance of our protocol showed a sensitivity of 87%, PPV of 22% and overtriage of 78%. The mechanism of injury as a TTA criterion had a sensitivity of 14%, PPV of 7% and overtriage of 93%. Physiological/anatomical criteria and interfacility transfer showed higher PPV and less overtriage. Undertriage (no TTA despite ISS > 15) was identified in 23 patients (13%), 18 of whom were hospital transfers. CONCLUSION: A TTA protocol based on physiological, anatomical and interfacility transfer criteria seems to yield a higher precision than, in particular, that based on mechanism of injury criteria. Because of substantial overtriage in our hospital, the TTA protocol needs to be re-evaluated.