Objective: Some major trauma (Injury Severity Score [ISS] >15) patients transported to a secondary hospital in Perth do not survive. We sought to describe this cohort and assess preventability.Methods: A cohort study from a previously developed cohort of trauma deaths in Western Australia from 1 July 1997 to 30 June 2006. A preformatted data sheet was used to collect a range of descriptive, time, physiological, and autopsy data. Trauma scores were calculated. Preventability was assessed using three approaches, based on ISS, Trauma Revised Injury Severity Score (TRISS) and individual case review.Results: There were 74 major trauma deaths, mean age 55.6 +/- 26.3 years (range 3-95). Thirty-seven (50%) were motor vehicle crashes. The mean Revised Trauma Score was 3.84 +/- 3.09 (0-7.84), median ISS 31 (interquartile range [IQR] 25-51), median TRISS 0.127 (IQR 0.031-0.772) and median time to death was 80 min (IQR 20 min-10 h 8 min). Severe head and chest injuries were the most common. Almost half (36, 48.6%) were receiving CPR on arrival to the hospital. The crude proportion of potentially preventable deaths, based on ISS, TRISS and case review, were 16.2%, 32.4% and 6.7%, respectively. However, these were predominantly elderly patients and a decision against resuscitation was recorded in 54%.Conclusions: The proportion of potentially preventable major trauma deaths at Perth secondary hospitals is low. The most notable group were the elderly after falls, and trauma system efforts should be focused on this group. Primary prevention of major trauma represents the biggest opportunity for improvements in trauma survival.