Objective To study the feasibility of modified early warning scores (MEWS) for assessing the severity and death prediction in the pre-hospital traumatic patients.MethodsData of the prehospital traumatic patients admitted between January 1,2010 and December 31,2010 were collected and assessed onsite by using MEWS.Numeration data was presented in percentage by using chi-square test,and measurement data was xepressed in mean with standard deviation,and P ＜ 0.05 was considered to be difference with statistical significance.Observation was lasted for 90 days after admission to get final results as observation object and ROC curve was drew and calculated the area under the curve for predicting severity and death of patients.The patients without vital signs and unsuccessful resuscitations were not included in this study.ResultsThere were 1475 (87.95％) cases/times with score of 0 -2,and 202 (12.05％)cases/times with score of 3 - 13.In the non-survival group,MEWS were higher than that in the survival group with statistic significance ( P ＜ 0.01 ).When the area under ROC was 0.94,the optimal cutoff point for potentially severe patients was MEWS≥3 for predicting the death of severe pre-hospital traumatic patients with sensitivity of 85.7％,specificity of 88.6％,accuracy of 88.6％ and Youden of 0.743,showing high significance of the application of MEWS to assessing severity of traumatic patients and death prediction.ConclusionsMEWS used to assess the pre-hospital traumatic patients and predict death with high validity and accurate quantification is a simple,practical and easily operable method with strong application significance.