Longitudinal trends in trauma mortality and survival in Stoke-on-Trent 1992–1998

Abstract

Objectives: To identify factors related to mortality and to test the null hypothesis of no longitudinal trend in mortality in patients admitted to the North Staffordshire Hospital (NSH) with an Injury Severity Score (ISS) greater than 15, between April 1992 and March 1998. Design: Longitudinal prospective study of 18 factors, including age, sex, mechanism of injury, anatomical injury scores and year of admission. Outcome, based on mortality at discharge, was analysed in two ways: alive or dead at discharge (mortality) and time to death or discharge (survival). Results: A decreasing trend ( P<0.01) in mortality with year of admission was detected on the log-odds scale. The trend could not be explained by a case-mix analysis, which allowed for the 17 other factors. Using multiple logistic regression analysis (mortality) and Cox proportional hazards analysis (survival), eight factors were identified as determinants of outcome: age, head AIS score, chest AIS score, abdominal AIS score, calendar year of admission, external injury AIS score, mechanism of the injury and primary receiving hospital. Conclusions: The observed improvement in survival in severely injured patients must result from the interplay of factors not controlled in this analysis or improvements in patient care or both.