Assessment of polytraumatized patients according to the Berlin Definition: Does the addition of physiological data really improve interobserver reliability?

Abstract

Background Several new definitions for categorizing the severely injured as the Berlin Definition have been developed. Here, severely injured patients are selected by additive physiological parameters and by the general Abbreviated Injury Scale (AIS)-based assessment. However, all definitions should conform to an AIS severity coding applied by an expert. We examined the dependence of individual coding on defining injury severity in general and in identifying polytrauma according to several definitions. A precise definition of polytrauma is important for quality management. Methods We investigated the interobserver reliability (IR) between several polytrauma definitions for identifying polytrauma using several cut-off levels (ISS 16, 18, 20, 25 points, and the Berlin Definition). One hundred and eighty-seven patients were included for analyzing IR of the polytrauma definitions. IR for polytrauma definitions was assessed by Cohen’s kappa. Results IR for identifying polytrauma according to the relevant definitions showed moderate agreement (<0.60) in the ISS cutoff categories (ISS 16, 18, and 20 points), while ISS 25 points just reached substantial agreement (0.62) and the Berlin Definition demonstrated a correlation of 0.77 which is nearly perfect agreement (>0.80). Conclusion Compared with the ISS-based definitions of polytrauma, the Berlin Definition proved less dependent on the individual rater. This underlines the need to redefine the selection of severely injured patients. Using the Berlin Definition for identifying polytrauma could improve the comparability of patient data across studies, in trauma center benchmarking, and in quality assurance.