Comparison of the prognostic validity of three simplified consciousness assessment scales with the Glasgow Coma Scale

Background Various tools simpler than the Glasgow Coma Scale (GCS) have been proposed for the assessment of consciousness. In this study, the validity of three coma scales [Simplified Motor Scale, Modified GCS Motor Response, and AVPU (alert, verbal, painful, unresponsive)] is evaluated for the recognition of coma and the prediction of short- and long-term mortality and poor outcome. The predictive validity of these scales is also compared to the GCS. Methods Patients treated in the Department of Neurosurgery and the Intensive Care Unit in need of consciousness monitoring were assessed by four raters (two consultants, a resident and a nurse) using the GCS. The corresponding values of the simplified scales were estimated. Outcome was recorded at discharge and at 6 months. Areas Under the Receiver Operating Characteristic Curve (AUCs) were calculated for the prediction of mortality and poor outcome, and the identification of coma. Results Eighty-six patients were included. The simplified scales showed good overall validity (AUCs > 0.720 for all outcomes of interest), but lower than the GCS. For the identification of coma and the prediction of long-term poor outcome, the difference was significant ( p