Glasgow coma scale (GCS), GCS-pupil score (GCS-P), and full outline of unresponsiveness (FOUR) score in traumatic brain injury patients: A comparative study of inter-observer variability and prognostic ability

Abstract Background and Aims: Various scores and scales have been devised to assess and prognosticate patients with altered sensorium. Recognizing an easy, quick and reliable score and implementing it uniformly can facilitate better assessment of patients. Our study was designed to identify the best score among Glasgow coma scale (GCS), GCS-Pupil (GCS-P), and Full Outline of Unresponsiveness (FOUR) score in terms of inter-observer variability as assessed by junior resident (JR) medical officer and a nursing officer (NO), and prognostic ability as measured by the Glasgow Outcome Scale-Extended (GOS-E). Material and Methods: After Institutional Ethics Committee approval and trial registration, this prospective observational study was conducted in 136 patients with traumatic brain injury (TBI). GCS, GCS-P, and FOUR scores were assessed within 6 hours of admission by a JR and a NO. The GOS-E was evaluated at discharge and again at 3 months post-injury via telephonic follow-up. Results: The inter-observer agreement, assessed using the kappa value, was 0.670 for GCS, 0.718 for GCS-P, and 0.593 for the FOUR score. The area under the receiver operating characteristic curve was 0.989 for GCS (near perfect prediction), 0.975 for FOUR score (excellent prediction), and 0.718 for GCS-P (moderate prediction). The Kruskal-Wallis test was performed to assess the prognostic ability of three scores in comparison with the GOS-E score. GCS, GCS-P, and FOUR scores had P values of 0.001, indicating that the scores are comparable in predicting functional outcomes. Conclusions: GCS-P showed the highest inter-observer agreement, indicating superior consistency between assessors. GCS and FOUR score exhibit the highest accuracy in predicting functional outcomes.