A comparison of the Glasgow Coma Scale and the Reaction Level Scale (RLS85)

Abstract

The Glasgow Coma Scale (GCS) and the Reaction Level Scale (RLS85) were compared for rating neurosurgical patients in regard to ranking order of deficit severity, interobserver variability, and coverage for relevant factors. Four physicians, four registered nurses, and four assistant nurses performed 72 pairwise ratings on 47 neurosurgical patients. The rank correlation between the GCS sum score and the RLS85 was -0.94, suggesting the same ranking order of severity and indicating that the underlying concepts of somnolence, delirium, and motor responses in coma are evaluated in the same way. By the sign test, the RLS85 was shown to have better interobserver agreement than the GCS sum score and the eye-motor-verbal (EMV) profile. The interobserver grading disagreements in both scales were distributed over the entire range of responsiveness, and for the GCS sum score they were slanted to combined segments 9 to 15. The RLS85 showed full coverage of relevant factors, while 43 (60%) of the 72 test occasions in the GCS sum score and the EMV profiles showed untestable features, most often because of patient intubation. The pseudoscore (that is, the choice of value given to untestable features) affects interobserver agreement as well as the estimated overall patient responsiveness in the GCS sum score. Assessment by the order of applying the scales showed a significant effect on the GCS eye-opening scale (p = 0.01) and the GCS sum score (p = 0.03), indicating a sensitivity to environmental stimuli unrelated to the patient's status. This study demonstrates that basically the same information as that found in the separate eye, motor, and verbal scales of the GCS can be combined directly into the RLS85, which has better interobserver agreement and better coverage than the GCS sum score.