Observer variability in assessing the clinical features of subarachnoid hemorrhage.

Clinical assessment of patients with subarachnoid hemorrhage (SAH) is important both in determining management and in predicting outcome. A previous report showed considerable observer inconsistency when patients were graded with either the Hunt and Hess or the Nishioka system. This study evaluates observer variability in assessment of the individual clinical features from which these grading systems are derived. Assessment of the presence or absence of neck stiffness caused least inter-observer variability. Headache caused most variability, due to difficulty in grading its severity. Determination of the severity of a neurological deficit proved more reliable than deciding whether or not a deficit was present. The terms used to describe the level of consciousness in the Hunt and Hess and Nishioka systems were found to be significantly less consistent than the Glasgow Coma Scale. The authors suggest that when patients with SAH are assessed it is necessary to take into account the consistency with which observers can record a clinical feature, as well as its prognostic importance.