Interobserver agreement in the interpretation of EEG patterns in critically ill adults

J Clin Neurophysiol. 2008 Oct;25(5):241-9. doi: 10.1097/WNP.0b013e318182ed67.


The significance of rhythmic and periodic EEG patterns in critically ill patients is unclear. A universal terminology is needed to facilitate study of these patterns, and consistent observer agreement should be demonstrated in its use. The authors evaluated inter- and intraobserver agreement using the standardized terminology (Hirsch et al., J Clin Neurophysiol 2005;22:128-135) recently proposed by the American Clinical Neurophysiology Society. Trained electroencephalographers viewed a series of 10-second EEG samples from critically ill adults (phase I), a set of >/=20-minute EEGs from the same patient cohort (phase II), and then reevaluated the first sample set (phase III). The readers used the proposed terminology to "score" each EEG. For each possible term, interobserver agreement (phases I and II) and intraobserver agreement (phase III) were calculated using pairwise kappa (kappa) values. Moderate agreement beyond chance was seen for the presence/absence of rhythmic or periodic patterns and for localization of these patterns. Agreement for other terms was slight to fair. Inter- and intraobserver agreement were consistently lower for optional terms than mandatory terms. Even when standardized terminology is used, the description of rhythmic and periodic EEG patterns varies significantly. Further refinement of the proposed terminology is required to improve inter- and intraobserver agreement.

MeSH terms

  • Adult
  • Critical Illness / epidemiology*
  • Electroencephalography / standards*
  • Electroencephalography / statistics & numerical data
  • Humans
  • Neurology / standards*
  • Observer Variation
  • Subarachnoid Hemorrhage / physiopathology
  • Terminology as Topic*