Abstracts

Measurement of Interobserver Agreement Using the ACNS Subcommittee Proposal for Standardized Terminology for Rhythmic and Periodic EEG Patterns in Critically Ill Patients

Abstract number : 1.060
Submission category : Clinical Neurophysiology-Clinical EEG
Year : 2006
Submission ID : 6194
Source : www.aesnet.org
Presentation date : 12/1/2006 12:00:00 AM
Published date : Nov 30, 2006, 06:00 AM

Authors :
Paula A. Gerber, John F. Kerrigan, Steve S. Chung, Rama K. Maganti, Yu-Tze Ng, Andrew S. Little, and David M. Treiman

There is variable opinion regarding rhythmic EEG patterns in critically ill patients. Prior to meaningful research of treatment of these patterns, a universal terminology needs to be established. Also, consistent observer agreement should be demonstrated using said terminology. The ACNS Subcommittee on Research Terminology for Continuous EEG Monitoring recently proposed standardized terminology for rhythmic and periodic EEG patterns encountered in critically ill patients1. The objective of this study is to assess the degree of interobserver agreement using this terminology., Five epileptologists, all board-certified in clinical neurophysiology, were selected as evaluators. All five read the proposed terminology, and met to discuss the terminology and review the examples provided in the proposal. A total of 59 EEG samples (10 seconds each) from a previously established research data set, comprised of comatose patients with subarachnoid hemorrhage, were used2. The evaluators reviewed the samples in random order, and used the terminology to [ldquo]score[rdquo] each sample. The terminology addresses the following elements: presence or absence of rhythmic or periodic patterns, localization, morphology, rhythm, and frequency. Items not applicable to a 10 second EEG sample were not utilized. After [ldquo]scoring[rdquo] each sample, the results from each evaluator were tabulated and interobserver agreement was calculated using pairwise kappa statistics3,4 which were then averaged for each variable. In addition, the percentage agreement across all 5 observers was calculated., There were no samples for which all five observers agreed across all variables. Substantial agreement beyond chance3,4 was seen only for determination of whether a sample pattern was rhythmic/periodic, vs. excluded from classification due to lack of rhythmicity or the presence of an electrographic seizure (mean pairwise k=0.668; all five observers agreed 66.1% of the time). Moderate agreement was seen for localization (mean k=0.482) and frequency (mean weighted k=0.486). Agreement was fair with respect to assigning morphology as either periodic or rhythmic, which is a major element of the terminology (mean k=0.392). Agreement was also fair for the optional prefixes [ldquo]quasi-[ldquo] and [ldquo]frontally predominant[rdquo] (mean k=0.233, 0.383, respectively). Slight agreement was seen for the suffix [ldquo]-plus[rdquo] (mean k=0.117)., Despite using standardized terminology, significant variability exists between readers in the analysis of rhythmic and periodic EEG patterns. In particular, detailed descriptors (such as the term [ldquo]quasi-[ldquo]) show a higher degree of variability compared to basic descriptors (such as [ldquo]generalized[rdquo] vs. [ldquo]lateralized[rdquo]). Further refinement or enhanced definition of the terminology may be required to improve interobserver agreement.,
Neurophysiology