An Observational Study to Assess Inter-observer agreement for neonatal EEG Interpretation

Objectives: To assess interrater observer agreement (IOA) in neonatal electroencephalogram (EEG) with respect to background characteristics, presence or absence of age related graphoelements, transient EEG patterns, interictal discharge characteristics & ictal discharge characteristics including status epilepticus. Methods: It is an observational study consisting of 45 neonatal EEGs reported by two certified Pediatric neurologists performed over one year period. EEG of neonates up to 44 weeks of postconceptional age (regardless of gestational age) were included after obtaining institutional ethics approval. Variables studied were- 1. EEG background characteristics (continuity, symmetry, synchrony, variability, reactivity, dysmaturity and voltage) 2. Presence or absence of normal grapho-elements specific to neonatal EEG a. ( monorythmic delta activity, delta brushes, rhythmic temporal theta, anterior dysrhythmia, encoches frontalis) 3. Presence or absence of EEG transient patterns 4. Presence or absence of seizures on EEG 5. Location of epileptiform discharges 6. Presence or absence of status epilepticus on EEG Data was obtained as nominal data. Cohen's kappa coefficient (κ) was used to measure IOA. Results: Measurement of agreement was MODERATE for normal background and brief rhythmic discharges. It was FAIR for excessive background discontinuity, synchrony, symmetry, variability & in case of central or multifocal discharges. It was POOR for reactivity, dysmaturity, graphoelements, suppression-burst, transient activity & discharges in other areas. Conclusion: This study underlines that there is moderate to fair amount IOA with respect to background characteristics & maturity. Contrarily, there was significant disagreement between the observations with regards to transient & epileptic elements.
Keywords: Neonatal EEG, Inter-observer agreement, Neonatal seizures, Neonatal encephalopathy, Background dysmaturity