Measuring the effects of pre-test probability on out-patient first EEG investigation in children – A guide to evidence-based EEG triage in a pandemic

Abstract

IntroductionThe yield of epileptiform EEG abnormalities is lower in unselected Paediatric populations than in prospective studies of incident seizures or prevalent epilepsy studies. At a time of limited capacity, it is important to match available EEG resources to children who are most likely to benefit. In this study we evaluated a prospective triage tool for estimating the likelihood of epileptiform abnormality in children’s first out-patient EEG. MethodsWe prospectively triaged 1865 out-patient referrals to the largest Paediatric EEG laboratory in Ireland. Based on a structured algorithm, we dichotomized first EEG referrals into priority and non-priority groups and assigned one of 5 sub-levels based on anticipated EEG yield. EEGs were reported by a single Consultant in Clinical Neurophysiology. ResultsTriage designated 757 (41 %) EEG referrals as non-priority. Priority exceeded non-priority referrals for all age groups except children between 18 months and 3.5 years. EEGs showed a two-fold higher incidence of interictal epileptiform abnormalities for priority referrals (36 % vs 18 %, p < 0.001). Rates of interictal epileptiform abnormality correlated with the 5 sub-levels of triage (p < 0.01). Epileptiform yield was highest (39 %) for children over 5 years vs 17 % for those under 5 years (p < 0.00001); these rates increased to 49 % and 20 % respectively for priority referrals. ConclusionStructured pre-test triage of EEG referrals can identify children who have the greatest likelihood of epileptiform abnormality. In a mixed population of Paediatric referrals, the epileptiform yield of first time EEG is 49 % for children over 5 years who are referred with an appropriate EEG indication. This is subject to much variability with epileptiform yields as low as 13 % in younger children with non-priority referrals. The use of a structured triage algorithm can help to optimise utility of EEG in situations of limited laboratory capacity.