Problems in estimating the prevalence of epilepsy in children: the yield from different sources of information

A prevalence study of childhood epilepsy in central Oklahoma allowed a comparison of different sources of case-finding. The percentage of all epilepsy cases identified was lowest (9%) from physicians' office records. Higher rates were obtained from review of emergency room records (40%), hospital inpatient records (45%), electroencephalography (EEG) requisitions (68%) and outpatient clinic records (62%). Nearly one-third of the 1159 epilepsy cases were identified from only one source. EEG requisitions provided most of these cases. Although the percentage of all abstracts from one source that identified cases was lowest for EEG requisitions (43%), these abstracts were prepared with much less effort and time than were abstracts from hospital or clinic records, which had higher levels of efficiency (hospital = 52%, clinic = 73%). Hospital inpatient records identified a disproportionate number of children with newborn seizures, simple partial epilepsy, and partial seizures secondarily generalised, and children with multiple developmental handicaps. Fully 10% of all cases had an episode of status epilepticus and, as might be expected, emergency room and hospital inpatient records preferentially identified them. We conclude that case-finding for epidemiological studies of epilepsy in children in some communities is best done through review of EEG requisitions, supplemented by information from outpatient records.