Abstracts

Questionnaire surveys provide an efficient means for screening large populations to identify those with a positive history of epileptic seizures. However, the quality of information obtained in this way can be of concern given the inherent biases usually associated with self- or surrogate-reported data. This study was undertaken to assess the accuracy of self- and cotwin-reported history of seizures in three population-based twin registries. Questionnaire information obtained from adult twins included in the population-based Danish Twin Registry (DTR), Norwegian Twin Panel (NTP), and Mid-Atlantic Twin Registry (MATR) was used to identify twins with a history of epileptic seizures. Possible cases were identified on the basis of positive responses to queries regarding history of epilepsy, febrile seizures (FS), other seizures and staring spells in the respondant or their cotwin. Reported cases were verified by medical records and detailed clinical and family interviews. Self-reported epilepsy was verified in 81.9% of twins overall and in 86.1% of Danish, 75.6% of Norwegian and 80.7% of MATR twins. History of epilepsy was validated in 97.2% (MATR) and 98.7% (NTP) of twins when both pair members reported the case to be affected and in 81% (MATR) and 100% (NTP) of twins when the information was provided by the cotwin alone. A similiar pattern was observed for self-reported FS, where verification was possible in 85.9% of cases overall (92.1% (DTR), 76.3% (NTP) and 78.3% (MATR)). In pairs where both twins reported a history of FS in the case twin, it could be verified in 90.9% (MATR) and 98% (NTP) of cases, with 6.1% (MATR) and 2% (NTP) reports found to be a mistake/denial. For cotwin reports, it was possible to verify 64.5% (MATR) and 100% (NTP) as FS while 25.8% (MATR) were denied and 6.5% (MATR) were another medical condition. Self-reports of epileptic seizures were found to be most reliable in the Danish sample, whose members were much younger than NTP or MATR twins at the time the information on seizure history was collected . The percentage of twins who denied a history of epilepsy after a previous self-report was increased among MATR twins (8.6%) compared to that observed in either the NTP (4.5%) or the DTR (0.6%). The accuracy of self- and co-twin reported history of epilepsy and FS was high across populations. Differences between populations in percentage of reports verified demonstrate the difficulty encountered in validating retrospective self-reports of epileptic seizures which occurred in subjects 40 to 50 years ago. (Supported by a grant from the National Institutes of Health NINDS (NS31564).)