Consistency of seizure frequency estimates across time, methods, and observers.

Assessed the consistency of seizure frequency reports across observers, time, and methods. Thirty-two persons with intractable, partial seizures and 17 observers participated in a two-phase assessment procedure. During the first phase, subjects and observers provided retrospective seizure frequency estimates for two consecutive 30-day intervals. During the second phase, subjects and observers completed a daily seizure diary for 30 days. Patients' and observers' retrospective estimates were highly stable over a 2-month period. Substantial consistency was also found between two different recording methods (recall vs. diary) for subjects with epilepsy. In contrast, the observers' retrospective estimates were only moderately correlated with their prospective diary counts. Agreement between subjects' and observers' retrospective reports was also high. However, no significant association was found between their diary counts. Contrary to expectation, seizure frequency reports of patients with partial epilepsy showed moderately high consistency across time and between patient and observer pairs.