In Reply.—We appreciate the letter of Beringer and colleagues. The ideal method to generate a "valid epidemiologic risk" for submersion accidents in children would be to perform a prospective study, but such a study would be difficult if not impossible to perform from a practical standpoint. Drowning accidents are rare occurrences (5.6/100,000 population), and to study a rare occurrence in a subpopulation of children with epilepsy (0.5% to 1.5% of the population) in a prospective fashion would be a long and formidable task. The impressive consistency between the diverse locations and studies listed in our article for the risk of submersion accidents in children with epilepsy suggests that the risk estimates are legitimate and that our statements of an average risk four times that of normal children is appropriate.
The precise quantification of the magnitude of the increased risk is probably less important than the realization and acceptance of