To the Editor.
—To identify risk "factors for attempted suicide among epileptics" requires that all subjects have epilepsy. Nevertheless, in an attempt to achieve this goal, the authors of a recent report chose controls from among nonepileptics.1 To draw inferences from a case-control design requires persons with epilepsy who attempted suicide (cases) to be compared with persons with epilepsy who have not attempted suicide (controls). Nonepileptic controls have no place in this study.To use as controls persons who attempted suicide but who do not have epilepsy results in: (1) failure to identify factors related to an increased risk of attempted suicide, since both cases and controls have made the attempt, or (2) identifying factors that may be associated with epilepsy but not with attempted suicide, since it is epilepsy that distinguishes cases and controls. Such a design does not allow for identification of risk factors of attempted suicide
Leviton A, Cowan L, Kuban K. Epilepsy and Suicide. Arch Neurol. 1990;47(4):381. doi:10.1001/archneur.1990.00530040021008