When is an apparent excess of neurologic cases epidemiologically significant?