Can the epidemiology of schizophrenia be explained by variation in maternal influenza?