Urbanicity and schizophrenia. From statistical association to causality?