Preexisting cognitive impairment in intracerebral hemorrhage: Methodological issues