Assessing the consistency and validity of self-reported year of diagnosis among participants in a longitudinal study of people with multiple sclerosis.