The geographic distribution of multiple sclerosis: An examination of mathematical models☆