Are geographical differences in cardiovascular mortality due to morbidity differences or to methodological differences? The project "myocardial infarction in Mid-Sweden".