Social class and cause of death.

BACKGROUND Previous studies have shown that causes of death differ in their relationship to social class, but we lack a more comprehensive description of this variation. The present study provides a detailed and extensive list of social class differences for a large number of specific causes of death. METHODS All deaths between 1991 and 2003 in Sweden were linked with information on household social class from 1990. Relative death risks and excess mortality in groups of causes according to the European shortlist were estimated separately for men and women in eight classes using Cox Regression. RESULTS A clear mortality gradient among employees was found for the majority of causes, from low-relative death risks among higher managerial and professional occupations to relatively high risks for the unskilled working class. There is considerable variation in the strength of the association, from causes such as malignant melanoma, breast cancer and transport accidents among women, where no clear class differences were found. At the other extreme, mental and behavioural disorders, endocrine, nutritional and metabolic diseases and diseases of the respiratory system all show steep slopes for both men and women. Circulatory diseases and cancer together account for 15-20% of excess mortality. CONCLUSIONS Exceptions to the general pattern--causes of death in which higher social classes are exposed to greater death risks or in which there is no mortality gradient--are practically non-existent. There is nevertheless significant variation in the strength of the class differences in specific causes.