European societies are increasingly ageing, and a considerable share of this elderly population has a migrant background. In Belgium in particular, large-scale immigration took place quite early, resulting in a large migrant population. Since migrant groups can be seen as a natural experiment from an epidemiological point-of-view, studying immigrant’s mortality relative to the host population may provide important clues on disease aetiology. Epidemiological studies on immigrant mortality generally observe a mortality advantage in first-generation immigrants. Explanations for this are the ‘healthy-migrant effect’, with only the healthiest individuals emigrating; and the ‘salmon-bias hypothesis’ which expects migrants to return to their country of origin when they get ill. Yet, due to ‘acculturation’ over time and generation, mortality patterns of migrants may converge to those of the host population.

This study wants to expand the knowledge on immigrant’s health by testing these epidemiological theories on a nationwide-scale. We will use nationwide individually-linked data of the Belgian censuses of 1991 and 2001 and register data on mortality and emigration for the 1990s and 2000s. The database is a unique source of information containing data on mortality, emigration, causes of death, and background characteristics of all individuals legally residing in Belgium at the time of the census. Firstly, we will probe into absolute and relative mortality differences between immigrants of different origins and the Belgian host population, for all major causes of death. Secondly, we will calculate these mortality differences both in the 1990s and 2000s to assess the trend over time. Thirdly, we will analyse whether these mortality differences can be explained by length of stay, migration generation, or socioeconomic position. Thirdly, we want to add the aspect of gender into the question.