Mortality statistics in immigrant research: method for adjusting underestimation of mortality

BACKGROUND: It is difficult to carry out fair comparisons of the mortality of different ethnic groups in a population in register-based studies because sizeable numbers of immigrants who subsequently leave their new homeland fail to register this fact with the national registration authorities. In this article we present a method which attempts to address these problems. METHODS: Age-standardized mortality rates for native Swedes and immigrants in the age group 20-64 years were calculated for all individuals who either were included in the Swedish Population Censuses for 1985 or 1990, or who moved to Sweden during the period November 1990-1994. In order to define the population under scrutiny different sources of income are used as indicators of residence in the country. RESULTS: When an analysis is made of all nationally registered individuals, significantly reduced death rates are found among immigrants outside the north-east of Europe compared to those for Swedish-born people. Extremely low death rates are found for those born in Turkey, Southern Europe, Latin America, Asia, and Africa and for those who are younger and without any income. When the income criterion is introduced, there is a change so that the earlier significantly reduced relative death risks for immigrants born outside the north-east of Europe for some subgroups are no longer significantly lowered. CONCLUSION: This study has important implications for the interpretation of every study of mortality among immigrants based on official mortality statistics. Using information about income as an indicator of residence in the country appears to be a method which can be pursued further in order to achieve a more accurate understanding of mortality among immigrant groups.