Migration and mortality: identifying selection and exposure effects in local populations using longitudinal data

Philippe Bocquier, Université Catholique de Louvain
Carren Ginsburg, University of the Witwatersrand
Donatien Beguy, African Population and Health Research Center (APHRC)
Sulaimon A. Afolabi, University of the Witwatersrand
Orvalho Augusto, Manhiça HDSS
Kobus Herbst, Africa Centre for Health and Population Studies
Yempabou Bruno Lankoande, Université Catholique de Louvain et Institut Supérieur des Sciences de la Population (ISSP)
Frank Odhiambo, Kenya Medical Research Institute (KEMRI)
Mark Otiende, KEMRI Wellcome Trust Research Programme (KWTRP)
Bassiahi Abdramane Soura, Université de Ouagadougou
Pascal Zabré, Centre de Recherche en Santé de Nouna - Burkina Faso
Michael J. White, Brown University
Mark Collinson, University of the Witwatersrand

Migration has been hypothesised to be selective on health but this healthy migrant hypothesis has generally been tested at destinations, and for only one type of flow, from deprived to better-off areas. The circulatory nature of migration is rarely accounted for. This study examines the relationship between different types of internal migration and mortality in Health and Demographic Surveillance System (HDSS) populations in West, East, and Southern Africa, and uses Event History Analysis techniques to explore how the processes of selection and exposure explain the migration-mortality experienced in these contexts. The study confirms the importance of migration in explaining variation in mortality, and the diversity of the migration-mortality relationship over a range of rural and urban local areas in these regions. The study concludes that the pattern of migration-mortality relationship is mainly generated by the combination of two processes: selection and exposure. However, there are many exceptions to the expected patterns.