Mortality differentials between self-settled refugees and South Africans in rural South Africa: insights from the Agincourt HDSS, 1994–2018

Abstract

Low- and middle-income countries (LMICs) host most of the world’s refugees but often have limited mortality information systems. Consequently, evidence on the long-term mortality trajectories of self-settled refugees and the contribution of socioeconomic disadvantage to mortality inequalities remains scarce. To examine long-term mortality patterns, cause-specific mortality and the contribution of socioeconomic status (SES) to mortality differentials among Mozambican self-settled refugees, South Africans, Mozambican migrant workers and the population of Manhiça, Mozambique. We analysed longitudinal data from the Agincourt Health and Demographic Surveillance System (1994–2018) and the Manhiça Health and Demographic Surveillance System (1999–2018). Age-standardised death rates were used to describe mortality trends across population groups, while logistic and multinomial logistic regression models examined all-cause and cause-specific mortality after adjustment for age, sex, time period and household socioeconomic status. Mortality increased across all populations during the HIV pandemic, peaking in 2004–2008 before declining thereafter. During 1994–1998, mortality among Mozambican self-settled refugees was broadly comparable to that of South Africans but subsequently exceeded both South Africans and Mozambican labour migrants during the HIV pandemic. By 2014–2018, mortality among refugees had converged towards that of the South African host population. Mozambican labour migrants consistently experienced the lowest mortality. Adjustment for household SES largely explained the overall mortality difference between refugees and South Africans, although refugees remained at higher risk of HIV/AIDS and TB and other communicable disease mortality. Mortality among Mozambican self-settled refugees progressively converged towards that of the South African host population over the 25-year study period. Socioeconomic disadvantage explained an important proportion of the observed mortality inequalities, highlighting the importance of policies that reduce socioeconomic disparities while maintaining equitable access to healthcare.