Return to search

A longitudinal study of migration and it relation to AIDS/TB mortality in rural South Africa

A thesis submitted to the Faculty of Humanities, University of Witwatersrand, Johannesburg, South Africa in fulfilment of the requirements of the Degree of Doctor of Philosophy in the field of Demography and Population Studies. / Background: In exploring the relationship between migration and HIV/AIDS, a
focus of earlier studies was on the role of the mobile population in the
geographical spread of the disease. There has been a shift in this perception and
the focus now is on the implications of being a migrant. A body of literature
has developed on the risk of migrants contracting HIV, but only a few studies
have examined the AIDS/TB mortality risk as a consequence of migration, with
the results showing that migrants have higher chance of dying of AIDS/TB
compared to their non-migrant counterparts. However, these studies mainly
looked at the impact of migration on mortality due to AIDS/TB and did not
make provision for the presence of other causes of death. Therefore, this study
is geared towards investigating migration as it relates to death caused by
AIDS/TB, longitudinally, and in the presence of other causes such as non
communicable diseases, other infectious diseases, and external causes of death,
in rural South Africa. Specifically, the study addressed the following questions:
(i) What is the risk of dying from AIDS/TB among migrants in rural South
Africa in the presence of other causes of death? (ii) How does this relationship
compare with the relationship between migration and other causes of death? (3)
What are possible predictors of the relationship between migration and
AIDS/TB in the presence of other causes of death?

Method: This research project is part of a longitudinal study of the inhabitants
of the Agincourt sub-district, situated in the rural north-eastern part of South
Africa. The study utilises the Agincourt Health and Demographic Surveillance
System data spanning 12 years, starting from 1st January, 2000 to 31st
December, 2011. The main target group for the study is individuals aged 20 to
69 years at the date of analysis. The selected individuals are divided into the
following categories: (i) the return migrants who returned after spending a
period of time outside the study area; (ii) the in-migrants who moved into the
study location for the first time, and (iii) the permanent residents (non
migrants). A six month residence threshold period is used to distinguish
participants from ordinary visitors. The migration status categorical variable
was further expanded from three to five categories with in-migrant and return
migrant categories being split to accommodate short and long-term durations of
exposure. In the year 2000, the baseline year, a total of 25,621 individuals who
met the entry criteria were recruited into the study.

For data analysis, a Fine and Gray model is used, which is a variant of a Cox
proportional hazard model, to estimate the competing risk of dying among the
selected participants by sex. The causes of death (CoD) variable was
categorised into the following broad categories: “AIDS/TB”, “Non
Communicable Disease”, “External cause” and “Other infectious disease”, with
indeterminate causes coded as missing. The five categories of migration serve
as the independent variable, with permanent residence acting as the reference
group, while the broad Cause of Death categories are the main dependent
variables. Other dependent variables are: period, nationality, education and
socio-economic status.

Results: This first set of results aims to address the question on the risk of
AIDS/TB mortality among migrants in rural South Africa in the presence of
other causes of death. The findings are that male and female short-term return
migrants have significantly higher relative risk of dying of AIDS/TB death
when compared to their non-migrants counterparts with sub-hazard ratio (SHR)
of 4.87 (95% CI 4.17-5.72; P<0.001) and 5.44 (95% CI 4.64-6.38; P<0.001))
reported for both gender group respectively. For male and female long-term
return migrants, their SHR was 1.80 (95% CI 1.43-2.26; P<0.001) and 2.06
(95% CI 1.57-2.70; P<0.001) respectively. The results did not reveal significant
results for the in-migrants.

The second set of results aims to address the second research question, which
is, how does the relationship between migration and mortality caused by
AIDS/TB in rural South Africa in the context of other causes of death compare
with the relationship between migration and causes different from AIDS/TB.
The results show that Short-term return migrants have higher mortality than
non-migrants, whatever the four causes of mortality. For instance, the
competing risk of death due to AIDS/TB for short-term return migrants
compared to non-migrants showed a lower SHR for external cause of death,
namely 8.78 (95% CI 5.86-13.16; P<0.05) vis-à-vis non-migrants. This implies
that the difference in the relative risk of mortality between migrants and non
migrants is even higher for external causes than for AIDS/TB. The same is
applicable to the risk of death from other infectious diseases for females, which
has a SHR of 4.97 (95% CI 2.50-9.89; P<0.05) in the competing risk model.
The relative risk of death due to AIDS/TB for male is 4.87 (95% CI 4.14-5.72
P<0.001) while that of female is 5.44 (95% CI 4.64-6.38; P<0.001);
respectively.

With regards to the question on the possible predictors of the relationship
between migration and AIDS/TB in the presence of other causes of death, it is
shown that period is one of the predictors of the relationship between migration
and AIDS/TB mortality. And, it is relevant to the study participants who died
as a result of AIDS/TB, NCDs and other infectious diseases. In general, the risk
dwindles in the latter period when the antiretroviral drugs become available for
AIDS/TB. Nationality is also a determinant of the relationship and it is
applicable to those who lost their lives due AIDS/TB (female only), NCDs and
other infections (female). In all, the Mozambican nationals are less likely to die
in comparison with the South Africans. Educational status is a predictor and it
relevance cuts across virtually all the causes of death. The dominant pattern that
is revealed in this context is that the higher the level of education, the lower the
risk of death due to any of the causes. The predictive impact of SES can only be felt among the respondents whose death was due to AIDS/TB and NCDs
(female only).

Conclusion: With circular labour migration in South Africa showing no
evidence of declining and with the attendant mortality risks due to AIDS/TB
and other causes, and needs to be carefully considered - in policies aiming to
control mortality in South Africa. Disease-induced migration creates burdens
not only for the left-behind families in terms of their means of livelihood
through loss of remittances, but also for the burden on health care facilities in
the rural area. With short-term labour migrants being a high risk group, the
success of intervention programmes addressing the problem of HIV infection
and the resultant mortality implication, such as ‘treatment as prevention’
programmes, can only be guaranteed by recognising the risks incumbent on this
group of people and the influence of the larger communities. / XL2018

Identiferoai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/24476
Date January 2017
CreatorsAfolabi, Sulaimon Atolagbe
Source SetsSouth African National ETD Portal
LanguageEnglish
Detected LanguageEnglish
TypeThesis
FormatOnline resource (xxiv, 244 pages), application/pdf, application/pdf

Page generated in 0.003 seconds