Page 28 - IJPS-2-1

Monitoring adult mortality by type of residence in the absence of death registration: a perspective from Burkina Faso

on the Demographic and Health Surveys (DHS), show that the urban advantage is weakening over-
time. This is partly due to the development of slums associated with rapid and poorly managed urba-
nization in Sub-Saharan Africa (Kimani-Murage, Fotso, Egondi et al., 2014). This particular focus
on child survival fit well within the context of the Millennium Development Goals (MDGs), but
adult health is now as important on the global health agenda. Coupled with the strong desire to place
equity in a central role in the Sustainable Development Goal (SDG) framework (“no one should be
left behind”), disaggregation of adult mortality indicators by urban/rural location is now imperative
to track progress in a world that is becoming both older and more urban (United Nations, 2015).
A better monitoring of adult mortality by place of residence is more important than ever in
Sub-Saharan Africa; available evidence shows that rapid urbanization could be a threat to adult
health. In contrast to most rural areas where infectious and parasitic diseases remain the
leading causes of deaths, urban settings of Sub-Saharan Africa are often characterized by a
“double burden of diseases,” particularly among urban poor (Ramroth, Lorenz, Rankin et al., 2012;
Awini, Sarpong, Adjei et al., 2014; Agyei-Mensah and de-Graft Aikins, 2010). While wealthy adults
are at higher risk of non-communicable diseases due to lifestyle (smoking, alcohol consumption,
physical inactivity), and inadequate diet habits (diets too rich in sugar, salt, fat), the urban poor bear
simultaneously the burden of infectious and non-communicable diseases (Soura, Lankoande,
Millogo et al., 2014; Mberu, Wamukoya, Oti et al., 2015). For instance, slums are a favorable
ground for some infectious diseases such as malaria and tuberculosis, and at the same time their
inhabitants tend to consume cheaper energy dense food that promote obesity, a risk factor for
non-communicable diseases (Neiderud, 2015; Zeba, 2012). Furthermore, mental disorders and
injuries seem to be more prevalent in urban settings of Sub-Saharan Africa (Dyson, 2003;
Kobusingye, Guwatudde, and Lett, 2001). In summary, it is simplistic to always consider the urban
environment as a “safer place” for adults, compared to rural areas. Poor sanitary and
living conditions experienced by rural dwellers may be offset by health problems related specifically
to urban residence. Despite its relevance for the global health agenda, research on urban/rural differentials in adult
mortality in Sub-Saharan Africa is still largely limited by the scarcity of data. While much is known
about child mortality, the measurement of adult mortality is hampered by the lack of reliable data
and the absence of very robust estimation methods. In the absence of civil registration data, esti-
mates derived from surveys and censuses are prone to recall errors and selection biases (Reniers,
Masquelier, and Gerland, 2011). These problems are magnified when looking at differentials.
Census reports on the number of household members who died in the last 12 months are a com-
mon source of data on adult mortality by place of residence. However, these data are subject to many
errors, including omissions of deaths due to recall errors and household dissolutions after a death of
an adult, but also fieldworkers’ related errors, coverage errors, and errors on the reference period
(Timæus, 1991). Adult mortality rates published in census reports are sometimes also inferred
from child mortality rates combined with model life tables, but this practice is not recommend-
ed because mortality in children and adult do not always evolve in the same direction (Masquelier,
Reniers, and Pison, 2014). This approach is even more problematic when deriving estimates of adult
mortality by place of residence, because nothing guarantees that urban/rural differentials are inva-
riant by age. Child mortality rates are dominated by infectious diseases which are more prevalent in
rural areas, while chronic conditions that disproportionately affect adults are typical of urban areas.
Apart from census estimates, it is possible to derive mortality rates from survey reports on the
survival of close relatives, such as parents or siblings. Nevertheless, estimates based on these me-
thods are also not exempt of problems (Helleringer, Pison, Kanté et al., 2014; Reniers, Masquelier,
and Gerland, 2011). Mortality rates obtained from the survival of parents are related to the past and
dating estimates is only possible under some assumptions. Data on survival of siblings provide the
opportunity to directly estimate adult mortality but estimates may be altered by underreporting of
deaths. Nevertheless, beyond these general issues related to the estimation, deriving estimates of
adult mortality by place of residence is even more challenging since information on the place of res-

22 International Journal of Population Studies | 2016, Volume 2, Issue 1