Page 30 - IJPS-2-1

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

(INSD, 2009c). The capital is among the fastest-growing cities in the world. Even though the popu-
lation is essentially rural (77% in 2006), it is expected that urban population will take over in 2050
(United Nations, 2014). However, the health indicators in Burkina Faso are not very encouraging.
The under-five mortality rate is still high, reaching a level of 10.4% in urban areas and 15.6% in rural
areas (INSD, 2012). Data on adult mortality are scarce, but the probability 45q 15, estimated at 30.4%
in men and 26.8% in women in 2005, was higher than that observed in countries such as Senegal and
Niger (IHME, 2014). According to IHME (2015), in 2013, communicable diseases were the lead-
ing causes of death in adults aged 15–49 while non-communicable diseases represented the ma-
jor causes of death in adults aged 50–69.

2.2 Data
Adult mortality is measured as the probability of dying between 15 and 60 years of age ( 45q 15). Three
types of data sources are used in this paper to estimate mortality levels. The corresponding sample
sizes by urban/rural location are presented in the Appendix (Table A3). It is important to mention
that the definition of urban and rural residence changes over time, from one census to another. An
economic criteria was used in the 1996 census, but an economic and population size criteria was re-
tained in 2006 (INSD, 2009c). The definition used in DHS is based on the definition adopted in the
most recent census (INSD, 2004; INSD, 2012).

2.2.1 Census Reports on Recent Household Deaths
To date, four censuses have been conducted in Burkina Faso to monitor demographic trends (1975,
1985, 1996, and 2006). However, it is unclear whether data from the 1975 census have been properly
archived. Samples of individual-level data from censuses conducted in 1985 and 1996 are freely
available online through the Integrated Public Use Microdata Series (IPUMS), but the urban/rural
status is missing. Thus, only the census of 2006 is used here to estimate adult mortality, based on
data collected on the number of deaths in each household in the twelve months preceding the census.
2.2.2 Data on Sibling and Parental Survival from Demographic and Health Surveys (DHS)
Data from DHS conducted in Burkina Faso in 1993, 1998/1999, 2003, and 2010 are used in this
study. Funded by the U.S. Agency for International Development (USAID), these surveys are a key
data source to assess population dynamics and their health in countries lacking vital registration sys-
tems. The data are freely available online and are representative at the urban/rural level. In the 1993,
2003, and 2010 surveys, children aged less than 15 years were asked about the survival status of
their parents (mothers and fathers). In surveys conducted in 1998/1999, 2003, and 2010, each wom-
an interviewed was also asked to list all her siblings born to the same mother. For each sibling, in-
formation was collected on their date of birth, sex, survival status, current age for those who were
alive, age at death, and number of years since death for those who had died prior to the survey.
2.2.3 Orphanhood Data from the Migration Dynamics, Urban Integration and Environment
Survey (EMUIB)
Census and DHS data are complemented with a survey (EMUIB) conducted in Burkina Faso in
2000 by the Demography department of the University of Ouagadougou and has never been used so
far to estimate mortality. This survey was representative at the national and urban/rural levels. In
total, 9188 individuals aged between 15 and 64 years old were interviewed. The overall objective
was to provide reliable and relevant information on urban planning in Burkina Faso and topics such
as migration and employment were covered (for a full description of the survey design, see Poirier,
Piché, Le Jeune et al. (2001)). The questionnaire included a set of questions on parental survival and,
unlike in other surveys, the place of residence at the time of survey or at the time of death of parents
was also collected. This is an added value compared with DHS data where information on parents’

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