Page 85 - IJPS-10-4

International Journal of
Population Studies Migration and child mortality estimation

due to rural and urban migration in Kenya. We focused (ii) Rural non-migrants: Women who always resided in
on rural and urban regions primarily because of the their current rural areas or moved from one rural area
relatively high prevalence of rural-to-urban migration in to another.
Kenya, driven by a high rate of urbanization. According (iii) Urban-to-rural migrants: Women who moved from
to the United Nations (2019), the urban population in an urban region after some birth experience and
Kenya increased by 4.36% between 2010 and 2015 and resided in a rural region during the survey date.
was projected to increase by 4.23% between 2015 and (iv) Rural-to-urban migrants: Women who moved from
2020. Numerous studies have also highlighted disparities rural areas after some birth experience and were living
in child mortality between rural and urban regions or in an urban area at the time of the survey.
among various population segments defined by rural-
urban migration status (Bocquier et al., 2011; Brockerhoff, 2.3. Estimation of child mortality rates
1994; Issaka et al., 2017; United Nations, 2019; Yaya et al., In our investigation, we focused on three age-specific
2019). According to Schmertmann & Sawyer (1996), mortality rates in childhood, namely infant mortality,
the migration of women between regions with different one-to-four mortality, and under-five mortality rates. The
mortality regimes can lead to erroneous child mortality infant mortality rate is the probability of dying before the
estimates in those regions. first birthday, expressed as the number of children who die
before age one per 1000 live births in a given year. The one-
2. Data and methods to-four mortality rate is the probability of dying between
2.1. Data sources the first and fifth birthdays, computed as the number of
children who die after 1 year but before their fifth birthday
The data used in this study were pooled from six Kenya among 1000 children who survived to the first birthday.
Demographic and Health Surveys (KDHS) carried out The under-five mortality rate measures the probability
between 1989 and 2014. The KDHS are national surveys
that collect data for monitoring and evaluating the impact of a child dying before the fifth birthday, expressed as
of various demographic and health programs. The KDHS the number of children who die before reaching age five
datasets include data on birth history, age of women and among 1000 live births (Etikan et al., 2019; KNBS & ICF
Macro, 2015).
children, previous place of residence, and duration of stay
in the current place. Pooling estimates from several surveys Since the indirect method uses summary birth
and was necessary because a single survey produces a set of history data, the full birth history data from KDHS
seven mortality estimates, which would not be sufficient for were summarized into two variables: The total number
a statistical comparison. Several studies have used a similar of live births and the number of surviving classified by
approach of data pooling to estimate child mortality rates the mothers’ age group. The age of the women, which is
or to construct models that, in turn, use summary birth normally taken as the proxy measure of exposure (Arthur
history data to estimate child mortality indirectly (Ayele & Stoto, 1983; Bangura et al., 2016; Rajaratnam et al.,
et al., 2016; Hallett et al., 2010; Verhulst, 2016; Walker 2010), was classified into seven 5-year age groups: 15 – 19,
et al., 2012; Yadava & Tiwari, 2003). 20 – 24, 25 – 29, 30 – 34, 35 – 39, 40 – 44, and 45 – 49.
The Demographic and Health Survey data are There are four models used for indirect child mortality
available in the Statistical Package for the Social Sciences estimation: The North model, the South model, the East
(IBM, 2020). model, and the West model. Each of these estimation
models produces seven estimates for each of the three child
2.2. Migration status classification mortality rates: The infant mortality rate, the child mortality
Our focus was on two types of officially categorized rate, and the under–five mortality rate. These rates were
residential regions: Rural and urban regions. These regions computed using the QFIVE program (United Nations,
are mutually exclusive and exhaustive, and every cluster 2013). The program generates the estimates using
sampled for the survey belongs to either of the two regions. Trussell’s regression equations, which are based on the
During surveys, women were asked to state their previous Coale and Demeny model life tables (Coale & Demeny,
place of residence, categorized as either rural or urban. 1966) or using the Palloni-Heligman equations based on
Based on current and previous place of residence, women the United Nations life table models (United Nations,
fell into one of the four migration statuses: 1983). Estimates based on Trussell’s model were preferred
(i) Urban non-migrants: Women who either never because it is the third generation of Brass variant models
moved from their current urban residence or moved with higher flexibility. Trussell’s equations fit empirical
from one urban area to another. data better and are less affected by random errors, which

Volume 10 Issue 4 (2024) 79 https://doi.org/10.36922/ijps.1837