Page 131 - IJPS-2-2

Priyanka Dixit and Laxmi Kant Dwivedi

rent picture of the consistency in utilization of institutional delivery across successive births and
its determinants has been provided. Therefore, the objectives of this paper are: (i) to address the
issue of whether women were consistent in delivering births in an institution over successive preg-
nancies, and (ii) to examine the factors associated with the consistent utilization of institutional deli-
very in India.
2. Data and Methods

2.1 Data Sources

Data for this study come from the third round of the cross sectional National Family Health Survey
(NFHS) in 2005–2006 (hereafter NFHS-III). The NFHS samples are nationally representative and
are designed to produce estimates at the national and state levels in India. Information related to ma-
ternal and child health care is available in each successive round of the NFHS. The NFHS surveys
utilize a hierarchical (systematic multistage stratified) random sample design from area-based sam-
pling frames to select clusters (the primary sampling units). Clusters are single-stage, equal probabil-
ity samples of segments of equal size. Once clusters are selected, households within clusters are se-
lected in order to obtain a nationally representative probability sample of individuals (Pamuk, Fuchs,
and Lutz, 2011). Overall, the NFHS adopted a two-stage sample design in rural areas and three-stage
sample design in urban areas (see IIPS and ORC Macro, 2007 for further details regarding sampling).
The NFHS has designed a weight to account for the differential non-response rates of household in-
terviews in the domains. The details of the sampling weights are given in the NFHS reports of vari-
ous rounds. To date, four waves of the NFHS have been conducted.

2.2 Measurements
In NFHS-III, mothers who gave birth during the five years preceding the survey were asked about
the place of delivery. Information regarding place of delivery is available for all of the succes-
sive births during the five years preceding the survey. Births delivered at health facilities — such as
public institutions, NGO trusts, and private hospitals — are termed ‘institutional delivery, ’ whe-
reas births delivered at own homes, parents’ homes, and other homes are termed ‘home delivery’
(Dixit, Dwivedi, and Ram, 2013a).
It is difficult to identify the factors that are most influential in the decision to utilize institutional
delivery services across three recent successive births. Based on previous literature (Babalola and
Fatusi, 2009; Baru et al., 2010; Stephenson et al., 2006; Thind et al., 2008), this study included
the following socioeconomic and demographic factors: region (central, north, east, northeast, west,
or south), place of residence (rural or urban), religion of the head of the household (Hindu, Muslim,
or others), caste of the head of the household (scheduled tribe, scheduled caste, or others), wealth
index for the household (poorest, poor, middle, rich or richest), educational status of women and
their partners (illiterate, literate but below primary school, primary school completed but below mid-
dle school, middle school completed but below high school, or high school or above), loss of a child
(no child lost or at least one child lost), sex composition of children (no son and no daughter, number
of sons greater than daughters, number of sons less than daughters, or equal sons and daughters),
ever having pregnancy terminated (no or yes), child’s age in months (0–23, 24–35, 36–47, or 48–59),
and birth order (second order birth, first order, third order, or 3+ order). The first category of each of
these independent variables is considered as the reference category in the multivariate analyses.
2.3 Analytical Strategy

The main focus of the study is to assess the consistency of institutional deliveries over successive
pregnancies. However, the consistent utilization of services can be influenced by community-level
(Primary Sampling Unit (PSU), or village) factors, beyond individual level factors. Due to this
nested structure, conventional single-level logistic regression models may not be appropriate. Alter-

International Journal of Population Studies | 2016, Volume 2, Issue 2 125