Page 101 - IJPS-2-2

Abha Gupta, Pushpendra Kumar and Olalemi Adewumi Dorcas

was classified into primary, secondary, and above secondary completion. The marital status of moth-
ers was grouped into currently married and others (married but Gauna not performed, separated,
deserted, divorced and widowed). Ages of the women at first birth were grouped into less than 20
years and above 20 years old. Birth order of the women ranged from 1 to 4 or higher. The sex of
the child included boy and girl. The mass media exposure was categorized as ‘no mass media expo-
sure’ and ‘any mass media exposure’ (included those sources from which women received infor-
mation about full ANC such as watching the TV, listening to the radio and reading the newspapers).
Household wealth index was calculated by combining household amenities, assets, and durables us-
ing factor analysis (Rutstein and Johnson, 2004). Region was measured by district, namely Garhwa,
Palamu, Chatra, Hazaribagh, Kodarma, Giridih, Deoghar, Godda, Sahibganj, Pakur, Dumka,
Dhanbad, Bokaro, Ranchi, Lohardaga, Gumla, western Singhbhum, eastern Singhbhum, Simdega,
Seraikella, Latehar and Jamatara.

2.4 Analytical Approach
The analytical approach considers that inequalities in maternal health care utilization are
mainly caused by socio-economic differences among the population (Obiyan and Kumar, 2015;
Singh, Kumar, Rai et al., 2014; Tsawe, Moto, Netshivhera et al., 2015). These socio-economic diffe-
rences are also considered to influence full ANC utilization in Jharkhand state. Therefore, concentration
index, proposed by Wagstaff and colleagues (1991), was calculated to estimate the socio-economic
inequality in full ANC utilization. It was also decomposed to quantify the factors which led to such
disparities. The decomposition analysis evaluated the proportional contribution of each factor in
generating imbalances. Decomposition and other estimates were computed using STATA version 12
(StataCrop LP, College Station, Texas 77845, USA).
2.4.1 Concentration Index
Concentration curve is generally used to identify the socio-economic inequality in a health variable.
It examines whether inequality exists in one group or not. However, it does not estimate the magni-
tude of inequality (O’Donnell, VanDoorslaer, Wagstaff et al., 2008). Therefore, in this paper, a con-
centration index is used as a method to measure the degree of socio-economic inequality in the utili-
zation of full ANC services. It can be computed as twice the area between the concentration curve
and the line of equality (the 45-degree line). The zero value of the concentration index indicates that
there is no socio-economic inequality. A negative value means the disproportionate concentration of
full ANC among the poor group while a positive value indicates the concentration of full ANC
among the rich group (Szabo, Hajra, Baschieri et al., 2016; Wagstaff, Paci, and VanDoorslaer, 1991).
The index value lies between –1 and 1. For measuring the socio-economic inequality in full
ANC care, the concentration index (C) can be obtained by using the following formula:
2 n 1
1
C = ∑ hr −− (1)
ii
Nµ i= 1 N
where h denotes the health sector variable, µ is its mean and r = i/N is the fractional rank of
i
i
individual in the socio-economic distribution with i= 1 for the poorest and i = N for the richest. A
more convenient formula for the concentration index is given in equation (2) which defines concen-
tration index in terms of covariance between health variable and a fractional rank in socio-economic
distribution (Kakwani, Wagstaff, and VanDoorslaer, 1997; Van Doorslaer and Koolman, 2004).
2
C = cov w (y r (2)
)
ii
µ
th
where y and r are respectively the health status of the i individual and the fractional rank of
i
i
th
the i individual (for weighted data) regarding the index of household economic status; µ is the
(weighted) mean of the health variable in the sample and cov denotes the weighted covariance.
w
International Journal of Population Studies | 2016, Volume 2, Issue 2 95