Study context

The study included four South-Asian countries Afghanistan, Bhutan, Nepal, and Pakistan (Punjab and Sindh provinces). All countries belonged to lower or lower-middle-income countries with high incidences of poverty and social and health inequalities [18]. Among the countries selected in this study, poverty headcount ratio at national poverty lines (percentage of population) was the highest for Afghanistan (54.5%) and the lowest for Bhutan (8.2%). The estimated total population was 37.17 million in Afghanistan, 0.75 million in Bhutan, 28.09 million in Nepal and 212.22 million in Pakistan in 2018 [19]. South Asian nations have low health standard compared to other regions [18]. The life expectancy at birth was the lowest for Afghanistan with 64 years, followed by 67.11 years for Pakistan, 70.48 years for Nepal and 71.46 years for Bhutan [19].

Data source and sampling

The study used data from the Multiple Indicator Cluster Survey (MICS) from four South-Asian countries; namely Afghanistan, 2010–11; Bhutan, 2010; Nepal, 2014; and Pakistan (Punjab, 2014 and Sindh, 2014). The MICS is based on nationally representative samples from Afghanistan, Bhutan and Nepal; however, Pakistan consists of two independent studies from Punjab and Sindh provinces. The surveys were based on a cross-sectional study design and multi-stage sampling methods [20]. At first, enumeration areas (EA) were selected systematically with probability proportional to their size and the required number of samples was selected from each EA in the second stage. Methodology and sampling design details are described elsewhere [4, 21,22,23,24]. A total of 95,616 households were included in this study with an overall response rate of 97.4%, with 129,785 women of reproductive age (15–45 years) being interviewed which accounted for an 89.5% response rate. A total of 26,029 women who were either married or living together with their partners and had a pregnancy outcome (live births) 2 years prior to the survey were included in this study (Additional file 1: Table S1).

Outcome variables

The outcome variables included current utilization of contraceptive methods, at-least one ANC visit, completed four or more ANC visits, institutional delivery and PNC services. The contraceptive methods included both traditional and modern methods. We included both having at-least one ANC visit and four or more ANC visits as the outcome variables. The childbirth at government hospitals, primary health care center, private hospitals, private clinics or health institutions managed by non-governmental organizations (NGOs) were included as an institutional delivery. The PNC indicator included women who had their health check-up within two days of the most recent birth, either at home or at health institutions. The PNC visit information is only available for Nepal and Pakistan.

Exposure variables

IPV usually denotes physical, sexual, and emotional abuse and controlling behaviours by their intimate partners [1]. For the purpose of this study, women’s justification of IPV denotes a condoning attitude towards physical violence perpetrated by their partners. The exposure variable, women’s justification of IPV, was measured through standard tools used by MICS and Demographic Health Surveys (DHS) [6, 20]. It was based on a series of questionnaires that collected information on women’s attitudes towards wife-beating by their husband/partner under different conditions. Women were asked if wife-beating is justified for different conditions. They were specifically asked if wife-beating is justified for going out without informing their husband, neglecting children, arguing with their husband, refusing to have sex with their husband and burning food. A dichotomous variable was created to document women’s justification of IPV if they justified wife-beating for any one of the conditions presented to them.

The variable capturing the levels of women’s justification of IPV was generated to measure the extent to which women justified the conditions for wife-beating, ranging from 0 (not justifying any option) to 5 (justifying all five options). The definition of all outcomes, exposure and other covariates used in the study are provided in Additional file 1: Table S2.

Statistical analysis

The datasets from all four selected countries were merged for the purpose of analysis. The combined dataset constitutes a hierarchical structure with more than one level of clustering, for example in the combined dataset, households are clustered within primary sampling units (i.e. sampling clusters), sampled clusters within countries and countries within a regional level. Therefore, modelling of the outcome variables should take into account the correlations within clusters that vary between them [25].

The study used a generalized linear mixed model with random effects at both cluster and country level with the households nested within the country. The detail description of the methods is provided elsewhere [26] but a brief description of the multi-level model is provided below.

Let, \(y_i\left(j\right)t\) denote the response for woman who lives in cluster i and country j, with 1 = occurrence of event/outcome and 0 = No occurrence of event for the outcome variables. Then, a multilevel model with random effect \(v_i\left( j \right)\\) for clusters and \(\u_\left( j \right)\\) for country and fixed effects for explanatory variables is given by the following equation (27).

(\(\textLogit [\textP(y_i\left( j \right)t ) = 1)] \, = \varvecx_\varveci\left( \varvecj \right)\varvect^\varvecT \beta + u_\left( j \right) + v_i\left( j \right)\).

where \(\varvecx^\varvecT\) represents the vector of explanatory variables for women t living in cluster i and country j. \(\beta\) represents the fixed effect parameter that have conditional interpretations given the random effect. \(v_i\left( j\right)\) denotes level-1 (cluster level) random effects that account for variability among respondents, i.e. women, within a cluster. \(u_\left( j\right)\) denotes level-2 random effects that accounts for variability among countries. The random components \(u_\left( j\right)\) and \(v_i\left( j\right)\) are assumed to be independent with distributions N (0, \(\sigma _u^2\)) and N (0, \(\sigma _v^2\)) respectively.

We calculated the aggregated effect size estimates of contraceptive and maternal health care service utilization associated with women’s justification of IPV at the regional level. The country-level analysis was performed using a multivariable logistic regression model. The model was also used to analyse differences in the time (months) for the first contact with healthcare facilities after pregnancy associated with women’s justification of IPV. A likelihood ratio test was used to evaluate a linear trend in utilization of contraceptive methods and maternal health care services associated with increasing levels of women’s justification of IPV. We also checked for interaction between women’s justification of IPV and the area of residence on the outcome variables.

We controlled for a wide range of confounding variables in the models that were identified in the existing literature [28,29,30]. Variables such as women’s age, women’s education status, area of residence and household wealth quintiles were entered in the model as categorical variables. Likewise, women’s age at first marriage/union, age of husband and number of children ever born were entered in the model as continuous variables.

Sampling weights for women were used to adjust for the complex survey sampling design and non-proportionate selection probability in the analysis. We adjusted for the country level weights in addition to the women weights for pooling the results at the regional level [31]. The P values are 2-sided and statistical significance level set at less than 0.05. This study used Stata 14.1 (Stata Corp, College Station, Texas) for data analysis.

By admin

https://muehlenbar.de/wp-includes/slot-online/

https://the-chef.co/sbobet/

Slot Gacor Terbaru

Link Slot Gacor

Sbobet

Situs Slot Gacor

Slot Gacor 2022

Login Sbobet

Daftar Sbobet

https://www.dcosmeticclinics.com.au/wp-includes/sbobet/

https://thetastesoflife.com/wp-includes/sbobet/

https://www.townshipofsugargrove.com/wp-includes/slot-gacor/

https://texasmamaboutique.com/wp-includes/slot-gacor/

https://bizu-me.com/wp-includes/slot-gacor/

https://tiketa.co.za/wp-includes/slot-gacor/

situs slot gacor

slot gacor gampang menang

slot gacor terbaru

daftar sbobet

slot bonus newmember

link sbobet

sbobet login