Factors associated with place of delivery in rural Nepal

The outcome variable of the study was institutional delivery. Institutional delivery was defined as the practice of delivery/giving birth by a mother in a health institution between 15 July, 2010 and 14 July, 2011. The outcome variable was validated by asking the mother for the name of the health institution she visited and about the type of health workers who assisted her during delivery.

The explanatory variables were studied under three categories: individual, household and health service factors. The roles of individual factors, which included mother’s current age, parity, past experience of adverse pregnancy outcome, literacy, occupation and antenatal care practice, were assessed. Literacy and occupational status of the study participants’ husbands’ was also assessed. The age of the respondent as reported by the individuals was recorded. For analysis purposes, age was categorized into four groups (less than 19, 20–24, 25–30 and above 30 years). Parity was classified as primiparous (mothers who had delivered only one child) and multiparous (mothers who had delivered more than one child).

Literacy status was identified by asking mothers whether they had attended formal school, and if they had, then the total number of years of formal schooling was recorded and the respondent was later categorised into one of the four levels: primary, secondary, intermediate, or bachelor and above. If the mother did not go to school, she was requested to read a card with a short statement on nutrition in order to identify whether she was just literate to a degree or illiterate entirely. During descriptive analysis, six levels of education were recognised: illiterate, just literate, primary, secondary, intermediate, bachelor and above. However, for bivariate analysis, these levels were reduced to two categories: illiterate and literate. The literacy status of the respondent’s husband was measured based on the response of the mother. Occupation was broadly classified into four categories: agriculture, labour, business and service. During bivariate analysis, occupation was regrouped into labour-intensive occupations (agriculture and labour) and non labour–intensive occupations (service/business). Experience of adverse pregnancy outcome was assessed by asking mothers whether they had experienced stillbirths or abortion during their lifetime. Mothers were requested to recall the frequency of antenatal care visits during last pregnancy, along with the services they received. The mothers were then categorized as either having had less than four visits, and those who had four or more visits.

In household factors, the association of ethnicity, socio-economic status (wealth quintile), women’s autonomy, and family type were examined. The Health Management Information System section of the Department of Health Services of Nepal has classified ethnicity into the following categories: Dalit, disadvantaged Janjati (indigenous people), relatively advantaged Janjati, other excluded groups including Terai based ethnic group (except Terai Brahmins), religious minorities and Brahmin/Chhetri. During multivariate analysis, the ethnic groups were classified into two groups: privileged and underprivileged ethnic groups. Brahmin/Chhetri and relatively advantaged Janjati (Newar, Thakali and Gurung) were classified as privileged ethnic groups, while Dalit, disadvantaged Janjati, Terai caste (excluding Terai Brahmins) and religious minorities were classified as underprivileged ethnic group; which was based on economic and educational differences between the two groups [14]. Socio-economic status was ascertained by measuring each group’s placement over the wealth quintile (used in the Demographic Health Survey), which involved principle component analysis (PCA) of details of household assets and expenditures [15]. Households were categorised into five ranked groups; the first rank representing the poor and the fifth rank representing the least poor. Women’s autonomy was measured by the help of 12 variables reflecting autonomy areas such as control over finances, decision-making power and the extent to which women have freedom of movement; adapted from a study in Pakistan [5]. Using PCA techniques, households were categorised into three ranked group. The first rank represented low autonomy group to third rank representing high autonomy group.

At the health institution level, the association of time taken to reach health institution, exposure to maternal health messages during pregnancy, and perceived behaviour of service provider during antenatal care with institutional delivery was assessed. For analysis purposes, the time taken to reach the health institution was divided into two categories: less than 60 minutes and more than 60 minutes, based on a study in Nepal [11]. Exposure to message was defined as the exposure of the respondents to messages relating to maternal health and service utilization during pregnancy through mass media and health workers. Mothers were requested to recall and rate the behaviour of health workers during their first antenatal care visit. The behaviour of health workers was classified as favourable or unfavourable, as perceived by the respondents.