Data sources
This study has utilized the data from the 2016 Nepal Demographic and Health Survey (NDHS) datasets. The NDHS is a nationally representative cross-sectional survey conducted under the aegis of the Ministry of Health of the Government of Nepal. In the 2016 NDHS, 11473 households were selected as samples of which 11,040 were interviewed. Similarly, from among 13,089 women aged 15–49 identified for individual interviews, only 12,862 women were successfully interviewed with the response rate of 98%. To select the household for the survey, stratified two-stage cluster sampling was used in rural areas whereas three-stage in urban ones. In rural areas, wards were selected as primary sampling units (PSUs) in the first stage and the households were finalized in the second stage. In urban areas, wards were selected as PSUs in the first stage, one enumeration area (EA) was selected from each PSU in the second stage, and households were selected from sample EAs in the third stage. Next, face-to-face interviews of eligible women aged 15–49 in the sample households were conducted with structured questionnaires.
There were six questionnaires administered in the 2016 NDHS: the household questionnaire, the women’s questionnaire, the men’s questionnaire, the biomarker questionnaire, the fieldworker questionnaire and the verbal autopsy questionnaire. This study used the data collected through women’s questionnaire which was utilized to collect information from among the married women aged 15–49 years. The women’s questionnaire sought the information of background characteristics, reproductive history and child mortality, knowledge and use of family planning methods, fertility preferences, delivery care, child health, women’s works, antenatal delivery and postnatal care, husband’s background characteristics and domestic violence [18]. The Individual Record (IR) data file was employed for this study. We confined the analyses to the 1593 young women who were currently married and non-pregnant for modern contraceptive use and 1606 young women who had birth in the 5 years preceding the survey for ANC visits and SBA for their most recent birth.
Variables
Dependent variables
Here, three reproductive health indicators were selected as dependent variables for the analysis. The first dependent variable is the current use of modern contraception which is dichotomous. Young women who were currently using modern contraception were coded as 1 and otherwise as 0. The modern contraceptive methods include male and female sterilization, injectables, intrauterine devices, pills, implants, male condoms, locational amenorrhea, and emergency contraceptives.
The second dependent variable is ANC visits for the most recent pregnancy in the 5 years preceding the survey, which is also dichotomous. The number of ANC visits were coded as 1 if women attended four or more ANC visits before their most recent birth and coded as 0 if there were fewer than four visits. Then the third dependent variable is the assistance of the SBA which is also dichotomous. Here, women assisted by the SBA during most recent delivery were coded as 1 and otherwise as 0. The SBAs include doctors, nurses, and midwives [18, 19].
Independent variables
The independent variables of this study are women’s age, education, ethnicity, occupation, wealth quintile, birth order, number of living children, participation in household decision-making and exposure to media. The age groups of women in this study are 15–19 and 20–24. Mother’s age at birth was classified into two age groups- < 20 and 20–24. Education level achieved by young women was classified into four categories: no education, primary, some secondary, and SLC and above. Moreover, caste/ethnicity variable was categorized into five groups- Brahman/Chhetri, Terai caste, Dalit, Janajati, and Muslim/other. Women’s working status was grouped into two categories- not working, and working. In the same way, wealth quintile variable was recorded into five categories- poorest, poor, middle, richer, and richest. Wealth quintile is a composite measure of household economic status composed of key household assets. Household wealth quintile was constructed using data of household assets such as televisions, bicycles, materials used for house construction, sources of drinking water, sanitation facilities and other wealth related characteristics [20]. Birth order was measured in three categories- 1, 2, and 3 or more. Similarly, the number of living children was measured in three categories- 0, 1–2, and 3 or more whereas fertility preference of young women was categorized into three groups- wants no more children, wants more children, and undecided. Women’s participation in household decision-making was studied in terms of three decisions: women’s own healthcare, large household purchases, and visit to family or relatives. Each of these three aspects of decision-making had five response options: respondent alone, respondent and husband/partner jointly, husband/partner, someone else and other. These responses were coded as 1 for yes (respondent alone, and respondent and husband/partner jointly) and 0 for no (husband/partner, someone else, and other). Then, a single composite variable was constructed by grouping women into two categories: women who participated in at least one decision were categorized as participated one or more decisions (=1) otherwise as not participated (=0). Media exposure is dichotomous variable. Finally, women exposed to either radio, television or newspapers/magazines at least once a week were coded as exposed (=1) and those exposed to none as not exposed (=0).
Methods of data analysis
This study employed three levels of data analyses. Firstly, univariate analysis was carried out to analyze the selected socio-demographic characteristic of young women aged 15–24. Secondly, the chi-square test was carried out to examine the association between the modern contraceptive use, at least four ANC visits and the use of SBA, and explanatory variables such as socio-demographic, economic, women’s participation in household decision-making and media exposure. Third, multivariate analysis with logistic regression was performed to examine the factors influencing the use of reproductive health services. Likewise, factors significantly associated with outcome variables at p < 0.05 were included in multivariate logistic regression. Multicollinearity assessment was employed before the multivariate analysis. The results were shown in adjusted odds ratios (OR) with 95% confidence interval (CI). The data were analyzed by using STATA version 15.1. All calculations were weighted using standard sample weight of NDHS. The STATA survey command ‘svy’ was applied to adjust for stratified sample design.