Wealth based inequality in the prevalence of teenage pregnancy and its correlates: Evidence from National Family Health Survey, 2019-21 | BMC Pregnancy and Childbirth

Data Source

This study utilizes data from the most recent round of the National Family Health Survey (i.e., NFHS-5) conducted in 2019–2021. NFHS-5 was conducted under the direction of the Ministry of Health and Family Welfare, Government of India, with the International Institute for Population Sciences, Mumbai, as the nodal agency. It is a nationally representative, cross-sectional survey that provides information on fertility, infant and child morbidity & mortality, family planning, and other important aspects of women-child health and nutrition. The survey includes 476,561 households from rural and 160,318 households from urban areas. More information on the sample design and survey details can be obtained from the NFHS-5 Report [10].

This study uses Individual (women’s) files. Based on the outcome variable of interest, women in the age group 15–19 years have been selected for the present study. A total of 122,477 women were eligible for the study. The sample focused on currently pregnant women and/or those with at least one child or ever had a pregnancy.

Variables

Outcome Variable

Teenage childbearing is the main outcome variable of this study. It can be defined as the teenage girls (15–19 years in this study) either currently pregnant or ever pregnant irrespective of their pregnancy outcome. A dummy variable was created and coded as 1 if the teenager was pregnant at the time of the survey or experienced pregnancy and 0 otherwise.

Independent Variables

Based on previous pieces of literature, the following variables are used as predictors of teenage pregnancy and have been classified as individual and community-level factors:

Individual Level Variables

The age of respondents was taken in single completed years at the time of the survey. Current marital status was recoded into two categories (1 if currently married and 0 = otherwise). Religion has subcategories of ‘Hindu’, ‘Muslim’, and ‘Others’ based on their religious beliefs. Caste/Ethnicity was recoded into four categories- ‘SCs’ (Scheduled Castes), ‘STs’ (Scheduled Tribes), ‘OBCs’ (Other Backward Castes), and ‘Others’. Educational attainment of women was categorised into ‘without any formal education’, ‘primary education’, ‘secondary education’, and ‘higher education’. Wealth index was categorized into five quintiles (poorest to richest) derived based on assets and amenities in the household. Employment status was recoded as ‘currently working’ and ‘currently not working’. Watching television (TV), listening to the radio and reading the newspaper at least once a week were merged and a dichotomized variable was generated as “having mass media exposure” and “not having mass media exposure.” The household head’s age and sex were categorized as less than 30 years, 30–44 years, 45–59 years and 60 or above & male and female respectively. The respondents were asked if they ever tried to avoid or delay pregnancy, knowledge of modern methods of contraception and their responses were recoded as yes or no. Based on the response to the question “Have you ever heard family planning on either TV or radio or magazine/newspaper” the variable ‘exposure to family planning mass media’ was coded as (none, one, two, three). Information on knowledge about the ovulatory cycle was recoded as incorrect, correct and don’t know.

Community Level Variables

The place of residence was dichotomized into urban and rural. The region was the geographical location where the population lived. The states and Union Territories of India have been categorized into six regions, namely North, Central, East, Northeast, West, and South.

Statistical analysis

Descriptive Statistics like frequency and percentage have been used to describe the distribution of population. Further, bivariate analysis was computed to examine the prevalence of teenage pregnancy by various socio-economic and demographic characteristics. Also, Pearson’s chi-square test was used to identify the significance of such associations. Since the chi-square test doesn’t indicate the strength (effect size) of a statistically significant relationship, Cramer’s V was used. It is calculated by dividing chi-square values by sample size and taking the square root of this value. The outcome variable of the study is dichotomous so binary logistic regression was applied to determine the association between teenage childbearing and other independent variables in terms of crude odds ratio with a 95% confidence interval. Also, the marginal effect and its 95% confidence interval were calculated. The study found no evidence of multicollinearity among the independent variables using the Variance Inflation Factor (VIF) (mean VIF = 1.14).

To measure the socioeconomic inequality in teenage pregnancy, this study used a concentration index. The concentration curve is a plot of the cumulative share of teen pregnancy against the share of the population in ascending order of the wealth index. The magnitude of the concentration index is twice the area between the concentration curve and the line of perfect equality (45-degree line). If the concentration curve lies above (or below) the line of perfect equality, it suggests that teenage pregnancy is concentrated among the poor (or rich).

The value of the concentration index was calculated using formula-\(\:CI=\frac{2}{\mu\:}\text{cov}\left(h,\:r\right)\)

Where CI indicates the concentration index, is the mean of teenage pregnancy, h is the teenage pregnancy and r is the cumulative percentage that total adolescents represent over the total population after ranking teenage pregnancy by the wealth index.

The concentration index value ranges from − 1 to 1 for continuous outcome variables. Since the outcome variable of this study is binary, the value of the concentration index depends on µ. For the prevalence of teenage pregnancy, an Erreygers CI (ECI) [11], a modified version of the normal CI is used to measure the socio-economic inequality which can be defined as

.

Where CI(y) is the generalized CI and \(\:\mu\:\) is the mean teenage pregnancy. This normalized value of the concentration index is used to quantify socioeconomic inequalities in teen pregnancy. Data analysis used appropriate weights to ensure the representation of the sample at the regional and national levels. STATA 16 was used for analysis.