The burden of unintended pregnancies among Indian adolescent girls in Bihar and Uttar Pradesh: findings from the UDAYA survey (2015–16 & 2018–19)

The data used in the present study is taken from the Population Council’s longitudinal survey called 'Understanding the Lives of adolescents and Young Adults (UDAYA) conducted in 2015–16 and 2018–19. UDAYA is a one-of-a-kind cross-sectional and longitudinal state representative survey that gathered information on various socioeconomic, demographic, and personal factors and shed light on factors determining transitions to adulthood [15]. UDAYA is based on both cross-sectional and longitudinal designs. The cross-sectional component entails interviewing independent samples of unmarried boys in ages 10–14 and 15–19, unmarried girls in ages 10–14 and 15–19 and married girls in ages 15–19 in both rural and urban settings in Bihar and Uttar Pradesh at two points in time. The first round of cross-sectional surveys was conducted in 2015–16 and the second round was conducted in 2018–19. The longitudinal component comprises: (1) re-interviewing in 2015–16 the sample of unmarried girls and boys and married girls in ages 15–19 who were first interviewed in 2007 as part of the Youth Study and were at the time of the re-interview in ages 23–27 in Bihar; and (2) re-interviewing in 2018–19 the sample of adolescents who were interviewed for the first time in 2015–16 in Bihar and Uttar Pradesh. A total of 10,433 adolescents in Bihar and 10,141 adolescents in Uttar Pradesh were interviewed in 2015–16, and a follow-up interview in 2018–19 was executed [16].

Unintended pregnancy among adolescent girls and adult females was the outcome variable of the study. The question on the intendedness of the pregnancy was asked to only currently married adolescents who were currently pregnant at the time of the survey. They were asked the question, “You are currently pregnant; did you want this pregnancy now, later or not at all?

The response included three options (a) pregnancy wanted at the time of conception (wanted), (b) pregnancy wanted later (mistimed) and (c) pregnancy not wanted at all (unwanted). These three categories were refined and divided into two categories: intended (wanted) and unintended pregnancies (mistimed and unwanted). The outcome variable was coded as 0 = “Intended pregnancy” and 1 = “Unintended pregnancy”.

Currently pregnant adolescents and young adults of age 16–23 were included in the study. The study’s final sample included 659 (286 at Wave 1 and 373 at Wave 2) adolescents and young adults in Uttar Pradesh and 1321 (586 at Wave 1 and 735 at Wave 2) adolescents and young adults in Bihar who were currently pregnant at the time of the survey. The study is based on a secondary dataset; hence, no ethical approval from any institutional board was required. The Institutional Review Board approved the study and its data collection by the Population Council. It also ensured the confidentiality of the participants and informed consent was sought from the respondents during the survey.

In order to determine the factors affecting unintended pregnancies among adolescents in Bihar and Uttar Pradesh, socioeconomic, demographic, health and contraception-related variables were taken into the analysis. The socioeconomic and demographic factors included caste (categorized as SC/ST & Others), wealth index (categorized as Rich, Middle & Poor), place of residence (categorized as Urban & Rural), religion (categorized as Hindu & Non-Hindu), education of the respondent (categorized as illiterate &, literate), age of the respondent (categorized as 16–19 & 20–23), age of the spouse (categorized as <  = 18, 19 to 25, 26 and more & don’t know), husband’s year of schooling (categorized as no schooling, 1–7 years, 8–9 years & 10 years and above), media use (categorized as No & Yes) and internet use (categorized as No & Yes). The demographic and contraception-related factors include ever given birth to a live child (categorized as no, yes & first time pregnant), number of wanted children (categorized as, less than 2 & more than 2), knowledge and use of contraception (categorized as no & yes), side effects of contraception (categorized as no & yes) allowed to go to a health facility (categorized as No & Yes), can obtain information about contraceptives from ASHA/ANM (categorized as confidant & not confidant), knowledge about termination of pregnancy (categorized as No & Yes), discussion about number of children (categorized as No & Yes), heard about SATHIYA (social workers focusing on improving the lives of girls across local communities by providing them education on their health related issues) (categorized as No & Yes), discussion about the number of children (categorized as no & yes) and heard about Adolescent Friendly Health Clinic (categorized as No & Yes).

The study starts with the univariate analysis (sample distribution) for all the variables taken in the analysis for Wave 1 and Wave 2 in Uttar Pradesh and Bihar. Afterwards, a bivariate analysis was conducted to examine the association between the dependent and the independent variables. A Chi-square test tested the degree of association for the same. Finally, a binary logistic regression model was used to analyze the adjusted effect of the predictors on unintended pregnancy among adolescent and young adult females represented as, where \({X}_{i}\) are predictor variables and \({b}_{1}, {b}_{2}, {b}_{3}\dots \dots {b}_{i}\) represents the coefficient of each predictor variable included in the model. The logistic regression was interpreted in terms of odds ratio, which showed how likely or unlikely the outcome was to be present among those with X = 0 and those with X = 1. The statistical software used for the entire process of data analysis was STATA (Version 15).

Table 1 presents the description of the profile of the respondents in Uttar Pradesh at Wave 1 and Wave 2. At both Wave 1 and Wave 2, most females belonged to the age group 16–19 (99.7% & 95.9%). Caste-wise, most of the respondents belonged to Other caste, followed by the Scheduled caste at both Wave 1 and Wave 2. Most adolescents belonged to the Hindu religion at both periods, while the number of literate adolescents decreased from Wave 1 to Wave 2. More than half of the adolescents at Wave 1 belonged to the poor wealth index, whereas 51.9 per cent of the adolescents belonged to the rich wealth index at Wave 2. Most respondents resided in the rural areas at Wave 1 and Wave 2. At Wave 1, 79.2 per cent of adolescents were exposed to media, which increased to 89.2 per cent at the time of Wave 2. Internet use among adolescents drastically increased from 0.9 per cent at Wave 1 to 24.8 per cent at Wave 2. The majority of the adolescents were first-time pregnant at Wave 1. At the same time, 60.4 per cent had ever given birth to a live child at the time of Wave 2. 64.4 per cent and 67.7 per cent of adolescents preferred less than 2 children at Wave 1 and Wave 2, respectively. Most adolescents were not allowed to go to a health facility at both Wave 1 and Wave 2. Most adolescents have heard about contraception at both periods, while 98.4% did not hear about SATHIYA at Wave 2. Almost 91 per cent of the females had ever used any method of contraception at Wave 1, which decreased to 36.3 at Wave 2. Most adolescents felt confident about obtaining information about contraceptives from ASHA/ANMs at Wave 1 and Wave 2. More than half (51.5%) of the adolescents at Wave 1 thought contraceptive methods would cause side effects, while 53.3 per cent at Wave 2 said no to this notion. Most respondents at both periods believed that a woman could not terminate the pregnancy. At Wave 1, almost 57 per cent of the respondents discussed the number of children with their husbands, which decreased to only 17.4 per cent at Wave 2. Lastly, most adolescents did not hear about the friendly health clinics at Wave 1 (94.2%) and Wave 2 (95.6%).

Table 2 presents the description of the profile of the respondents in Bihar at Wave 1 and Wave 2. At Wave 1, the majority of the respondents (46.6%) were 20–23-year-old while at Wave 2 (84.8%) majority were 16–19-year-old. Most of the respondents belonged to Other caste and Hindu religion and were literate at both Wave 1 and Wave 2. At Wave 1, most respondents hailed from the rich wealth quantile (44.4%), whereas the majority belonged to the poor wealth quantile (48.1%) at Wave 2. Most of the respondents were rural residents and were exposed to media during both periods. At Wave 1, 39.8 per cent of the adolescents were first-time pregnant, while 62.2 per cent had ever given birth to a live child at the time of Wave 2. More than 60 per cent of adolescents preferred less than 2 children at both Wave 1 and Wave 2.

Many respondents were not allowed to attend a health facility at either period. Most adolescents have heard about contraception at both periods. In contrast, only 1.2% had heard about SATHIYA at Wave 2. 92 per cent of the respondents had ever used any method of contraception at Wave 1, while it was only 18.6 per cent at Wave 2. Most of the adolescents felt confident about obtaining information about contraceptives from ASHA/ANMs at both periods. More than 60 per cent of the adolescents at both periods responded that contraceptive methods would cause side effects. Most respondents at both periods believed that a woman could not terminate the pregnancy. At Wave 1, almost 55 per cent of the respondents discussed the number of children with their husbands, which decreased to only 24 per cent at Wave 2. Most respondents (74.5%) reported their spouse’s age as 19–25, whereas the majority did not know about their spouse’s age at Wave 2 (67.6%). Lastly, more than 93 per cent of the adolescents did not hear about the adolescent-friendly health clinics at both waves.

Figure 1 shows that 40.1 per cent of all currently pregnant adolescents and young adults reported their pregnancy as unintended (mistimed and unwanted) in Uttar Pradesh at Wave 1 of the survey, which decreased to 34.2 per cent at Wave 2. On the contrary, almost 99 per cent of all currently pregnant adolescents and young adults in Bihar reported their pregnancy as unintended at Wave 1, which decreased to 44.8 per cent at Wave 2.

Table 3 presents the bivariate analysis in which age and unintended pregnancy were positively associated among adolescents, wherein the prevalence of unintended pregnancy was highest among the 20–23 age group in Uttar Pradesh at Wave 2. Religion-wise, the highest prevalence of unintended pregnancy was observed among Non-Hindu adolescents at both Wave 1 (48.0%) and Wave 2 (38.2%). It was also observed that the prevalence of unintended pregnancy was higher among illiterate adolescents at Wave 1 (78.5%). Wealth and unintended pregnancies were positively related as an increase in unintended pregnancies was observed with an increase in the wealth quantile among the adolescents of Uttar Pradesh at Wave 1 (44.8%). At Wave 2, a significant rural–urban difference was observed as adolescents from urban areas showed a significant preponderance of unintended pregnancy (38.6%). Exposure to media and internet use was significantly associated with a lower prevalence of unintended pregnancy among adolescents at both periods. At Wave 1, a higher prevalence of unintended pregnancy was found among first-time pregnant adolescents (41%). At both Wave 1 and 2, unintended pregnancy was higher among adolescents who preferred less than two children. Adolescents who were not allowed to go to a health facility reported higher unintended pregnancies at both periods. At Wave 2, adolescents who had not heard about contraception reported higher unintended pregnancy (34.3%). Subsequently, the prevalence of unintended pregnancy was also higher among the respondents who did not use any method of contraception (55%) at Wave 1. Unintended pregnancy was observed high among females who thought contraceptive methods would have side effects (36.6%) at Wave 2. The prevalence of unintended pregnancy was higher at both waves among women who thought a pregnant woman could not terminate a pregnancy. At both Wave 1 and Wave 2, a higher prevalence of unintended pregnancy was found when the husband and wife did not discuss about the number of children (48.7% & 35.1%). A positive relationship was found between the husband’s education and age with unintended pregnancy at Wave 2. Lastly, a higher prevalence of unintended pregnancy was observed among adolescents who had not heard about adolescent-friendly health clinics at both periods (41.1% & 35).

Table 4 presents the bivariate analysis results starting with age significantly associated with unintended pregnancy in Bihar at the time of both Wave 1 and Wave 2. A higher prevalence of unintended pregnancy was found among non-Hindu adolescents at Wave 1 (53.5%) and among Hindu adolescents at Wave 2 (46%). Adolescents who were literate and those who resided in urban areas showed a higher preponderance of unintended pregnancy at Wave 1 and Wave 2, respectively. Wealth index was positively associated with unintended pregnancy as unintended pregnancy showed a decrement among adolescents with a decrease in wealth quantiles. Media exposure and internet use were significantly associated with a lower prevalence of unintended pregnancy among adolescents at both periods (46.1% & 43.5%). At both Wave 1 and 2 a higher prevalence of unintended pregnancy was found among adolescents who had given birth to a live child. Adolescents who preferred less than two children showed a higher preponderance of unintended pregnancy at both waves. A higher prevalence of unintended pregnancy was observed among adolescents who were not allowed to go to a health facility at both periods (48.6% & 45.4%). A higher preponderance of unintended pregnancy was found among adolescents who have not heard about contraception at Wave 2 (45.5%). Moreover, the prevalence of unintended pregnancy was also higher among the respondents who did not use any method of contraception at both periods. At Wave 1, adolescents confident in obtaining information about contraceptives from ASHA/ANM showed less preponderance of unintended pregnancy (45.8%). The prevalence of unintended pregnancy was higher among those who thought contraceptive methods would have side effects (36.6%) at both Wave 1 and Wave 2. Similarly, the prevalence of unintended pregnancy was higher at both waves among women who thought pregnant women could not terminate a pregnancy. Unintended pregnancy was higher among respondents who did not discuss the number of children with their spouses during both periods. Lastly, a higher prevalence of unintended pregnancy was observed among adolescents who had not heard about adolescent-friendly health clinics during both periods (47% & 45).

Table 5 presents the results of binary logistic regression analyses showing the relation of pregnancy intention of adolescents and young adults' most recent birth by controlling for socioeconomic and demographic variables. The results indicated that factors like husband’s schooling, age of spouse, allowed to go to a health facility, and termination of pregnancy were not significantly associated with the pregnancy intention status of the most recent birth (mistimed or unwanted) at both Wave 1 and 2, when all covariates were adjusted.

Older adolescents were 31 per cent more likely to have higher unintended pregnancies at Wave 1, whereas adolescents aged 20–23 were 36 per cent less likely to have unintended pregnancies than their counterparts at Wave 2. Adolescents of Other caste were more likely to have unintended pregnancy at Wave 1 while they were 15 per cent significantly less likely [OR = 0.95, 95% CI = 0.73–1.24] to have an unintended pregnancy compared to SC\ST caste at Wave 2. Religion-wise, adolescents from non-Hindu religion were 58 per cent and 27 more likely to have unintended pregnancy at both periods. Literate adolescents were 36 per cent more likely than the illiterate ones to have reported their current pregnancy as unintended at Wave 1 [OR = 1.36, 95% CI = 0.48–3.85]. On the other hand, the literate adolescents were 58 per cent less likely than the illiterate ones to have reported their current pregnancy as unintended at Wave 2. Wealth and unintended pregnancy showed a significant negative relationship at both periods. Adolescents from poor wealth quantiles were significantly more likely to have unintended pregnancies than their affluent counterparts at Waves 1 and 2. Adolescents from rural backgrounds were 19 per cent more likely to have an unintended pregnancy than their urban counterparts at Wave 2. In case of media exposure, adolescents exposed to media were 16 per cent and 26 per cent less likely to report their pregnancy as unintended than their counterparts at Wave 1 and 2, respectively. Comparatively, those who used the internet were 3 per cent significantly less likely to report an unintended pregnancy at Wave 2. Adolescents who had ever given birth to a live child were significantly less likely to have unintended pregnancies than those who had not given birth at both periods. Adolescents who wanted more than two children were 5 per cent significantly more likely to have an unintended pregnancy than their counterparts [OR = 1.05, 95% CI = 0.82–1.35]. Adolescents who heard about contraception and SATHIYA were 11 per cent and 55 per cent significantly less likely to have unintended pregnancies at Wave 2, respectively. Adolescents who were not confident about getting contraceptive information from ASHA/ANM and those who thought contraceptive methods would have caused side effects were significantly more likely to have unintended pregnancy at Wave 2. Adolescents who discussed the number of children with their spouses were 50 per cent less likely and 33 per cent more likely to have unintended pregnancy at Wave 1 and 2, respectively. Lastly, adolescents who heard about adolescent-friendly health clinics were 26 per cent less likely and 36 per cent more likely to have unintended pregnancies at Wave 1 and 2, respectively.

The strength of the study lies in the fact that it uniquely focuses on the drivers of unintended pregnancy among the young population in those two states which lag behind on the indices of reproductive health and family planning matters. Another strength could be attributed to the fact that it derives data from a survey which is dedicated completely towards understanding the health concerns of this young and vulnerable population in Bihar and Uttar Pradesh. Despite these strengths, there are a few limitations as well. Findings should not be extrapolated to the entire population of the country because the data is not representative of all of India and is restricted to just two socioeconomically backward states. Additionally, because our study is quantitative in nature, we cannot capture the unique social and cultural perspectives on the intendedness of pregnancy among young people that might be acquired from a qualitative investigation.