Menu
Menu
Menu
Menu
Adequate infant and young child feeding (IYCF) improve child survival and growth. Globally, about 18 million babies are born to mothers aged 18 years or less and have a higher likelihood of adverse birth outcomes in India due to insufficient knowledge of child growth. This paper examined factors associated with IYCF practices among adolescent Indian mothers. This cross-sectional study extracted data on 5148 children aged 0–23 months from the 2015–2016 India National Family Health Survey. Survey logistic regression was used to assess factors associated with IYCF among adolescent mothers. Prevalence of exclusive breastfeeding, early initiation of breastfeeding, timely introduction of complementary feeding, minimum dietary diversity, minimum meal frequency, and minimum acceptable diet rates were: 58.7%, 43.8%, 43.3%, 16.6%, 27.4% and 6.8%, respectively. Maternal education, mode of delivery, frequency of antenatal care (ANC) clinic visits, geographical region, child’s age, and household wealth were the main factors associated with breastfeeding practices while maternal education, maternal marital status, child’s age, frequency of ANC clinic visits, geographical region, and household wealth were factors associated with complementary feeding practices. IYCF practices among adolescent mothers are suboptimal except for breastfeeding. Health and nutritional support interventions should address the factors for these indicators among adolescent mothers in India.
This study utilized data extracted from the 2015–2016 India National Family Health Survey (NFHS-4); also referred to as the 2015–16 India Demographic and Health Survey (DHS), conducted by the International Institute for Population Sciences, Mumbai, India. Details of the methodology and sampling procedure of the survey can be found elsewhere [23]. Sociodemographic, household characteristics, and data on infant and young child feeding practices were collected from a sample of respondents (adolescent mothers aged between 15 and 19 years). A multistage cluster sampling design was used for the survey (which adopted a standardized questionnaire), from NFHS-4. The study was limited to children who were alive, of singleton births, last-born, aged 0–23 months and lived with the respondent. The survey yielded a weighted total of 5148 children with an average response rate of 94%. The study outcomes were the 2008 IYCF (BF and CF) indicators prescribed by the World Health Organization (WHO) [24]. The study was based on the recall of mothers regarding the food they fed their child within the 24 h preceding the survey. In this study, we considered four key BF indicators because exclusive breastfeeding and early initiation of breastfeeding are protective factors for child mortality and morbidity, while predominant breastfeeding and bottle feeding increases the risk of diarrhea and respiratory illness [25]. The selected BF indicators for our study are defined below: The CF indicators considered in the study are defined below: The independent variables were composed of socio-demographic and economic characteristics of children and their parents. The choice to use these variables was informed by previous literature [26,27,28,29] and their availability in the India 2015–2016 NFHS dataset. They were classified into three levels: individual-, household- and community-level factors. Individual-level factors consisted of characteristics of the mother (religion, age, work status, education, literacy, body mass index (BMI), age, marital status, place and mode of delivery, delivery assistance, number of antenatal visits, postnatal checks, access to the media, and power over earning and decision making), the father (occupation) and the child (sex, age, birth weight, birth order, birth interval, illness, and perceived size at birth). The wealth index, number of living children, quality of the source of drinking water, and quality of the type of toilet facility constituted the household-level factors, while the community-level factors were composed of the type of residence, geographic region, and type of caste or tribe. In the NFHS, the principal components analysis [30] was used to construct the household wealth index. The latter was calculated as a score of ownership of household assets, such as transportation device, ownership of durable goods, and household facilities. Furthermore, the index was divided into three categories, namely, poor, middle, and rich (detailed information on the definition and categorization of potential confounding variables used in the study is provided in Supplementary Table S1). The strategy for the analyses in this study was in line with that of previously published research [31,32,33]. Preliminary analyses involved the assessment of frequencies and cross-tabulations to estimate the prevalence of all the IYCF indicators used in the study. An estimation of the prevalence and corresponding confidence intervals of IYCF indicators then followed. IYCF indicators used in the study were categorized as binary (yes as ‘1′ and no as ‘0′) and we then conducted univariable and multivariable survey logistic regression analyses to examine the association between the study variables (individual-, household- and community-level factors) adjusted for clustering and sampling weights. Survey logistic regressions that adjusted for cluster and survey weights were used to identify unadjusted odds ratios of all the study outcomes. A three-staged modelling technique was adopted for the survey multivariable analyses in which level factors were entered progressively into the model to assess associations with the study outcomes. In the first stage, individual-level factors were entered into the baseline multivariable model to examine their association with the study outcomes. Thereafter, a manually executed elimination method was used to determine factors associated with IYCF at a 0.05 significance level (Model 1). In the second stage, household-level factors were added to Model 1, and those factors with p-values < 0.05 were retained (Model 2) after a manually executed elimination method was conducted. In the third stage, community-level factors were added to Model 2. As before, those factors with p-values < 0.05 were retained (Model 3). Only those factors significantly associated with IYCF at a 5% significance level in Model 3 were reported in the study. In the final model, we tested and reported any co-linearity. We then calculated the odds ratios with 95% confidence intervals derived from the adjusted logistic regression models, which were used to determine the level of association of the factors of possible confounding variables, and all analyses were conducted using Stata version 14.0 (Stata Corp, College Station, TX, USA).
