Using a gender lens to understand eating behaviours of adolescent females living in low‐income households in Bangladesh

Abstract Adolescence is a critical period characterized by rapid physical, psychological, and social development and growth. In Bangladesh, high rates of undernutrition persist among adolescent females living in low‐income households. Prevalence of adolescent marriage and pregnancy is extremely high, with almost half of Bangladeshi women giving birth by 18 years of age. Qualitative research was carried out from April to June 2017 to examine individual, social, and environmental factors influencing eating behaviours of female adolescents between 15 and 19 years of age living in low‐income families in urban and rural settings in Bangladesh. Methods included freelisting exercises (33), key informant interviews (11), in‐depth interviews (24), direct observations (16), and focus group discussions (12). Findings show that household food insecurity necessitates adjustments in meal food quality and frequency. Gender norms prescribe that females receive small meal portions and make sacrifices in food consumption so that male family members can eat more. Work and school schedules cause long breaks between meal consumption, restricting food intake of adolescent females for extended periods. Gender discrimination and its manifestations likely amplify susceptibility to psychological stresses in adolescent females. An inferior social position makes adolescent females living in food insecure households vulnerable to undernutrition, with factors affecting food deprivation increasing as they approach childbearing. Policies to increase age of marriage and reduce adolescent pregnancy must continue. Programmes must ensure that school‐going adolescents eat adequately during the school day. Prolonging school education and strengthening the economic viability of women should alter cultural expectations regarding marriage age and normative female roles.


| INTRODUCTION
Adolescence is a transitional period from childhood to adulthood characterized by rapid physical, psychological, and social development and growth (Das et al., 2017). Early physiological changes involve puberty and sexual development and, in later phases, pubertal maturation (Das et al., 2017). Adolescents experience significant physical growth accounting for about 50% of adult weight and 15% of adult height (Spear, 2002). These dramatic physical changes create increased demands for nutrients and energy (Das et al., 2017). Simultaneously, adolescents undergo psychosocial changes, including growing concerns about physical appearance, greater autonomy, and modifications in roles and responsibilities; all of which can impact on food intake and nutritional status (Das et al., 2017;Story, Neumark-Sztainer, & French, 2002). Research shows that adolescents frequently have poor eating habits, increasing risks for iron deficiency anaemia, micronutrient deficiencies, suboptimal linear growth, and undernutrition, as well as overweight in many contexts (Akseer, Al-Gashm, Mehta, Mokdad, & Bhutta, 2017;Christian & Smith, 2018). Poor adolescent eating behaviours and undernutrition are predictive of adult eating practices and health (Larson, Neumark-Sztainer, Hannan, & Story, 2007;Pedersen, Holstein, Flachs, & Rasmussen, 2013). Determinants of nutritional problems vary according to the sociocultural and economic context and physical environment where adolescents live.
Despite rising incomes and access to a variety of foods, widespread malnutrition persists in low-and middle-income countries, particularly in urban slums and rural areas (Abarca-Gómez et al., 2017;Pörtner & Su, 2018). In these settings, girls are often more likely than boys to drop out of school at a young age, increasing the likelihood of social isolation, early marriage, and childbearing when girls are still maturing (Gibbs, Wendt, Peters, & Hogue, 2012;WHO, 2014a). Adolescent mothers are at greater risk for severe delivery complications, with nutritional status prior to conception critical for maternal and infant outcomes Han, Mulla, Beyene, Liao, & McDonald, 2011). An early start to childbearing has been shown to trigger negative health and social consequences, perpetuating the intergenerational cycle of malnutrition (Branca, Piwoz, Schultink, & Sullivan, 2015;Temmerman, Khosla, Bhutta, & Bustreo, 2015;WHO, 2011WHO, , 2014b. For decades, undernutrition has been a major public health problem in Bangladesh, with the prevalence of malnutrition historically reported to be higher among females than their male counterparts (Chen, Hug, & D'Souza, 1981;Choudhury, Hanifi, Rasheed, & Bhuiya, 2000). Recent data show that 31% of ever-married females aged 15-19 are underweight (BMI < 18.5), and 26% of adolescent girls 10-  (Leroy, Ruel, Sununtnasuk, & Ahmed, 2018).
Energy intake is also inadequate (Leroy et al., 2018). Early marriage and childbirth are high with about 60% of women age 20-24 marrying and close to half of all Bangladeshi women giving birth by 18 years of age and before reaching physical maturity (National Institute of Population Research and Training [NIPORT] et al., 2016).
There has been a dearth of research investigating adolescent eating behaviours and their determinants in Bangladesh. The few studies conducted have shown that limited dietary diversity, suboptimal energy and micronutrient intake, and undernutrition are higher in the lowest wealth quintiles, with household wealth associated with the probability of better nutritional status (Akhter & Sondhya, 2013;Alam, Roy, Ahmed, & Ahmed, 2010;Leroy et al., 2018;National Institute of Population Research and Training [NIPORT] et al., 2016). Data from rural areas indicate that education and empowerment of female household heads are insufficient to improve nutritional status of adolescents, suggesting that resources may be too constrained (Leroy et al., 2018). To our knowledge, potential contributions of more distal social and cultural factors contributing to adolescent food intake in Bangladesh have not been explored.
Gender has proven to be a key criterion when assessing decisions related to intrahousehold food allocation (Haddad, 1999). Research conducted in Nepal found that women were systematically not allowed to consume more expensive, prestigious foods (Gittelsohn, Thapa, & Landman, 1997). A study carried out in India showed that discrepancies in the value attributed to males and females caused discrimination in the allocation of foods for girls (Das Gupta, 1987). In Ethiopia, study results demonstrated that preferential treatment of adolescent boys led to gender-based differentials in food insecurity experienced by girls (Hadley, Lindstrom, Tessema, & Belachew, 2008).

Key messages
• Gender norms interact with economic, social, and structural factors to make female adolescents vulnerable to poor dietary intake and undernutrition.
• Female socialization involves making sacrifices related to food consumption so that male family members can eat more.
• Work of family wage earners and school schedules foster long breaks between meals, causing adolescent females living in low-income households to skip meals • Gender discrimination and its manifestations appear to amplify psychological stresses in adolescent females living in food insecure households.
• Prolonging school education and strengthening the economic viability of women should alter cultural expectations regarding normative female roles in households and society.
Using gender as an overarching framework, qualitative research was conducted to examine individual, social, environmental, and structural factors influencing eating behaviours of female adolescents living in low-income families in urban and rural settings in Bangladesh.

| Study setting and population
Data collection was carried out between April and June 2017 in urban slums in the cities of Dhaka and Chittagong and rural areas in the districts of Barisal and Kurigram (see Figure 1). Selection of research sites was designed to include low-income urban and rural settings that differed geographically, topographically, and culturally. Dhaka and Chittagong, the most populated cities in Bangladesh, are infiltrated with slums inhabited by immigrants in search of cash employment.
Barisal is a southern fishing and agricultural district located on flood plains, and Kurigram, a northern farming district, is one of the poorest and most food insecure regions of Bangladesh.

| Study design, sampling, and methods
Two study areas in each research site were included, with slums in Dhaka and Chittagong selected randomly. In Kurigram and Barisal, one village close and one far from the main district town were selected purposively. Adolescent respondents ranged from 15 to 19 years of age and were divided into three categories typical of this age group, including students, girls who had discontinued school, and married adolescents. We only sampled unmarried girls residing with their parents. The research used a mix of qualitative methods administered across the four research sites as described below.
Free-listing exercises were employed to assess the cultural domain of meals. School-going girls were identified on the road while going to school or in their homes, whereas non-school-going and married adolescents were approached near or in their households. The initial target was to conduct 30 exercises.
Key informant interviews were conducted with professionals working directly or indirectly on issues related to adolescence at the national level or in research communities to examine determinants of eating behaviours. Key informants were selected purposively based on their expertise and knowledge of adolescent health, nutrition, and social behaviour, with the goal to continue interviewing until reaching data saturation.
In-depth interviews were administered with school-going, nonschool-going, and married adolescents. The aim was to explore individual motivations for food consumption, patterns of food intake, environmental and social factors guiding eating behaviours, and the social context of eating. We also examined perceptions of health and work responsibilities. School-going girls were identified in or nearby schools, whereas the non-school-going and married respondents were approached in households with the target to interview eight females representing each of the different adolescent categories. Prior to indepth interviews, day-long direct observations of subsamples of indepth interview respondents were conducted. Observations started in the morning and followed respondents' activities throughout the day with the objective to assess meal and snack food consumption, as well as household cooking patterns and food distribution. We also observed work and school schedules and social interactions. Observation data were referred to when conducting in-depth interviews with the same adolescent respondents.
Group discussions were administered with fathers and mothers of adolescent females ages 15-19 years, as well as adolescent males 15-19 years of age. During discussions, we explored determinants of food acquisition and consumption, perceptions of intrahousehold food distribution, and household decision-making and responsibilities.
Participants were selected with assistance from community key informants and opinion leaders, with the aim to include six to 10 participants in each group.

| Data collection procedures
Data were collected by five research assistants, two supervisors, and an anthropologist responsible for field activities. Prior to data collection, a 5-day training on data collection methods and ethical procedures was held in Dhaka. Study instruments were tested in a slum and modified based on the field experience.
Teams of reseach assistants travelled to the study sites sequentially, starting in Dhaka and subsequently going to Barisal, Kurigram, and Chittagong. In each site, initial data collection involved freelisting exercises and key informant interviews. Key informant interviews were open-ended, designed to last approximately 1-hr, and typically FIGURE 1 Map of Bangladesh indicating the districts where the study was conducted conducted in the informant's work place. A guide was used, with ensuing questioning incorporating emerging study findings. Subsequently, adolescents were selected for observations and in-depth interviews.
Research assistants employed a form to record observational data. A semi-structured guide was followed to conduct in-depth interviews, with interviews limited to 1.5 hrs. When needed, researchers returned to households for additional information. Group discussions lasting no longer than 2 hrs were held at the end of data collection in each site. A guide based on study themes and preliminary results of other research methods was used.
All interviews were carried out in Bengali, the Bangladesh national language. Key informant and in-depth interviews and group discussions were audio recorded. Efforts were made to maintain privacy when administering all data procedures.
Research investigators were in regular communication during data collection to discuss study results and modifications of instruments. A debriefing meeting of investigators was held after completion of data collection.

| Data analysis
Freelisting exercises were analysed on Anthropac 4.983 software (Borgatti, 1996). Salience was derived using a salience index (Smith's S) defined as S = ((L − Rj + 1)/L)/N, where L is the length of each list, Rj is the rank of item J in the list, and N is the number of lists in the sample (Borgatti, 1999). Items with a higher salience score were considered to have greater significance (Bernard, 1988;Weller, 1984).
The audio recorded interviews and observational data were translated into English from Bengali and transcribed. Transcripts were reviewed by senior researchers during the course of the study. After data collection, a coding system was developed for key informant, in-depth interviews, observations, and group discussions based on reviews of transcripts. Coding categories were derived from initial research themes, as well as emerging concepts. Coding of interview transcripts was done on ATLAS.ti, a text-organizing software (ATLAS.ti Scientific Software Development, 2013), whereas observational and group discussion data were coded in matrixes. Content analysis was used to identify trends of concepts in and across individual codes. The combination of data, environmental and methodological triangulation, facilitated data analysis across the different research sites and methods and across and between key informants and adolescent respondents.

| Research ethics
The study protocol was reviewed and ethical approval granted by the Institutional Review Board Bangladesh Medical Research Council, Dhaka, Bangladesh. Verbal informed consent and assent from guardians of respondents under 18 years of age were obtained from all freelisting and in-depth interview respondents, key informants, and focus group participants prior to data collection.

| Background information
We administered 33 freelisting exercises across the research sites.
Seven key informants composed of community-based health workers, school teachers, and opinion leaders were interviewed in study communities. National level informants included two NGO representatives working on adolescent and maternal health and nutrition, a researcher with gender expertise, and a social marketer. Eight adolescents from each respondent category participated in in-depth interviews, with five or six of each observed. Observations lasted from about 7 a.m.
to 5 p.m. Due to security and logistic reasons, research assistants were unable to observe evening meals, which often occurred late at night.
In each site, we administered discussions with separate groups of mothers and fathers of adolescent girls and adolescent boys. Fathers were involved in income-generating activities, whereas mothers were working women and housewives. Adolescent boys were school-going, working, and nonworking. Table 1 presents the number of respondents according to each data collection method.
The average age of adolescents in respondent categories was between 16 and 18 years ( Table 2). All respondents were Muslim.
School-going and non-school-going adolescents had more years of education than those in the married group, with one married respondent never attending school. Mean months of marriage was 15.5; two married respondents had children, and one was pregnant.
Half of households where adolescent respondents lived had more than one earning member, particularly those in cities. Most urban households had women engaged in wage labour, whereas females residing in rural sites were not earning money. One school-going and one married respondent generated small wages through home-based manufacturing activities. Common occupations of male household heads were small business, day labour, driver, and service employee, such as garment worker or security guard. No urban-based families possessed land in their home districts. All families in rural areas owned homestead land and about half had small agricultural plots where they produced rice for consumption.

| Freelisting
Exercises identified 87 foods eaten during meals with rice, fish, green leaves, lentil, egg, vegetables, and potato the most salient. Typical of a cultural domain, procedures elicited a core set of more significant and a wide range of less important meal items, with the top 15 ranked foods listed in Table 3. Vegetables and types of fish were meal foods most frequently listed.

| Eating patterns of the female adolescents
At the time of the study, most adolescents reported eating three meals daily, although some from Kurigram claimed their families ate twice.
Prior to breakfast, some families offer a "light" meal consisting of biscuits, crispy rice (chaal bhaja) 1 or puffed rice (muri) for members rising early. Adolescents described family meals as dominated by rice with curry 2 dishes, with most families eating either fish or eggs daily.
Chicken is eaten once weekly or bi-monthly, often Friday, with beef consumed only on special occasions.
We observed that all respondents consumed breakfast, with schoolgoing girls eating earlier in the morning (Table 4). Breakfast most often consisted of leftover rice with curry, with a few adolescents also eating meat or fish. Some respondents substituted rice for roti (traditional flatbread made of whole meal flour) eaten with curry or lentils. One girl only ate rice with chilli. Food portions differed from those of male members who ate full plates of rice or several pieces of roti with curry, whereas the girls ate small amounts of rice or one or two pieces of roti with or without curry. Several school-going girls claimed to skip breakfast often, stating they did not like leftovers in the morning.
Lunch was consumed between 1:00 and 5:00 pm, with those eating a late breakfast consuming lunch later in the day. School-going adolescents ate lunch late in the afternoon after returning home from school; this was sometimes the first food eaten since breakfast. During observations, most respondents ate fish or animal source foods for lunch, whereas the others consumed vegetable curry and lentils with rice. One married respondent who attended a special festivity consumed different meats and fish. Although foods consumed varied little with other family members, we observed that portions of fish or meat were less than that given to males. An exception was the breastfeeding respondent whose food portions were observed to be comparable with those of her husband. We observed one married respondent refuse to eat chicken her husband intentionally placed on her plate.
Respondents reported that dinner is eaten between 8 pm and midnight. Many respondents claimed to have little appetite at dinner time, eating small amounts when the meal is comprised of leftovers or less desirable foods such as green leafy vegetables, missing dinner particularly when lunch is consumed late in the afternoon, or falling asleep before dinner is served.
Adolescents indicated that snacks generally entail leftover meals, packaged or street foods such as chanachur (a spicy dry mix) and biscuits, or homemade rice-based items. Snacks were described as "light" foods eaten in small amounts between meals to appease hunger or when meal foods are unavailable, the meal is delayed, or for breakfast.
Few respondents were observed eating snacks between meals. Those 1 Uncooked rice is fried in a hot pan, which makes the rice crispy.
2 Curry is usually a dry or gravy dish which includes one or more vegetable items cooked with spices and oil. Sometimes driedfish is mixed with the vegetable, but this is mostly for taste.
Curry can also be combined with a pulse.  snacks consumed were limited in portions or quality (Table 4). However, during interviews, over half of the respondents claimed to eat snacks regularly between meals, particularly leftovers or puffed rice, when they are hungry. A couple of married respondents reported occasionally storing packaged biscuits at home to address hunger or as a dinner substitute. Most school girls received pocket money occasionally from parents to purchase snack foods.
Most respondents reported frequently experiencing hunger between meals, with married respondents emphasizing hunger between lunch and dinner. Adolescents across all groups reported periodically eating meals consisting of rice with chilli.

| Gender constraints on food consumption
This section highlights ways in which gender norms define male and female roles, which in turn influence food acquisition, decisions surrounding intrahousehold food allocation, consumption patterns, and discrepancies related to actual food intake, placing adolescent females at risk for poor health and undernutrition.

| Economic hardship and food insecurity
Parents participating in focus groups stated that food purchasing is driven by food availability and especially cost. Most adolescent respondents and group discussion participants reported that staple foods such as rice, potato, onion, and lentil are purchased in bulk weekly or less frequently, whereas several families in Kurigram and newly married nuclear families relying on low-paying and irregular daily wages bought staples daily. Families in rural areas producing rice for family consumption reported periodic shortages. We were told that fresh vegetables are generally obtained on a daily basis, particularly green leaves, which can be purchased for little money, or in rural settings, gathered. Due to the high price, fruit is only bought on special occasions. Eggs are purchased multiple times weekly, with rural households also raising chicken for eggs. Fish is procured once or more weekly, but meat is rarely bought due to the cost. According to key informants, low-income families buy chicken organs, legs, and skin discarded by hotels and sold in markets near slum communities as a meat substitute. Focus group participants indicated that male household heads are responsible for obtaining staples, animal food sources,  During observations, we found married respondents to be demure and obedient towards their in-laws and husband. Several married respondents reported being victim to verbal and emotional abuse inflicted by their mothers-in-law and husbands, and in one instance, the mother-in-law periodically restricted food as punishment.
Due to their economic roles, bigger physiques, and perceptions that males are more active, men and adolescent males were reported to be given both larger quantities and better-quality foods, with boys also favoured because they are expected to provide money and care for their parents in old age. This non-school-going girl (IDI_6) said,

| Daily schedules, household responsibilities, and mobility
School girls followed a structured schedule attending school and tutoring from early in the morning. They returned home late in the afternoon when some reported that the better lunch portions like fish might no longer be available. School-going girls performed household tasks after school and on weekends. Unmarried, non-school-going girls started and ended the day doing chores, with breaks particularly in the late afternoon, when they watched television or chatted with friends living nearby. Married adolescents worked in the household from morning to evening or nighttime when dinner is served. Those living in extended households had more responsibilities due to numbers of family members. Much of their day involved responding to ongoing requests, particularly by the mother-in-law, father-in-law, and husband, and ensuring their comfort. Household chores across adolescent respondents included food preparation, serving meal foods, washing dishes, fetching water, sweeping, washing clothes, and providing childcare for younger siblings or, in the case of two married respondents, their own children. Observations confirmed that most married adolescents did not leave the household and had limited contact with other people. Many desired the comfort provided by natal family members and social interactions with friends. Married respondents did, however, occasionally visit their natal family.

| Gender expectations, nutrition and health
Most adolescents expressed concerns about their health. Unmarried respondents commonly complained of fatigue, weight fluctuations, dizziness, and social tension, claiming that health problems interfere with studying and going to school, as well as performing household chores. These respondents often attributed inadequate food intake to small meal portions, long gaps between meals, and poor appetite, particularly when given frequently served meal foods and during menstruation. Tensions related to looking attractive and subscribing to the modest behaviour expected of adolescent girls, as well as concerns about their future schooling, falling in love and maintaining family dignity, and marriage, were also perceived to reduce appetites. School-going girls worried about having to terminate school due to costs, exposure to sexual harassment while travelling to school, or an arranged marriage, with all strongly opposed to marrying young. Although unmarried girls no longer in school preferred postponing marriage, many predicted marrying soon and voiced concerns of being victim to violence inflicted by in-laws, causing ten- considered an economic burden until they leave the household (Chowdhury, 2004;Parsons et al., 2015). In preparation for marriage, Bangladeshi girls are socialized to adhere to cultural principles associated with a good woman which involves being modest, obedient, and respectful, characteristics attractive to potential suitors (Chowdhury, 2004;Schoen, 2015). Female socialization involves making sacrifices, which in this study entailed consuming smaller food portions so that male family members ascribed with higher value can eat more.
Respondents gained satisfaction by offering food to males, acknowledging family economic constraints and their less consequential status.
In the context of low-income households, the interaction of food insecurity and prevailing gender norms works synergistically to limit dietary intake. Biologically, adolescent males have increased energy requirements compared with females the same age, and in food secure settings, boys would be expected to eat larger food portions (Spear, 2002). However, in these food insecure environments, inequities in food distribution based on gender (not biological sex differences) result in females given inadequate food to meet their nutritional needs. Although favouring males in or expected to enter the paid work sector may seem from the families' perspective to be a logical coping strategy, our results suggest that gender discrimination in the allocation of limited household food resources makes adolescent females more vulnerable to hunger and malnourishment. Other research carried out in South Asia has suggested that economically active family members and those expected to contribute to the household welfare in the future receive preferential treatment in relation to food resources (Gittelsohn, 1991;Senauer, 1990). Low female status has been shown to impact on household decision-making involving resource allocation, thus limiting women's ability to influence important familial decisions related to health and nutrition and perpetuating the intergenerational cycle of undernutrition (Haddad, 1999). Many studies have demonstrated that maternal education is positively associated with nutritional status in children and adolescents (Madjdian, Azupogo, Osendarp, Bras, & Brouwer, 2018;Makoka & Masibo, 2015 (Amin & Huq, 2008). Whereas family prestige, social pressure, and government programmes encourage female schooling, the dowry may be higher for better educated females, who may also be considered older prospective brides (Field & Ambrus, 2008;Parsons et al., 2015;Schuler, Bates, Islam, & Islam, 2006;Streatfield, Kamal, Ahsan, & Nahar, 2015). Social tenets opposing premarital sex and maintaining family dignity and concerns about female safety also foster early marriage (Chowdhury, 2004;Rashid, 2006;Schuler et al., 2006;Streatfield et al., 2015).
Limited food intake was most apparent among newly married females who are considered strangers when they enter their in-law's household. Married respondents, who claimed to be subject to scrutiny and criticism, developed strategies such as skipping meals, eating alone, consuming small amounts, and sacrificing food portions for others, all aimed at meeting expectations regarding their family role and avoiding attention. Low incomes and limited assets in newly formed households further restrict food consumption of married adolescents. The combination of their young age, perceived inferior sex, and economic dependence make new brides highly vulnerable to discrimination and suppression, which in turn affects food intake, as elucidated by married adolescents in our sample. Another study conducted in Bangladesh reported limited decision-making power and economic assets affect access to food resources by newly married adolescent women (Nguyen et al., 2017). Social status changes with the arrival of a first baby, marking acceptance as a family member (Rashid, 2006). In our study, the only married woman regularly provided more and special food items was a mother who reported significant changes in her treatment after conception and childbirth. , which appeared to be frequent occurrences in our sample, underlining the commonality of these practices.
The combination of poor food intake and strenuous work schedules may explain why many respondents complained of ill health, with most asserting that their limited food consumption does not meet their physical requirements. A common complaint was lack of appetite with some describing meal foods as undesirable. Sociocultural norms and economic factors force adolescent females to assume adult roles, leading to a variety of psychological stresses including concerns about growing work responsibilities, their impending marriage, and for married respondents, household finances at a prematurely young age.
Married respondents claimed a deterioration in their health since marriage, mentioning a multitude of socioemotional and physical stresses, including unhappiness, weight loss, social isolation, and abuse in the form of food deprivation. Limitations on mobility and interpersonal interactions deprived adolescents of entertainment and social outlets, leading to social isolation. These restrictions are linked to purdah, rules designed to limit interactions between males and females and maintain gendered behaviour (Chowdhury, 2004;Schoen, 2015). Numerous studies show increased risk to common mental health conditions (Gebreyesus, Endris, Hanlon, & Lindtjørn, 2018;Hadley & Patil, 2006), weight loss and dietary modifications (Rose, 1999;van Liere, Ategbo, Den Hartog, & Hautvast, 1995), and poor health in food insecure households (Casey et al., 2005;Cook et al., 2004;Cook et al., 2013;Siefert, Heflin, Corcoran, & Williams, 2004). A study in the United States found that food insecurity among adolescents was associated with chronic depression and contemplation of suicide (Alaimo, Olson, & Frongillo, 2002). In Bangladesh, gender discrimination and its manifestations likely amplify susceptibility to psychological conditions among adolescent girls living in food insufficient households.
In conclusion, findings illuminate sociocultural and economic factors that constrain food intake in adolescent females at a time when nutritional requirements are elevated. On a continuum, newly married adolescents, who have more demanding workloads and are expected to bear children, were the most extreme example, with married respondents consistently eating irregular, small food portions. In order to improve the nutritional status of adolescent girls, both gender norms and food insecurity influencing dietary intake must be recognized and addressed. In Bangladesh, the prevalence of adolescent marriage and pregnancy is among the highest in the world, signifying the strength of social and economic factors that guide early marriage and roles relegated to women (Loaiza & Liang, 2013;UNICEF, 2017).
Policies and programmes must continue to be implemented to increase marriage age and reduce adolescent pregnancy. In addition, the government and its partners need to develop programmes that ensure that school-going children and adolescents eat adequate nutritious foods during the school day. Such efforts should ensure that nutritious foods are available in and around school settings and include educational outreach to parents and teachers. Research examining psychological health and food intake among adolescents merits exploration. Over the long-term, increasing job opportunities, prolonging school education, and strengthening the economic viability of young women should influence the value ascribed to girls and women and thus alter cultural expectations regarding age at marriage and initiation of childbearing and normative female roles in households and society. Increased female empowerment should impact on household decision-making, including decisions related to food intake of adolescent girls.