INVESTING IN ADOLESCENT GIRLS' NUTRITION IN BANGLADESH: Situation Analysis of Trends and Ways Forward DISCUSSION PAPER July 2019 Malay Kanti Mridha Mokbul Hossain Tanvir Hassan Ipsita Sutradhar Samiun Nazrin Bente Kamal Akib Khan Nizam Uddin Ahmed Rudaba Khondker Piyali Mustaphi Ireen A Chowdhury Alayne M Adams Ziauddin Hyder INVESTING IN ADOLESCENT GIRLS’ NUTRITION IN BANGLADESH: Situation Analysis of Trends and Ways Forward Malay Kanti Mridha, Mokbul Hossain, Tanvir Hassan, Ipsita Sutradhar, Samiun Nazrin Bente Kamal, Akib Khan, Nizam Uddin Ahmed, Rudaba Khondker, Piyali Mustaphi, Ireen A Chowdhury, Alayne M. Adams, and Ziauddin Hyder July 2019 1 Health, Nutrition, and Population (HNP) Discussion Paper This series is produced by the Health, Nutrition, and Population Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. For information regarding the HNP Discussion Paper Series, please contact the Editor, Martin Lutalo at mlutalo@worldbank.org or Erika Yanick at eyanick@worldbank.org. Rights and Permissions The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street, NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. © 2019 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW, Washington, DC 20433 All rights reserved. ii Health, Nutrition, and Population (HNP) Discussion Paper Investing in Adolescent Girls’ Nutrition in Bangladesh: A Situation Analysis of Trends and Ways Forward Malay Kanti Mridha,a Mokbul Hossain,b Tanvir Hassan,c Ipsita Sutradhar,d Samiun Nazrin Bente Kamal,e Akib Khan,f Nizam Uddin Ahmed,g Rudaba Khondker,h Piyali Mustaphi,i Ireen A. Chowdhury,j Alayne M. Adams,k and Ziauddin Hyderl a BRAC JPG School of Public Health, Bangladesh b BRAC JPG School of Public Health, Bangladesh c BRAC JPG School of Public Health, Bangladesh d BRAC JPG School of Public Health, Bangladesh e BRAC JPG School of Public Health, Bangladesh f IDinsight, Zambia, & World Bank Consultant g Shornokishoree Network Foundation (SKNK) h Global Alliance for Improved Nutrition (GAIN), Bangladesh i UNICEF, Bangladesh j UNICEF, Bangladesh k Department of International Health, Georgetown University, Washington, DC, USA & UNICEF Consultant l The World Bank Group, Washington, DC, USA This paper was prepared with funding from the World Bank and the United Nations Children’s Fund Abstract: Adolescents are among the age groups most vulnerable to malnutrition; their situation requires priority attention. However, information on adolescent nutrition in Bangladesh is limited. Using data from the Food Security and Nutrition Surveillance Project (FSNSP), we examined the nutritional situation of adolescent girls, including regional and urban-rural patterns in undernutrition and overnutrition, dietary diversity, household food security, as well as their growth dynamics. Our analysis focused on data collected from 2012 to 2014. The total sample size was 15,740 adolescent girls age 10 to 19 years, of which one-third were early adolescents (age 10 to14 years) and one-tenth lived in urban areas. We found that among younger adolescent girls (age 10 to 14), the proportion of moderate to severe thinness declined from 35 to 28 percent between 2012 and 2014, and rates of overweight and obesity were consistently low. For older adolescent girls (age 15 to 19), the proportion of moderate to severe thinness remained low across the study period, while rates of overweight and obesity increased from 13 to 23 percent between 2012 and 2014. Overall, 17 percent of younger adolescent girls were stunted in 2012, decreasing to 11 percent in 2014. Study findings also highlighted substantial regional variations in both age groups. Of particular concern was a decrease in dietary diversity. The proportion of younger adolescent girls falling into the poor dietary diversity group increased from 54 percent in 2012 to 60 percent in 2014, and for older adolescent girls, a similar pattern was evident, with rates increasing from 53 to 64 percent. The analysis of growth dynamics indicated substantial deficits relative to healthy norms in the younger adolescent period. Study findings emphasize the importance of leveraging critical developmental entry points through high-impact adolescent nutrition interventions. These investments will help ensure a future healthy workforce and a healthy next generation of children in Bangladesh. iii Keywords: adolescent, undernutrition, overnutrition, dietary diversity, food security Disclaimer: The findings, interpretations, and conclusions expressed in the paper are entirely those of the authors, and do not represent the views of the World Bank, its executive directors, or the countries they represent. Correspondence Details: S. M. Ziauddin Hyder, The World Bank Group, 1818 H Street, NW, Washington, DC; 202-458-9721; zhyder@worldbank.org. iv Table of Contents ACRONYMS ................................................................................................................. VII EXECUTIVE SUMMARY ......................................................................................... VIII ACKNOWLEDGMENTS ............................................................................................ XII CHAPTER 1: INTRODUCTION.................................................................................... 1 1.1 IMPORTANCE OF REDUCING ADOLESCENT MALNUTRITION IN BANGLADESH........ 1 1.2 THE FOOD SECURITY AND NUTRITION SURVEILLANCE PROJECT .......................... 3 1.3 OBJECTIVES OF ANALYSIS ..................................................................................... 4 CHAPTER 2: METHODS ............................................................................................... 4 2.1 SAMPLE DESIGN .................................................................................................... 4 2.2 DATA COLLECTION ............................................................................................... 6 2.3 DATA MANAGEMENT ............................................................................................ 7 2.4 DATA ANALYSIS ................................................................................................... 8 2.5 ETHICAL CONSIDERATIONS ................................................................................. 11 CHAPTER 3: RESULTS ............................................................................................... 12 3.1 CHARACTERISTICS OF ADOLESCENT GIRLS IN BANGLADESH .............................. 12 3.2 GEOGRAPHIC DISTRIBUTION OF MALNUTRITION AMONG ADOLESCENT GIRLS ... 13 3.3 FACTORS ASSOCIATED WITH ADOLESCENT UNDERNUTRITION AND OVERNUTRITION ............................................................................................................ 18 3.4 STATUS OF DIETARY DIVERSITY AMONG ADOLESCENT GIRLS AND FACTORS ASSOCIATED WITH DIETARY DIVERSITY ......................................................................... 27 3.5 STATUS OF FOOD SECURITY AMONG ADOLESCENT GIRLS AND FACTORS ASSOCIATED WITH FOOD SECURITY ............................................................................... 32 3.6 GROWTH DYNAMICS OF ADOLESCENT GIRLS...................................................... 36 3.7 POLICIES AND STRATEGIES SUPPORTING ADOLESCENT NUTRITION AND WELL- BEING 41 CHAPTER 4: DISCUSSION ......................................................................................... 43 CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS ............................... 45 ANNEX 1 ......................................................................................................................... 50 ANNEX 2 ......................................................................................................................... 52 Figure 2A.1: Trends in BMI among Early and Late Adolescent Girls in Rajshahi between 2012 and 2014........................................................................ 52 Figure 2A.2: Trends in BMI among Early and Late Adolescent Girls in Khulna between 2012 and 2014 ........................................................................................ 52 Figure 2A.3: Trends in BMI among Early and Late Adolescent Girls in Barisal between 2012 and 2014 ........................................................................................ 53 Figure 2A.4: Trends in BMI among Early and Late Adolescent Girls in Dhaka between 2012 and 2014 ........................................................................................ 53 v Figure 2A.5: Trends in BMI among Early and Late Adolescent Girls in Sylhet between 2012 and 2014 ........................................................................................ 54 Figure 2A.6: Trends in BMI among Early and Late Adolescent Girls in Chattogram between 2012 and 2014 .................................................................. 54 Figure 2A.7: Trends in BMI among Early and Late Adolescent Girls in Rangpur between 2012 and 2014 ........................................................................ 55 ANNEX 3 ......................................................................................................................... 56 Figure 3A.1: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Rajshahi between 2012 and 2014 .......... 56 Figure 3A.2: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Khulna between 2012 and 2014 ............. 56 Figure 3A.3: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescents Girls in Barisal between 2012 and 2014............ 57 Figure 3A.4: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Dhaka between 2012 and 2014 .............. 57 Figure 3A.5: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Sylhet between 2012 and 2014 ............... 58 Figure 3A.6: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Chattogram between 2012 and 2014 ..... 58 Figure 3A.7: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Rangpur between 2012 and 2014 .......... 59 REFERENCES ................................................................................................................ 60 vi ACRONYMS AOR Adjusted Odds Ratio BAZ BMI-for-Age Z BCC Behavior Change Communication BMI Body Mass Index DGFP Directorate General of Family Planning DGHS Directorate General of Health Services DP Development Partner FANTA Food and Nutrition Technical Assistance FCS Food Consumption Score FDS Food Deficit Scale FSNSP Food Security and Nutrition Surveillance Project GDP Gross Domestic Product HAZ Height-for-Age Z HFIAS Household Food Security Access Scale HKI Helen Keller International HNP Health, Nutrition, and Population HPNSIP Health, Population, and Nutrition Sector Investment Program INFS Institute of Nutrition and Food Sciences LMIC Lower- and Middle-Income Countries MDG Millennium Development Goal MOHFW Ministry of Health and Family Welfare MUAC Mid-Upper Arm Circumference NGO Nongovernment Organization NNS National Nutrition Services OP Operational Plan OR Odds Ratio NPAN 2 National Plan of Action on Nutrition II PDA Personal Digital Assistant PIP Program Implementation Plan RMNACH Reproductive, Maternal, Neonatal, Adolescent, and Child Health SES Socioeconomic Status SD Standard Deviation SPSS Statistical Package for Social Science TALC Teaching Aides at Low Cost UN United Nations WHO World Health Organization vii EXECUTIVE SUMMARY A window of opportunity for lifelong health and well-being Investing in adolescent girls’ nutrition is fundamental for economic growth and improved health in Bangladesh. Well-targeted investments to improve the nutritional status of adolescent girls will contribute to the country’s human capital and economic growth and break the intergenerational cycle of malnutrition. The policies of the government of Bangladesh are broadly supportive of adolescent nutrition. However, nutrition interventions targeted directly toward the specific needs of adolescent girls are relatively limited, partly due to a lack of evidence to justify required investments. This important study examines the nutritional situation of adolescent girls, including regional and urban-rural patterns in undernutrition and overnutrition, dietary diversity and household food security, as well as their growth dynamics during a critical period of physical and mental development. Data from the Food Security and Nutrition Surveillance Project (FSNSP) provide a unique opportunity to understand the nature and geographic dimensions of the nutritional status of adolescent girls, and how it is evolving over time. The FSNSP collected 15 rounds of surveillance data from a national representative sample of children under-five years, adolescent girls, and women of reproductive age, across 13 ecological zones and seven divisions of Bangladesh from 2010 to 2015. The analysis focused on data collected from 2012 to 2014. The total sample size was 15,740 adolescent girls, age 10 to 19 years, of which one-third were early adolescents (age 10 to 14) years, and one-tenth lived in urban areas. Areas of improvement and ongoing challenges The findings of the study highlight nutritional improvements and areas of concern. Among younger adolescent girls (age 10 to 14), the proportion of moderate to severe thinness declined from 35 to 28 percent between 2012 and 2014, and rates of overweight and obesity were consistently low. For older adolescent girls (age 15 to 19), the proportion of moderate to severe thinness remained low across the study period, while rates of overweight and obesity increased from 13 to 23 percent between 2012 and 2014. The findings highlighted substantial regional variations in both age groups. For example, in Rajshahi and Khulna divisions, patterns resembled national trends, but in Dhaka division, a steep increase and comparatively high rates of overweight and obesity were observed. In Barisal and Sylhet divisions, there was no increase in overweight and obesity among older adolescent girls. In both age groups of adolescent girls, those living in rural areas had a higher proportion of moderate to severe thinness and a lower proportion of overweight and obesity than those in urban areas. Overall, 17 percent of younger adolescent girls were stunted in 2012, decreasing to 11 percent in 2014. However, among older adolescent girls, stunting (short for their age) rates were much higher and increased from 38 percent in 2012 to 42 percent in 2014, reflected in the slowing of linear growth during this period. Differences in rural and urban areas were also notable. Among younger adolescent girls, the prevalence of stunting decreased from 18 to 13 percent in rural areas. In urban areas, the rate of decline was much steeper, almost halving from 15 percent in 2012 to 8 percent in 2014. viii Table ES1: Changes in Under- and Overnutrition by Division and Area of Residence Early adolescence (10–14 years) Late adolescence (15–19 years) Division Undernutrition Overweight Undernutrition Overweight Rajshahi Decreased No change Decreased Increased Khulna No change No change Decreased No change Barisal No change Increased No change Decreased Dhaka Decreased No change Decreased Increased Sylhet No change No change No change No change Chittagong No change Increased Increased Increased Rangpur No change No change Decreased Increased Rural Decreased Increased Decreased Increased Urban Decreased Inconsistent Decreased Increased In younger adolescent girls, undernutrition was associated with lower maternal education, while overnutrition was associated with greater household wealth. In older adolescent girls, undernutrition was associated with household food insecurity, whereas overnutrition was associated with household food security and urban residence. Of particular concern was worsening trends in dietary diversity. The proportion of younger adolescent girls falling into the poor dietary diversity group increased from 54 percent in 2012 to 60 percent in 2014, and for older adolescent girls, a similar pattern was evident, with rates increasing over 10 percent from 2012 to 2014—from 53 to 65 percent. Household food insecurity and lower wealth quintile were consistently associated with poor dietary diversity during early adolescence. In late adolescence, poor dietary diversity was also associated with rural residence, small household size, and lower maternal education. Figure ES1: Prevalence of Inadequate Dietary Diversity among Early (10–14 Years) and Late (15–19 Years) Adolescents 70 65 60 60 54 53 55 57 50 40 10-14 years 30 15-19 years 20 10 0 2012 2013 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh ix Food insecurity in the households of both younger and older adolescent girls improved over the three years under investigation. Among younger adolescent girls, the proportion falling into the severe food insecurity category declined from 54 percent in 2012 to 27 percent in 2014, and from 44 to 21 percent in the households of older adolescent girls. Likewise, the proportion of households with food security increased for both age groups over the three years under study. For both adolescent age groups, greater maternal education, dietary diversity, and household wealth had a positive association with household food security. The analysis of growth dynamics revealed steady gains in height from age 10 to 14 years, then a leveling off as full adult stature is attained. A slowing in the rate of growth faltering was apparent after 15 years, with predicted height-for-age Z (HAZ) scores leveling off at -1.6 to -1.8 below the norm. However, the average rate of growth faltering is -22 percent in the early adolescent period compared to -4 percent among older adolescents. For measures of undernutrition, rates are high at the start of early adolescence, decline between ages 12 to 13, then level off and decline slowly through the late adolescent period. Trends and implications for action This situation analysis of the nutrition and food security status of adolescent girls and their families revealed trends and geographic variations. Most encouraging was evidence of secular declines in severe food insecurity across the country. At the same time, results indicated a declining yet unacceptably high prevalence of moderate to severe thinness among younger adolescent girls and a trend toward increasing rates of overweight and obesity among late adolescents. Moreover, there is an indication of a shift from undernutrition to normal body mass index (BMI), and from normal BMI to overweight and obesity. The ideal programmatic goal is to increase the number of adolescents with normal BMI by reducing both undernutrition and overnutrition. Prevention of adolescent undernutrition and overnutrition will lead to increased work capacity and productivity, optimal maternal health and birth outcome, and prevention of non-communicable diseases (NCDs). Striking differences were apparent between different divisions throughout the country and between girls living in rural and urban areas. Smaller-scale studies suggest an even higher level of undernutrition among pregnant adolescents between 15 to 19 years of age, and similar levels of severe thinness in rural areas in particular. Another worrisome finding was a consistent increase in the proportion of early and late adolescents with poor dietary diversity over the three-year period of study. There is a need to develop region-specific plans that respond to the region-specific need for adolescent nutrition. Moreover, both early and late adolescents need to be included in the plan, though most early adolescents can be reached at school. Given national gains in household food security and female education, and known positive associations between food security, dietary diversity, and maternal schooling, trends, indicating worsening dietary diversity warrant further research and policy attention. These trends also raise concerns about the limitations of standard definitions of food security that fail to take dietary diversity into account. Key entry points for intervention Analysis reveals critical developmental entry points that need to be leveraged. Perhaps most crucial is the need to focus on the early adolescent period when nutritional x needs are greatest, and the most rapid period of growth faltering occurs. This can be done through school-based nutrition education and fortification programs, mid-morning snacks, and community-based nutrition interventions. The formation or support of existing school or community-based adolescent clubs and social media platforms offer promise as a means of involving both younger and older adolescents more directly in their own health and nutrition. Healthy dietary habits, daily physical activity, and sexual and reproductive health are also potential areas of focus, which can be animated by youth in terms of their design, delivery, and evaluation. The role of parents in supporting youth nutrition is similarly important, and can be facilitated through the dissemination of easy, affordable, and healthy family meals, and regular exercise through radio and television spots, and cell phone messaging. Enable public and private sector action Myriad policies and strategies relevant to adolescent health and nutrition are in place, as are legal, educational, social, and cultural rights and protections for girls and women. These can be mobilized to enable investments in adolescent girls’ nutrition. Of particular note are the Second National Plan of Action on Nutrition II (NPAN2), and the second Country Investment Plan (CIP2). Implementing these policies and strategies requires firm political commitment and a country investment plan focused on improving adolescent nutrition. Of particular importance is the National Plan of Action for Adolescent Health Strategy (2017–2030), which recognizes adolescent nutrition and the imperative for multiministerial engagement. Investments in planning, executing, and sustaining interventions that enable this plan must recognize the importance of responding to region- specific nutrition and food security needs, and the role of the Ministries of Trade and Agriculture in ensuring the variety and seasonal availability of healthy local foods. The powerful private sector food industry must also be incentivized to produce affordable nutritious food that appeals to young people, and to help diversify diet in a healthy direction. Partnerships with the private sector through existing business networks would help expedite engagement in terms of corporate social responsibilities and the creation and promotion of healthy products to youth audiences. Moreover, the production of unhealthy foods should be discouraged, and there should be laws to increase taxes on unhealthy foods and drink, for example, sugar-sweetened beverages. Prioritize intersectoral and multilevel investments A cost-benefit analysis of potential interventions is recommended to help prioritize intersectoral and multilevel investments, and the development of information systems that monitor the nutritional status, dietary behaviors, and activity levels of both adolescent girls and boys. These data will help measure the impact of nutrition- related investments by division, rural or urban residence, age group, and gender. By investing in adolescent girls’ nutrition interventions that yield high impact, Bangladesh will reap a triple dividend of benefits in the present, in a future healthy workforce, and in parents nurturing the next healthy generation of children in Bangladesh. xi ACKNOWLEDGMENTS This report is a collaborative effort of BRAC James P. Grant School of Public Health, the World Bank Group, the United Nations Children’s Fund (UNICEF), Shornokishoree Network Foundation (SKNF), and the Global Alliance for Improved Nutrition (GAIN). Data were furnished by the Food Security and Nutrition Surveillance Project (FSNSP), implemented by the BRAC James P Grant School of Public Health, Helen Keller International, Bangladesh, and Bangladesh Bureau of Statistics, and funded by the European Union. The authors are grateful to the World Bank for publishing this report as an HNP Discussion Paper. xii CHAPTER 1: INTRODUCTION 1.1 IMPORTANCE OF REDUCING ADOLESCENT MALNUTRITION IN BANGLADESH Adolescence constitutes a critical period for the achievement of optimal health and nutrition across the life course. Consequences of adverse exposures during this period of rapid growth and development can impact early adulthood and later life (Patton et al., 2016). The increased nutritional needs and diet-related behaviors established during the adolescent period are especially important (Akseer et al., 2017). A growing body of evidence links adolescent nutritional inadequacies and cardio-metabolic risk factors, like obesity, to the development of chronic disease, and premature mortality in adulthood (Saydah et al., 2013; Twig et al., 2016). For adolescent mothers, the risks of poor nutrition may also extend to the next generation and impact both fetal and infant growth, and later childhood development outcomes (Victora et al., 2008; UNFPA, 2013). Yet until recently, the health and nutrition of adolescents have been largely overlooked on the global health agenda (Patton et al., 2016; WHO, 2018). Numbering over 1.8 billion, they represent the largest cohort of young people in history (UNFPA, 2014). As such, the global population health benefits of optimizing nutrition in this age group, are enormous, and demand urgent attention Bangladesh has made significant strides in economic development despite political turbulence and continued vulnerability to natural disasters. From 1994 to 2016, the country averaged a steady annual gross domestic product (GDP) growth rate of 5.7 percent, and maintained relatively low inflation and stable domestic debt, and interest and exchange rates. Bangladesh has also made remarkable progress in many aspects of human development. In education, Bangladesh improved access, and reached the Millennium Development Goal (MDG) of gender parity at the primary and secondary levels. In Health, Nutrition, and Population (HNP), Bangladesh has achieved impressive declines in infant and child mortality rates, and in the maternal mortality ratio, attaining the MDGs 4 and 5, respectively. Improvements in nutrition outcomes have been occurring gradually since the 1990s but undernutrition rates remain “very high”. As of 2014, 36.1 percent of children under five years of age were stunted [short for their age], 14.3 percent were wasted [low weight for height], and 32.6 percent were underweight [low weight-for-age]). Food and nutritional insecurity among adolescent girls and adult women continues to pose significant public health challenges. In 2014, more than 50 percent of adolescent girls and adult women had diets deemed inadequate in terms of macro- and micronutrients. Moreover, recent data highlight a growing dual burden of malnutrition among adolescent girls, with disproportionate shares of underweight adolescents in poorer socioeconomic households and overweight or obese adolescents in wealthier households. Fragmentation, inefficiency, and weak governance remain major constraints to the implementation of multisectoral nutrition interventions in Bangladesh. To overcome some of these challenges, the government of Bangladesh has mainstreamed the implementation of a set of high-impact nutrition interventions into health and family welfare planning services, and produced the Second National Plan of Action on Nutrition II 1 (NPAN2) involving 18 ministries. Commitment to implementing evidence-based policies to accelerate the improvement of nutritional outcomes is present and growing. Greater emphasis on investments in public health nutrition approaches that emphasize prevention and mainstreaming nutrition into health, education, social protection, and agriculture sectors is necessary to achieve further gains in nutrition-related outcomes. Improving nutrition also requires in-depth understanding of the multiple causes of undernutrition at different stages of the life cycle and the optimum utilization of available evidence to guide resource allocation and its efficient use in nutrition programming. Adolescent girls are among the age group most vulnerable to malnutrition, and their situation is of particular concern. Apart from early childhood, adolescence is the second most crucial period of physical growth and change in body composition, physiology, and endocrine systems, during which time nutritional needs and risks are elevated. Adolescence is the last opportunity to reverse the growth faltering experienced during childhood and to support the growth spurt and skeletal development that are key to breaking the intergenerational cycle of undernutrition and poverty. One-fourth of adolescent girls in Bangladesh are stunted, and many are deficient in critical micronutrients. The growth deficit in this cohort, among adolescent girls in Bangladesh is as follows: 4 percent were severely short, 22 percent were moderately short, and 0 percent were tall; whereas corresponding figures from a “healthy” reference population were 0, 2, and 16 percent, respectively. Micronutrient deficiencies affecting adolescent girls are an additional concern, given their potential consequences for later reproductive outcomes. Investing in nutrition interventions for adolescent girls would help Bangladesh maximize the “demographic dividend” and contribute to the human capital and economic growth of the country. The size and growth of this demographic group and their potential contribution to the economy and realizing the demographic dividend is a resource of immense value in Bangladesh’s journey toward middle-income status. This dividend can be further enhanced by adequate investment in adolescent nutrition so that optimal stocks of health capital are available to the workforce. For example, a healthy diet can influence school performance due to its positive impacts on student health, cognition, concentration, and energy levels. Good nutritional status of adolescent girls is particularly important due to early marriage and childbearing in Bangladesh and to other practices associated with the persistence of childhood undernutrition. Children are more likely to be of low birthweight and to remain malnourished throughout the life cycle if their mothers were malnourished during the adolescence. In 2014, almost one-third of the adolescents age 15 to 19 were already mothers or pregnant with their first child. Importantly, the incidence of underweight in women of childbearing age was also the highest in this cohort. The health and nutrition of adolescent girls has remained relatively underresearched and underinvested in, in most lower-middle-income countries (LMICs). In Bangladesh, the FSNSP provides the most comprehensive source of dietary and anthropometric data on this age cohort, spanning the period from 2011 to 2015. As such, it presents a unique opportunity to generate evidence on the trends and factors 2 associated with malnutrition among adolescent girls useful to program development and the design of effective interventions. Analyses of FSNSP data will help identify knowledge gaps concerning the food and nutritional security of this comparatively neglected age group, which, in turn, could guide new data collection and research efforts, potentially in partnership with other nongovernmental organizations (NGOs), and bilateral and United Nations (UN) organizations working on this issue. The following section describes the FSNSP data and their relevance to the nutrition of adolescent girls. 1.2 THE FOOD SECURITY AND NUTRITION SURVEILLANCE PROJECT The FSNSP is a unique source of seasonal, nationally, and subnationally representative estimates of food security and nutrition in Bangladesh, focused on pregnant and nonpregnant women, adolescent girls, and households with and without children. Using state-of-the-art methods and indicators, the FSNSP also provides up-to-date information on the nutrition situation of households from specific agroclimatic zones that are at a greater risk of food and nutrition insecurity. The FSNSP and its predecessor, the National Nutrition Surveillance Program, have sought to quantify important behaviors that support a healthy population, and monitor seasonal variations and yearly trends in key nutrition indicators over the last 17 years. Providing a robust evidence base for decision making in the areas of food and nutrition, which complements routine government data collection, FSNSP data are widely used by policy makers, practitioners, and researchers alike. FSNSP used a three-stage sample design to ensure representativeness for 13 geographic regions (or strata). Six regions corresponded to special surveillance locations marked by endemic food and nutritional insecurity, owing to specific geographic and socioeconomic features, for example, northern regions are prone to river erosion or droughts, and coastal areas are vulnerable to saline intrusion and seasonal cyclones. The rest of the strata corresponded to the country’s seven administrative divisions— comprising areas that preclude the ones already included in the aforementioned special surveillance zones. Surveys were carried out over three different seasons each year: the post-aman harvest period (January to April), the height of the monsoon (May to August), and the post-aus harvest season (September to December), thereby enabling seasonal estimation of food and nutrition security. On average, more than 25,000 households were surveyed each year. A household was included in the survey if a female family member between the age of 10 and 49 years or a preschool child (i.e., below five years of age), was living in the household at the time of the interview. From each sampled household, one nonpregnant woman or an adolescent girl (10 to 49 years of age) was randomly selected for anthropometric measurements of height, weight, and mid-upper arm circumference (MUAC), and information on dietary diversity using a structured questionnaire. On average, more than 4,500 adolescent girls were surveyed and measured per year. In addition to anthropometric data and dietary recall over the last 24 hours, a separate questionnaire was also administered at the household level to collect information on household composition, socioeconomic status (SES), and food security, among other things. 3 1.3 OBJECTIVES OF ANALYSIS This secondary analysis of FSNSP data seeks to support the government of Bangladesh in making evidence-based decisions on multisectoral nutrition programming for adolescent girls. It addresses critical knowledge gaps with respect to the distribution, causes, and consequences of adolescent malnutrition focused on five objectives: 1. Determine the geographic distribution of adolescent girls’ malnutrition 2. Explore the factors associated with adolescent girls’ undernutrition and overnutrition 3. Investigate dietary diversity among adolescent girls and its associated factors 4. Assess household food insecurity of adolescent girls and its associated factors 5. Model the growth dynamics of adolescents to identify periods of faltering 6. Identify current policies and programs pertinent to adolescent nutrition, and make recommendations for action CHAPTER 2: METHODS This section describes the methodology of the FSNSP survey and how adolescent data were analyzed. Though FSNSP also collected data in 2010 (three rounds) and 2015 (one round), only twelve rounds from 2011 to 2014 were considered in this analysis (three rounds each year), as methods remained consistent during this period. Furthermore, while data were also representative of thirteen zones in Bangladesh (six highly vulnerable zones and seven divisions), we analyzed data at the division level. 2.1 SAMPLE DESIGN The FSNSP used a three-stage sampling design to reduce travel time and to provide a representative sample for each of 13 regions during 2011 to 2014. The six highly vulnerable regions were the coastal belt, eastern hills, haor, Padma chars, northern chars, and northwest, which coincided with Barisal, Chattogram, Dhaka, Khulna, Rajshahi, Rangpur, and Sylhet divisions. The vulnerable regions were hard-to-reach areas in Bangladesh and were inhabited by a large number of poor people. 4 Figure 1: FSNSP Surveillance Regions Source: Food Security and Nutrition Surveillance Project, Bangladesh In the first stage of sampling, 12 upazillas (subdistricts) were randomly selected per round from the six highly vulnerable zones (2 upazillas per zone), and 22 upazillas were selected from seven divisions—excluding upazillas already included in the sampling frame of the highly vulnerable zones. The number of upazillas selected from the divisions was proportionate to the number of remaining upazillas in the division and ranged from one to eight. From each of the 13 surveillance zones and for each round, upazillas were selected by rotation to reduce random variation in estimates between rounds, as recommended for surveillance systems by the United Nations, and is commonly done in labor participation surveillance. In the second stage, mohallas (villages) in each upazilla were broken into units of equal size before selection. There was no stratification of rural and urban areas during the second stage of selection; four communities were chosen at random and without replacement from all the mohallas in each selected upazilla. The third stage of sample selection occurred in the field. The data collection team approached the assigned mohalla starting from the first eligible house from a randomly assigned approach road (north, south, east, or west), as determined by a random number generator until 24 households were selected systematically and interviewed. The next and subsequent households for interview were chosen systematically by skipping four households from the previously interviewed household and, in a "zigzag" fashion, selecting households from both sides of the road. In situations where the identified household was not eligible for inclusion or refused participation, the 5 next household that met the inclusion criteria was selected. Households were considered eligible for surveillance if there was at least one woman in the household age 10 to 49 years or a preschool child, that is, below five years of age. In every household sampled, one nonpregnant woman or adolescent girl (10 to 49 years of age) was randomly selected for height/ weight/ mid-upper arm circumference (MUAC) measurement and was asked a series of questions about dietary consumption. In addition, all currently pregnant women had their MUAC measured and were asked about their dietary patterns and care they received during pregnancy. All children less than five years of age in the household were weighed and measured, but only the caretaker of the youngest child in each household answered questions about child feeding and morbidity relevant to that child. The sample size estimation was based on 14 key indicators: 4 child indicators, 4 women indicators, and 6 household indicators. These indicators were as follows: 1. Seasonal proportion of acute malnutrition; that is, weight for length/ height <-2 Z score among < five-year–old children 2. Seasonal proportion of underweight; that is, weight-for-age <-2 Z score among < five-year–old children 3. Annual proportion of underweight; that is, weight-for-age <-2 Z score among < five- year–old children 4. Annual proportion of stunting; that is, height/ length-for-age <-2 Z score among < five-year–old children 5. Seasonal proportion of women with chronic energy deficiency; that is, BMI < 18.5 kg/m2 6. Annual proportion of women with chronic energy deficiency; that is, BMI < 18.5 kg/m2 7. Seasonal proportion of women with overweight and obesity; that is, BMI > 23.0 kg/m2 8. Annual proportion of women with overweight and obesity; that is, BMI > 23.0 kg/m2 9. Seasonal proportion of households with food insecurity 10. Annual proportion of households with food insecurity 11. Seasonal proportion of households with food deficits 12. Annual proportion of households with food deficits 13. Seasonal proportion of households with borderline food consumption pattern 14. Annual proportion of households with borderline food consumption pattern Between 2012 and 2014, total target households for data collection in each round were 9,024 in number. 2.2 DATA COLLECTION Surveillance data were collected through structured interviews using paper-based questionnaires and proprietary survey software (Surveymaster Version 1.0 and Version 2.0, Helen Keller International) administered using commercially available personal digital assistants (PDAs; Hewlett Packard, HP iPAQ 112, USA). The questions were the same on the paper questionnaire and the PDA. Approximately two-thirds of the data were collected using PDAs. To the extent possible, surveillance questionnaires and protocols employed by FSNSP were based on existing global standards and guidelines, for example, Household Food Insecurity Access Scale (HFIAS), Food Deficit Scale (FDS), and Food Consumption Score (FCS). Data collectors received two weeks of initial training on the study objectives and methods, interview techniques, use of PDAs for questionnaire administration, 6 anthropometric measurements, and maintaining anthropometric instruments. Before each surveillance round, a one-week refresher training was conducted to share lessons learned from the field and discuss any changes in the questionnaire. Before each round of data collection, field testing and back translation of questionnaires were undertaken for any altered sections of the questionnaire. After each field test, monitoring officers and other supervisors revised the questionnaire and the standard operating procedures in collaboration with the enumerators. Midway through each round of data collection, a one- or two-day refresher training was also organized to reinforce skills and knowledge. In each selected household, the weight of children, women, and adolescent girls was measured to the nearest 0.1 kg, using a portable electronic weighing scale (TANITA Corporation Japan, model HD-305). The height of women, adolescent girls, and children older than two years of age, and the recumbent length of children younger than two years were measured to the nearest 0.1 cm, using a locally made height and length board. The mid-upper arm circumference of children, adolescent girls, and women (both pregnant and nonpregnant) was measured to the nearest 0.2 mm, using a numerical insertion tape produced by Teaching Aides at Low Cost (TALC). All anthropometric measurements were performed according to WHO guidelines, as specified in the Food and Nutrition Technical Assistance (FANTA) anthropometry manual. Monitoring officers supervised the activities of every team, and two field managers provided oversight of the data collection process. The three rounds of data collection were divided into two phases. The data collection teams spent four to six weeks at a time in the field. The monitoring officers visited each data collection team at random at least once a week to check questionnaires and ensure adherence to the questionnaire protocol. All data collected were reviewed and cross-checked by monitoring officers to ensure quality. Monitoring officers reviewed completed questionnaires at the time of survey so that any errors or inconsistencies identified could be corrected in the field. To verify the quality of data, quality control officers revisited a randomly selected subsample (around 5 percent) of interviewed households within 48 hours of the initial visit by the data collectors. To recheck, data collectors compared the surveillance data to the quality control data. Inconsistencies, if any, were reviewed by the director of the project, project coordinator, training officers, and the field manager to identify possible reasons for the discrepancy and to implement appropriate solutions. Quality control operations were supplemented by Bangladesh Bureau of Statistics (BBS) staff that performed a 10 percent postenumeration check using a shortened questionnaire. 2.3 DATA MANAGEMENT Data entry or import from PDA was done concurrently with data collection. Data obtained using paper questionnaires were entered using a data entry program developed in FoxPro software (Version 2.6), while PDA data were imported using Surveymaster (HKI, Version 2.0). One senior data management officer supervised data entry and cleaning, including the transfer of data from PDAs to computer and merging the data from paper questionnaires and PDAs using SPSS (IBM, Version 19.0). Afterward, data management officers reviewed, edited, and cleaned the data by performing a series of logic, frequency, and data range checks. Any inconsistencies identified were checked visually by comparing the electronic entry to the entry on the original questionnaire or to the data collectors’ notebooks. Data management officers consulted with field managers and monitoring 7 officers to understand any discrepancies found during the data cleaning process. Only the senior data management officer had permission to make necessary corrections. 2.4 DATA ANALYSIS Data analysis was done using Stata (StataCorp, Version 13.0). For Objective 1, which concerned the geographic and socioeconomic distribution of adolescent malnutrition, and for Objectives 2- 4, which focused on the status of nutrition, dietary diversity and food insecurity, respectively, descriptive analysis was performed. Estimates were adjusted using sampling weights that were constructed based on each adolescent’s probability of selection, and analyzed using the svy commands in Stata 13.0, to take into account the complex sampling design. Adolescents were divided into two categories: early adolescents (10 to 14 years) and late adolescents (15 to19 years). Analysis was carried out according to these two subgroups. For Objectives 2, 3, and 4, logistic regression was performed. For Objective 2, the outcome variables were undernutrition (defined as BMI <18.5 kg/m2), overnutrition (defined as BMI >23.0 kg/m2), and low MUAC (defined as MUAC <21.0 cm). The outcome variable for Objective 3 was poor dietary diversity (defined as consumption of <5 food groups in the last 24 hours), and for Objective 4, the outcome variable was any food insecurity (defined as mild, moderate, or severe food insecurity in the last four weeks). Before approaching the logistic regression, a comprehensive literature review identified key distal, intermediate, and proximal factors associated with malnutrition (undernutrition and overnutrition combined), dietary diversity, and food security. Proximal factors were variables closely related to the outcome (individual level factors), whereas distal factors were variables that captured broader or macro aspects of context that act via a number of intermediary factors at the household level. Three conceptual frameworks were developed that visualize the presumed interrelationships between distal, intermediate, and proximate factors with the outcome variable of interest (see Figures 2, 3, and 4). Variables that were not available in the FSNSP data but represent important explanatory factors are indicated in dark pink. Prior to logistic regression, bivariate analyses were conducted to assess the association of proximal, intermediate, and distal factors with each of the outcome variables. A multivariable logistic regression analysis was performed, including all variables having a bivariate association with the outcome variables p ≤ 0.1. The model fit was assessed using the Hosmer-Lemeshow test for goodness-of-fit. No interaction terms were included in the model. For Objective 5, the growth dynamics of adolescent girls were modeled using a pseudo-cohort method where cohorts are constructed using repeated cross- sections in the absence of longitudinal data at the individual level. Repeated cross- sections of data were combined across 12 rounds of FSNSP data from 2011 to 2014, representing approximately 4,500 adolescents, and growth curves for height, weight, stunting, and BMI were constructed. The validity of this approach was supported by the fact that adolescents were randomly sampled across consecutive years, and thus, were statistically similar for a given cohort over time. This analytical strategy has been employed in similar study settings where longitudinal data are not available. To further verify the robustness of the method, estimates were compared with findings from the IFPRI Bangladesh Integrated Household Panel Survey (2011–12 and 2015), which provides two data points for each individual in the survey. Reasonable correspondence was noted. 8 Growth curves were created for several measures of nutritional status including (1) height-for-age Z or HAZ scores, which represents a measure of chronic undernutrition; (2) Stunting, which measures the degree of chronic undernutrition with reference to a cutoff of HAZ <-2 (moderately or severely short for age; (3) BMI-for-age Z or BAZ scores, which captures acute malnutrition, and finally (4) Undernourished, which measures the degree of acute malnutrition with reference to a cutoff of BAZ <-2 (moderately or severely thin). Local polynomial smoothing was used to enable easy visualization of trends. Furthermore, growth dynamics were stratified by wealth quintiles to assess socioeconomic inequities in the adolescent growth trajectory. Figure 2: Conceptual Framework Showing Factors associated with Adolescent Malnutrition Source: Food Security and Nutrition Surveillance Project, Bangladesh 9 Figure 3: Conceptual Framework Showing Factors associated with Adolescent Dietary Diversity Source: Food Security and Nutrition Surveillance Project, Bangladesh 10 Figure 4: Conceptual Framework Showing Factors associated with Food Security of Adolescents’ Households Source: Food Security and Nutrition Surveillance Project, Bangladesh For Objective 5, we conducted a comprehensive desk review of the available grey literature relevant to the health and nutrition of adolescent girls, inclusive of policy documents, government reports, strategy papers, as well as research and program reports from local partners and stakeholders. Semi-structured interviews and consultative meetings with a variety of key informants assisted in directing us to relevant material, and provided additional supplementary insight. Among the parties consulted were representatives of the Ministry of Health and Family Welfare (MOHFW), Directorate General of Health Services (DGHS), Directorate General of Family Planning (DGFP), National Nutrition Services (NNS), Institute of Public Health Nutrition (IPHN), Institute of Nutrition and Food Sciences (INFS), international and local NGOs, development partners (DPs), and private sector actors working in the area of adolescent health and nutrition. 2.5 ETHICAL CONSIDERATIONS FSNSP staff, including data collectors, monitoring officers, and field managers explained the objectives and procedures of the research to the leaders of the selected districts, upazilla (subdistricts), and mohallas (villages). At the beginning of each interview, data collectors provided details about the purpose of the study and read a statement clarifying that participation was voluntary and that those who granted consent had the right to refuse to answer any questions and/ or discontinue the interview at any time. The interview took place after the participants had given their consent for the interview. Consent for measuring children less than five years of age and adolescents less 11 than 18 years of age was obtained from their caregivers. The Institutional Review Board of the BRAC James P. Grant School of Public Health approved the study. CHAPTER 3: RESULTS 3.1 CHARACTERISTICS OF ADOLESCENT GIRLS IN BANGLADESH We analyzed data from 15,740 adolescent girls—5,803 from three rounds in 2012, 4,621 from three rounds in 2013, and 5,316 from three rounds in 2014. As presented in Table 1, depending on the year of data collection, 34 to 37 percent of the girls were in early adolescence (10 to 14 years), and 10 to 13 percent of them were from urban areas. Among mothers of adolescent girls, 55 to 59 percent (depending on year) had secondary or higher education. The proportion of households with any form of food insecurity was 60, 38, and 28 percent in 2012, 2013, and 2014, respectively. Table 1: Sociodemographic Characteristics of Adolescent Girls in Bangladesh (2012–2014) 2012 2013 2014 (n = 5,803) (n = 4,621) (n = 5,316) Age (in years) Mean ± SD 15.37 ± 2.60 15.42 ± 2.62 15.59 ± 2.60 (All adolescents) Area of residence Rural 5,100 (87.89) 4,150 (89.81) 4,630 (87.10) Urban 703 (12.11) 471 (10.19) 686 (12.90) Division Rajshahi 839 (14.46) 583 (12.62) 613 (11.53) Khulna 406 (7.00) 312 (6.75) 333 (6.26) Barisal 458 (7.89) 309 (6.69) 632 (11.89) Dhaka 1,270 (21.89) 1,111 (24.04) 1,328 (24.98) Sylhet 657 (11.32) 384 (8.31) 595 (11.19) Chattogram 1,341 (23.11) 1,172 (25.36) 1,076 (20.24) Rangpur 832 (14.34) 750 (16.23) 739 (13.90) Wealth quintile Lowest 1,161 (20.01) 918 (19.87) 1,051 (19.77) Second 1,234 (21.26) 922 (19.95) 1,055 (19.85) Middle 1,241 (21.39) 917 (19.84) 1,060 (19.94) Fourth 1,142 (19.68) 935 (20.23) 1,071 (20.15) Highest 1,025 (17.66) 929 (20.10) 1,079 (20.30) Food security Food secure 2,321 (40.00) 2,865 (62.00) 3,838 (72.20) Mild to moderate food 699 (12.05) 258 (5.58) 257 (4.83) insecure Severe food insecure 2,783 (47.96) 1,498 (32.42) 1,221 (22.97) Religion Muslim 5,027 (86.63) 3,947 (85.41) 4,520 (85.03) Non-Muslim 776 (13.37) 674 (14.59) 796 (14.97) Maternal education Primary completed 865 (14.91) 695 (15.04) 784 (14.75) 12 No formal and partial 1,719 (29.62) 1,197 (25.90) 1,389 (26.13) primary Secondary or more 3,219 (55.47) 2,729 (59.06) 3,143 (59.12) Household size <5 members 2,161 (37.24) 1,800 (38.95) 2,043 (38.43) ≥5 members 3,642 (62.76) 2,821 (61.05) 3,273 (61.57) Mean ± SD 5.26 ± 1.96 5.21 ± 1.95 5.25 ± 1.97 Note: SD = Standard deviation. Source: Food Security and Nutrition Surveillance Project, Bangladesh The mean age of girls in the early adolescent group was approximately 12.5 years, and in the late adolescent group, it was 17.0 years (see Annex 1). 3.2 GEOGRAPHIC DISTRIBUTION OF MALNUTRITION AMONG ADOLESCENT GIRLS Table 2 describes the mean height, weight, and MUAC of adolescent girls. In the full sample, between 2012 and 2014, the mean height, weight, and MUAC were about 149 cm, 42 kg, and 23 cm, respectively. Among early adolescent girls, the mean height, weight, and MUAC were about 146 cm, 37 kg, and 22 cm, respectively; and for late adolescents, the mean height, weight, and MUAC were about 151 cm, 45 kg, and 24 cm, respectively. Table 2: Height, Weight, and MUAC of Adolescent Girls in Bangladesh (2012–2014) Participant type Year (n) 2012 2013 2014 All participants (n = 5,803) (n = 4,621) (n = 5,316) Height in cm 149.26 ± 7.07 149.44 ± 6.76 149.84 ± 6.67 (Mean ± SD) Weight in kg 41.71 ± 8.19 42.12 ± 8.31 42.57 ± 8.13 (Mean ± SD) MUAC in cm 23.05 ± 2.89 23.31 ± 2.93 23.48 ± 2.88 (Mean ± SD) Early 2012 2013 2014 adolescents (n = 2,110) (n = 1,694) (n = 1,832) Height in cm 145.81 ± 7.97 146.32 ± 7.51 146.91 ± 7.39 (Mean ± SD) Weight in kg 36.55 ± 7.82 37.20 ± 7.88 37.72 ± 7.91 (Mean ± SD) MUAC in cm 21.24 ± 2.70 21.60 ± 2.69 21.75 ± 2.72 (Mean ± SD) 2012 2013 2014 Late adolescents (n = 3,692) (n = 2,927) (n = 3,484) Height in cm 151.24 ± 5.61 151.25 ± 5.54 151.38 ± 5.68 (Mean ± SD) Weight in kg 44.65 ± 6.83 44.96 ± 7.14 45.12 ± 7.01 (Mean ± SD) MUAC in cm 24.08 ± 2.46 24.30 ± 2.58 24.39 ± 2.52 (Mean ± SD) Note: SD = Standard deviation. Source: Food Security and Nutrition Surveillance Project, Bangladesh 13 Figure 5: classifies undernutrition of early and late adolescent groups based on body mass index (BMI). Among early adolescent girls, the proportion of moderate to severe thinness declined from 35 to 28 percent between 2012 and 2014, and rates of overweight and obesity were consistently low. For late adolescent girls, the proportion of moderate to severe thinness remained low across the study period, while rates of overweight and obesity increased from 13 to 23 percent between 2012 and 2014. Figure 5: Trends in BMI among Early and Late Adolescent Girls in Bangladesh (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh Annex 2 presents division-specific figures showing trends in BMI among adolescent girls between 2012 and 2014 (Figures 2A.1 to 2A.7). In Rajshahi and Khulna divisions, patterns are mostly consistent with national trends of declining moderate to severe thinness among girls in the early adolescent period and increasing rates of overweight and obesity in later adolescence (Figures 2A.1 and 2A.2). Barisal division also saw a decline in comparatively high levels of moderate to severe thinness from 43 to 36 percent in early adolescent girls, and no increase in overweight and obesity in the older age group (9 percent in 2012 and 7 percent in 2014) (Figure 2A.3). Although similar overall trends were observed in the early adolescent period, what was remarkable in Dhaka division was the steep increase and comparatively high rate of overweight and obesity in the late adolescent group, rising from 18 percent in 2012 to 32 percent in 2014 (Figure 2A.4). Sylhet division also experienced declines in moderate to severe thinness among early adolescents; however, trends in the late adolescent period were quite different. In Chattogram, no increase in overweight and obesity was apparent in the late adolescent period (Figure 2A.6), while in Sylhet, overweight and obesity rates doubled (8 percent in 2012 to 19 percent in 2014) (Figure 2A.5), but at far lower absolute levels than Dhaka. Finally, in Rangpur, no substantial declines in moderate to severe thinness occurred in early adolescence. Rates of obesity and overweight among girls in late adolescence were low overall, increasing from 6 percent in 2012 to 9 percent in 2014 (Figure 2A.7). 14 Figures 6 and 7 below display differences in BMI of adolescent girls, comparing rural and urban areas. Girls in the early adolecent period living in rural areas had a higher proportion of moderate to severe thinness and a lower proportion of overweight and obesity than those in urban areas (Figure 13). The same is true for late adolecents, although the proportion of girls with moderate to severe thinness was substantially lower in both rural and urban areas compared to those in the early adolescent period (Figure 14). Figure 6: Trends in BMI among Early Adolescent Girls Living in Rural vs. Urban Areas (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 7: Trends in BMI among Late Adolescent Girls in Rural vs. Urban Areas (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh 15 Figure 8 shows the prevalence of stunting, comparing early and late adolescent girls at the national level. Overall, 17 percent of girls in early adolescence were stunted in 2012, and 11 percent were stunted in 2014. Rates were much higher among girls in the late adolescent period, increasing from 38 percent in 2012 to 42 percent in 2014. Figure 8: Trends in Height-for-Age Z Score (HAZ) among Early and Late Adolescent Girls in Bangladesh (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh Annex 3 includes division-specific figures showing trends in stunting, comparing girls in early and late adolescent age groups over the period 2012 to 2014 (Figures 3A.1 to 3A.7). In Rajshahi, the prevalence of stunting was 10 percent in 2012 and 12 percent in 2014 among girls in the early adolescent period. Among older adolescents, the stunting prevalence was substantially higher but decreased slightly from 32 percent in 2012 to 29 percent in 2014 (Figure 3A.1). In Khulna, the prevalence of stunting hovered between 6 and 13 percent over the period 2012 to 2014 among early adolescent girls, and for those in the older adolescent group, the prevalence declined from 31 to 24 percent (Figure 3A.2). In contrast, among early adolescent girls in Barisal, stunting rates decreased quite dramatically from 18 percent in 2012 to 5 percent in 2014, but remained between 27 to 37 percent in the older adolescent age group (Figure 3A.3). In Dhaka and Sylhet divisions, similar trends in early adolescent stunting were apparent. The prevalence of stunting in Dhaka declined from 21 percent in 2012 to 13 percent in 2014, and from 19 to 14 percent during the same period in Sylhet. Among older adolescents, the prevalence of stunting increased in Dhaka from 41 to 50 percent from 2012 to 2014 (Figure 3A.4), and declined in Sylhet, from 51 to 41 percent over the same period (Figure 3A.5). 16 In Chattogram, the prevalence of stunting was 16 percent in 2012 and 6 percent in 2014 among early adolescent girls, and remained between 31 to 37 percent among girls in the late adolescent age group (Figure 3A.6). In contrast, in Rangpur, the prevalence of stunting in the early adolescent group changed little over the period 2012 to 2014, 13 percent and 11 percent, respectively. However, among older adolescents, stunting prevalence increased from 36 to 41 percent (Figure 3A.7). Stunting trends for early and late adolescent age groups in rural and urban areas are presented in Figures 9 and 10, for years 2012 to 2014. Among early adolescents, the prevalence of stunting decreased from 18 to 13 percent in rural areas; however, in urban areas the rate of decline is much steeper, almost halving from 15 percent in 2012 to 8 percent in 2014. Figure 9: Trend of Height-for-Age Z Score (HAZ) Categories among Early Adolescent Girls in Rural and Urban Areas (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh 17 Figure 10: Trend of Height-for-Age Z Score (HAZ) Categories among Late Adolescent Girls in Rural and Urban Areas (2012–2014) Source: Food Security and Nutrition Surveillance Project, Bangladesh 3.3 FACTORS ASSOCIATED WITH ADOLESCENT UNDERNUTRITION AND OVERNUTRITION Factors associated with undernutrition among early adolescents. Table 3 presents results from bivariate and multivariable analyses of factors associated with undernutrition (underweight = BMI <18.5 kg/m2) among early adolescent girls. Table 3: Factors associated with Undernutrition among Early Adolescent Girls (2012–2014) 10–14 years (Early adolescent girls) 2012 2013 2014 (n = 2,110) (n = 1,694) (n = 1,832) Variables OR AOR OR AOR OR AOR (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 1.28 1.27 0.95 1.18 1.01 1.18 (0.93– (0.90– (0.66– (0.79– (0.71– (0.81– 1.76) 1.78) 1.39) 1.76) 1.46) 1.72) Khulna 1.37 1.50 0.78 0.76 0.69 1.37 (0.88– (0.95– (0.48– (0.45– (1.81– (0.83– 2.11) 2.36) 1.28) 1.28) 1.12) 2.26) Barisal 1.34 1.46 0.87 0.99 1.17 1.36 (0.89– (0.94– (0.57– (0.63– (0.82– (0.94– 2.01) 2.25) 1.33) 1.54) 1.68) 1.96) Sylhet 1.74 1.38 1.42 1.29 1.46 1.41 (1.18– (0.92– (0.91– (0.81– (1.01– (0.96– 2.56)* 2.06) 2.21) 2.05) 2.12)* 2.07) Chattogram 1.38 1.26 1.06 0.98 0.67 0.79 (1.03– (0.93– (0.79– (0.72– (0.50– (0.58– 1.85)* 1.72) 1.43) 1.35) 0.89)* 1.07) 18 Rangpur 0.97 0.85 0.87 0.86 0.87 0.96 (0.71– (0.61– (0.62– (0.61– (0.62– (0.67– 1.32) 1.19) 1.21) 1.22) 1.22) 1.36) Dhaka Ref Food security Mild food 1.27 1.14 1.23 0.96 1.25 — insecure (0.82– (0.72– (0.68– (0.52– (0.73– 1.96) 1.79) 2.20) 1.76) 2.15) Moderate 1.91 1.67 1.52 1.08 1.78 — food insecure (1.22– (1.05– (0.72– (0.49– (0.81– 2.99)* 2.68)* 3.22) 2.37) 3.92) Severe food 1.67 1.34 1.18 0.91 1.15 — insecure (1.35– (1.06– (0.94– (0.71– (0.91– 2.07)* 1.70)* 1.48) 1.18) 1.44) Food secure Ref Area of residence Rural 1.88 1.69 1.67 1.45 1.13 — (1.42– (1.23– (1.19– (1.00– (0.84– 2.50)* 2.32)* 2.34)* 2.09)* 1.52) Urban Ref Household size <5 members 0.95 — 0.62 0.65 0.81 0.81 (0.77– (0.49– (0.51– (0.65– (0.65– 1.17) 0.77)* 0.82)* 1.00)* 1.02) ≥5 members Ref Religion Muslim 1.25 — 1.50 1.44 1.88 1.63 (0.92– (1.13– (1.05– (1.43– (1.21– 1.71) 2.01)* 1.96)* 2.47)* 2.20)* Non-Muslim Ref Maternal education No formal and 3.49 3.49 3.73 3.67 2.84 2.86 partial (2.77– (2.74– (2.89– (2.82– (2.25– (2.25– primary 4.38)* 4.46)* 4.82)* 4.79)* 3.57)* 3.62)* Primary 1.63 1.59 1.80 1.83 1.57 1.50 completed (1.24– (1.21– (1.35– (1.36– (1.20– (1.14– 2.13)* 2.10)* 2.41)* 2.46) 2.05)* 1.98)* Secondary or Ref more Dietary diversity 5 or more 0.87 — 0.89 — 0.85 0.96 food items (0.71– (0.71– (0.69– (0.77– 1.07) 1.11) 1.05) 1.20) <5 food Ref item(s) Asset quintile Lowest 1.56 0.77 1.26 1.05 1.00 — (1.13– (0.53– (0.90– (0.72– (0.73– 2.15)* 1.12) 1.76) 1.52) 1.35) Second 1.65 1.05 1.34 1.13 1.12 — 19 (1.19– (0.73– (0.96– (0.78– (0.82– 2.27)* 1.50) 1.87) 1.63) 1.53) Middle 1.35 0.90 1.09 0.92 1.19 — (0.98– (0.64– (0.78– (0.64– (0.86– 1.85) 1.27) 1.54) 1.33) 1.65) Fourth 1.01 0.89 0.99 0.97 1.17 — (0.73– (0.64– (0.71– (0.69– (0.86– 1.40) 1.26) 1.38) 1.38) 1.61) Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; Cl =Confidence Interval; — = Not available. * p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh Bivariate results show significant associations (p < 0.05) between underweight and division of residence, food security, area of residence (i.e., rural or urban), household size, religion, maternal education, dietary diversity, and asset quintile in at least one of the years under study (2012 to 2014). Among these variables, only maternal education was significantly associated with undernutrition in all years. Multivariable analysis did not yield any consistent patterns in the association between undernutrition and most variables under study. Associations with underweight were found for at least one year under study for area of residence, food security, household size, and religion, but only maternal education is significantly associated with undernutrition in all three years under analysis. Compared to early adolescents whose mothers had secondary or higher education, girls whose mothers had no formal education or partial primary education had higher odds of being underweight (Adjusted Odds Ratio [AOR] 3.49 in 2012; AOR 3.67 in 2013; and AOR 2.86 in 2014). Factors associated with undernutrition among late adolescent girls. Table 4 shows results from bivariate and multivariable analyses regarding the factors associated with underweight among girls in late adolescence. Bivariate analysis indicates significant associations (p < 0.05) between underweight and division of residence, food security, area of residence, household size, religion, and wealth quintile in at least one of the three years under study (2012 to 2014). Among these variables, division of residence, food security, religion, and wealth quintile were significantly associated with underweight in all three years. Unlike the younger age group, at the bivariate level, maternal education and dietary diversity were not associated with undernutrition among late adolescent girls. Multivariable analysis yielded no consistent patterns between underweight and any of the variables under analysis with the exception of food security, which was significantly associated with underweight in both bivariate and multivariable analysis in all three years under analysis. Compared to the girls in late adolescence living in food secure households, those in severe food insecure households had higher odds of being underweight (AOR 1.19 in 2012; 1.26 in 2013; and 1.35 in 2014). Factors associated with overnutrition of early adolescent girls. Table 5 shows results from bivariate and multivariable analyses of factors associated with overnutrition (defined as BMI >23.0 kg/m2), in the early adolescent age group. Bivariate results show that overweight was significantly associated (p < 0.05) with every variable in the model, with the exception of division of residence, in at least one of the years under study (2012 to 20 2014). Among these variables, area of residence and wealth quintile were significantly associated with overweight in all three years. Multivariable analysis yielded few consistent patterns between overweight and the variables under study. For example, religion was associated with overnutrition only in 2013 and 2014, while maternal education was associated only in 2013. The only variables that was consistently associated with overnutrition in early adolescent girls was area of residence and wealth quintile. Compared to those living in rural areas, early adolescent girls living in urban areas had higher odds of being overweight (AOR 3.18 in 2012; AOR 2.59 in 2013; and AOR 2.73 in 2014). Similarly, early adolescent girls living in households falling into the richest quintile have higher odds of being overweight than those living in households belonging to poorer wealth quintiles. Moreover, the strength of association between overweight and wealth quintile increased as wealth quintile increased. Table 4: Factors associated with Undernutrition among Late Adolescent Girls (2012–2014) 15–19 years (Late adolescent girls) 2012 2013 2014 (n = 3,693) (n = 2,927) (n = 3,484) Variables OR AOR OR AOR OR AOR (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 1.45 1.76 1.04 1.08 0.98 1.00 (1.16– (1.40– (0.80– (0.83– (0.77– (0.79– 1.83)* 1.11)* 1.34) 1.40) 1.25) 1.28) Khulna 1.43 1.41 0.89 0.91 0.93 0.93 (1.08– (1.06– (0.65– (0.66– (0.69– (0.69– 1.90)* 1.88)* 1.22) 1.25) 1.25) 1.26) Barisal 1.44 1.40 1.26 1.24 1.15 1.10 (1.09– (1.06– (0.90– (0.88– (0.91– (0.86– 1.89)* 1.85)* 1.76) 1.73) 1.47) 1.40) Sylhet 1.33 1.27 1.36 1.34 1.38 1.31 (1.05– (0.99– (1.01– (0.99– (1.07– (1.02– 1.70)* 1.63) 1.84)* 1.82) 1.76)* 1.69)* Chattogram 0.91 0.91 0.86 0.83 0.72 0.74 (0.74– (0.74– (0.69– (0.66– (0.58– (0.59– 1.12) 1.12) 1.07) 1.04) 0.89)* 0.93)* Rangpur 1.21 1.16 0.96 0.93 1.03 0.98 (0.96– (0.92– (0.75– (0.72– (0.82– (0.78– 1.52)* 1.47) 1.22) 1.19) 1.29) 1.23) Dhaka Ref Food security Mild food 1.22 1.17 1.48 1.45 1.36 1.32 insecure (0.92– (0.87– (0.97– (0.95– (0.90– (0.87– 1.63) 1.56) 2.25) 2.22) 2.06) 1.99) Moderate food 1.47 1.41 1.01 0.97 1.02 0.91 insecure (1.10– (1.05– (0.60– (0.58– (0.57– (0.51– 1.97)* 1.89) 1.70) 1.65) 1.82) 1.63) 21 Severe food 1.23 1.19 1.26 1.26 1.40 1.35 insecure (1.06– (1.02– (1.07– (1.06– (1.19– (1.13– 1.41)* 1.39)* 1.48)* 1.50)* 1.66)* 1.61)* Food secure Ref Area of residence Rural 1.29 1.20 1.35 1.28 0.93 — (1.05– (0.97– (1.05– (0.98– (0.76– 1.59)* 1.49) 1.73)* 1.66) 1.14) Urban Ref Household size <5 members 1.04 — 0.89 0.89 0.99 — (0.91– (0.76– (0.76– (0.87– 1.19) 1.03) 1.03) 1.14) ≥5 members Ref Religion Muslim 1.41 1.36 1.18 1.20 1.40 1.25 (1.16– (1.12– (0.96– (0.97– (1.15– (1.01– 1.71)* 1.66)* 1.46) 1.49) 1.70)* 1.55)* Non-Muslim Ref Maternal education No formal and 1.00 — 0.99 — 1.05 — partial primary (0.84– (0.81– (0.87– 1.19) 1.21) 1.26) Primary 1.00 — 1.08 — 1.01 — completed (0.82– (0.86– (0.82– 1.22) 1.35) 1.25) Secondary or Ref more Dietary diversity 5 or more food 1.00 — 0.90 — 1.08 — items (0.87– (0.77– (0.93– 1.14) 1.05) 1.25) <5 food item(s) Ref Asset quintile Lowest 1.14 1.05 1.14 1.08 1.16 1.11 (0.91– (0.82– (0.90– (0.84– (0.93– (0.88– 1.42) 1.33) 1.45) 1.39) 1.45) 1.41) Second 1.29 1.16 1.32 1.24 1.42 1.36 (1.04– (0.93– (1.04– (0.97– (1.14– (1.08– 1.59)* 1.46) 1.67)* 1.58) 1.76)* 1.70)* Middle 1.25 1.17 1.11 1.06 1.33 1.30 (1.01– (0.94– (0.88– (0.83– (1.07– (1.04– 1.54)* 1.45) 1.40) 1.35) 1.65)* 1.62)* Fourth 1.09 1.08 1.07 1.05 1.11 1.10 (0.88– (0.87– (0.85– (0.83– (0.89– (0.88– 1.35) 1.34) 1.35) 1.34) 1.39) 1.37) Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; Cl =Confidence Interval; — = Not available. * p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh 22 Table 5: Factors associated with Overnutrition among Early Adolescent Girls (2012–2014) 10–14 years (Early adolescent girls) 2012 2013 2014 (n = 2,110) (n = 1,694) (n = 1,832) Variables OR AOR OR AOR OR AOR (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 1.04 — 1.08 — 1.40 — (0.26– (0.27– (0.49– 4.08) 4.27) 4.01) Barisal 0.28 — 0.67 — 0.62 — (0.03– (0.13– (0.19– 2.75) 3.38) 1.99) Dhaka 1.98 — 0.89 — 2.33 — (0.58– (0.25– (0.86– 6.75) 3.16) 0.31) Sylhet 0.84 — 0.38 — 1 (0.19– (0.06– 3.82) 2.34) Chattogram 0.51 — 0.55 — 0.68 — (0.12– (0.14– (0.24– 2.18) 2.10) 1.95) Rangpur 1.24 — 0.53 — 0.68 — (0.32– (0.13– (0.22– 4.74) 2.28) 2.12) Khulna Ref Food security Food 3.10 1.89 1.30 0.61 2.27 1.48 secure (1.64– (0.94– (0.68– (0.29– (1.15– (0.71– 5.84)* 3.79) 2.49) 1.28) 4.50)* 3.08) Mild food 0.64 0.43 Model did Model did 2.19 1.98 insecure (0.08– (0.05– not not (0.59– (0.52– 4.85) 3.36) converge converge 8.17) 7.56) Moderate 2.26 2.45 Model did Model did 1.30 1.07 food (0.73– (0.79– not not (0.16– (0.13– insecure 6.89) 7.64) converge converge 10.46) 8.80) Severe food insecure Ref Area of residence Urban 4.95 3.18 4.03 2.59 3.71 2.73 (2.72– (1.65– (2.02– (1.21– (2.20– (1.54– 9.03)* 6.13)* 8.05)* 5.53)* 6.25)* 4.86)* Rural Ref Household size ≥5 0.85 — 0.52 0.63 0.51 0.54 members (0.47– (0.28– (0.33– (0.31– (0.33– 1.54) 0.97)* 1.19) 0.83)* 0.89) 23 <5 members Ref Religion Non- 1.01 — 2.24 2.22 1.78 1.99 Muslim (0.39– (1.11– (1.06– (0.99– (1.07– 2.56) 4.52)* 4.64)* 3.21)* 3.71)* Muslim Ref Maternal education No formal 1.37 — 0.57 0.63 0.50 0.48 and partial (0.59– (0.15– (0.17– (0.22– (0.21– primary 0.25) 2.13) 2.38) 1.14) 1.11) Secondary 1.53 — 3.76 3.55 2.09 1.61 or more (0.71– (1.32– (1.23– (1.07– (0.81– 3.32) 10.7)* 10.2)* 4.09)* 3.20) Primary Ref completed Dietary diversity <5 food 0.41 0.58 0.63 0.74 0.61 0.87 item (0.23– (0.32– (0.34– (0.38– (0.37– (0.52– 0.74)* 1.08) 1.17) 1.42) 0.99)* 1.46) 5 or more food item Ref Asset quintile Second 0.66 0.58 0.66 0.62 1.08 0.97 (0.19– (0.16– (0.18– (0.17– (0.43– (0.37– 2.36) 2.08) 2.34) 2.24) 2.76) 2.51) Middle 1.21 0.87 1.14 1.05 0.78 0.62 (0.40– (0.29– (0.36– (0.33– (0.28– (0.21– 3.61) 2.66) 3.56) 3.36) 2.22) 1.80) Fourth 4.45 2.18 1.98 1.43 2.49 1.40 (1.77– (0.80– (0.72– (0.48– (1.12– (0.59– 11.1)* 5.95) 5.41) 4.27) 5.54)* 3.33) Highest 2.73 1.27 2.92 2.08 3.32 1.78 (1.00– (0.42– (1.12– (0.71– (1.53– (0.75– 7.45)* 3.83) 7.63)* 6.10) 7.22)* 4.24) Lowest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI=Confidence Interval; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh 24 Factors associated with overnutrition of late adolescent girls. Table 6 shows results from bivariate and multivariable analyses of factors associated with overnutrition among girls in late adolescence. Table 6: Factors associated with Overnutrition among Late Adolescents (2012– 2014) 15–19 years (Late adolescent girls) 2012 2013 2014 (n = 3,693) (n = 2,927) (n = 3,484) Variables OR AOR OR AOR OR AOR (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 0.66 0.69 0.85 0.754 1.05 0.91 (0.39– (0.40– (0.52– (0.46– (0.63– (0.55– 1.13)* 1.18) 1.38) 1.24) 1.74) 1.52) Barisal 0.97 1.05 0.78 0.819 0.86 0.95 (0.56– (0.59– (0.43– (0.44– (0.51– (0.56– 1.71) 1.86) 1.43) 1.51) 1.45) 1.61) Dhaka 1.28 1.19 0.77 0.730 1.17 1.10 (0.81– (0.74– (0.49– (0.46– (0.74– (0.69– 2.02) 1.89) 1.21) 1.16) 1.84) 1.74) Sylhet 0.79 0.99 0.68 0.663 0.46 0.49 (0.46– (0.57– (0.38– (0.37– (0.25– (0.26– 1.35) 1.72) 1.20) 1.19) 0.84)* 0.90)* Chattogram 1.08 1.19 0.80 0.867 1.11 1.18 (0.68– (0.74– (0.51– (0.55– (0.70– (0.73– 1.70) 1.90) 1.24) 1.36) 1.78) 1.89) Rangpur 0.70 0.82 0.44 0.485 0.78 0.84 (0.41– (0.48– (0.26– (0.29– (0.47– (0.50– 1.18) 1.40) 0.75)* 0.82)* 1.30) 1.41) Khulna Ref Food security Food secure 1.89 1.51 2.24 2.03 1.74 1.41 (1.49– (1.16– (1.65– (1.47– (1.28– (1.02– 2.40)* 1.98)* 3.04)* 2.80)* 2.37)* 1.94)* Mild food 1.01 0.95 1.46 1.34 1.03 0.85 insecure (0.59– (0.55– (0.67– (0.61– (0.45– (0.37– 1.74) 1.64) 3.18) 2.95) 2.33) 1.94) Moderate food 1.08 1.09 1.01 1.03 2.46 2.50 insecure (0.63– (0.63– (0.35– (0.36– (1.10– (1.11– 1.86) 1.89) 2.89) 2.96) 5.50)* 5.64)* Severe food Ref insecure Area of residence Urban 2.27 1.85 1.84 1.48 1.65 1.33 (1.73– (1.38– (1.33– (1.05– (1.25– (0.99– 2.98)* 2.48)* 2.55)* 2.08)* 2.19)* 1.80)* Rural Ref Household size 25 ≥5 members 0.94 — 0.94 — 0.89 — (0.75– (0.74– (0.72– 1.18) 1.20) 1.11) <5 members Ref Religion Non-Muslim 1.17 — 1.05 — 0.94 — (0.87– (0.76– (0.70– 1.57) 1.45) 1.27) Muslim Ref Maternal education No formal and 1.01 1.10 1.05 0.71 0.83 — partial primary (0.65– (0.70– (0.66– (0.44– (0.54– 1.58) 1.74) 0.42)* 1.13) 1.27) Secondary or 1.47 1.24 1.08 0.60 1.21 — more (1.02– (0.85– (0.76– (0.86– (0.86– 2.12)* 1.81) 1.54) 1.24) 1.70) Primary Ref completed Dietary diversity <5 food item 0.66 0.89 0.82 1.02 1.00 — (0.53– (0.70– (0.64– (0.79– (0.80– 0.83)* 1.12) 1.04)* 1.32) 1.27) 5 or more food Ref item Asset quintile Second 1.14 1.10 0.74 0.71 1.29 1.23 (0.76– (0.73– (0.47– (0.45– (0.87– (0.82– 1.70) 1.66) 1.15) 1.11)* 1.92) 1.84) Middle 1.16 0.99 0.99 0.86 1.33 1.19 (0.78– (0.65– (0.66– (0.57– (0.90– (0.80– 1.73) 1.50) 1.49) 1.30) 1.96) 1.78) Fourth 1.80 1.21 1.80 1.37 1.66 1.43 (1.24– (0.80– (1.24– (0.93– (1.14– (0.96– 2.61)* 1.84) 2.61)* 2.04) 2.43)* 2.12) Highest 2.24 1.45 1.66 1.20 2.54 2.13 (1.55– (0.95– (1.14– (0.80– (1.78– (1.45– 3.24)* 2.20) 2.40)* 1.78) 3.64)* 3.12) Lowest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI=Confidence Interval; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh Bivariate analysis revealed significant associations (p < 0.05) between overweight and division of residence, food security, area of residence (rural or urban), maternal education, dietary diversity, and wealth quintile in at least one of the years under study (2012 to 2014). Of these, food security, area of residence, and wealth quintile were significant across all three years. Multivariate analysis identified food security and area of residence as predictors of overweight in all three years under study. Compared to late adolescents living in the severe food insecure households, those living in food secure households had higher odds of overweight (AOR 1.51 in 2012; AOR 2.03 in 2013; AOR 26 1.41 in 2014). Likewise, compared to girls in rural areas, late adolescents living in urban areas had higher odds of overnutrition (AOR 1.85 in 2012; AOR 1.48 in 2013; and AOR 1.33 in 2014). 3.4 STATUS OF DIETARY DIVERSITY AMONG ADOLESCENT GIRLS AND FACTORS ASSOCIATED WITH DIETARY DIVERSITY Status of dietary diversity. Table 7 below displays the status of dietary diversity among early and late adolescent girls. Poor dietary diversity was defined as consumption of less than five food groups in the past 24 hours prior to interview. More than half of early adolescent girls fell into the poor dietary diversity group—53.76, 54.64, and 60.37 percent in 2012, 2013, and 2014, respectively. Similarly, the majority of late adolescent girls were also classified as having poor dietary diversity—52.70, 57.39, and 65.34 percent in 2012, 2013, and 2014, respectively. A worrisome increasing trend of poor dietary diversity was apparent in both early and late adolescent age groups. Table 7: Dietary Diversity Status of Early and Late Adolescent Girls (2012–2014) Early adolescents Late adolescents 2012 2013 2014 2012 2013 2014 (n=2,110) (n=1,694) (n=1,832) (n=3,693) (n=2,927) (n=3,484) n (%) Poor dietary 1,365 1,147 1,267 2,288 1,970 2,380 diversity (53.76) (54.64) (60.37) (52.70) (57.39) (65.34) Adequate 745 547 565 1,405 957 1,104 dietary (46.24) (45.36) (39.63) (47.30) (42.61) (34.66) diversity Source: Food Security and Nutrition Surveillance Project, Bangladesh Factors associated with dietary diversity among early adolescent girls. Table 8 shows results from bivariate and multivariable analyses of factors associated with poor dietary diversity in the early adolescent group. Bivariate results show that poor dietary diversity was significantly associated (p < 0.05) with division of residence, food security, area of residence (rural or urban), religion, maternal education, and wealth quintile, in at least one of the years under study (2012 to 2014). Among these variables, division of residence, food security, area of residence, maternal education, and wealth quintile were significantly associated with poor dietary diversity across all three years. In multivariable analysis, only food insecurity and wealth quintile were consistently associated with poor dietary diversity across the three years under study. Compared to those in food secure households, early adolescent girls living in the severe food insecure and mild to moderate food insecure households had higher odds of having poor dietary diversity (AOR 1.88 for severe and 1.56 for mild to moderate in 2012; AOR 2.48 for severe and 1.57 for mild to moderate in 2013; and AOR 2.16 for severe and 1.99 for moderate in 2014). In addition, early adolescent girls living in poorer households had higher odds of poor dietary diversity compared to those living in wealthier households. It is notable that the strength of association between poor dietary diversity and wealth quintile increased with a decrease in wealth of the households. 27 Table 8: Factors associated with Poor Dietary Diversity among Early Adolescent Girls (2012–2014) 10–14 years (Early adolescent girls) 2012 2013 2014 (n=2,110) (n=1,694) (n=1,832) Variables OR AOR OR AOR OR AOR (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 1.41 1.41 1.23 1.42 1.25 1.48 (1.06– (1.04– (0.85– (0.97– (0.87– (1.02– 1.88)* 1.90)* 1.76) 2.08) 1.78) 2.14)* Khulna 1.82 1.71 1.27 1.13 1.29 1.13 (1.23– (1.13– (0.77– (0.68– (0.80– (0.69– 2.71)* 2.58) * 2.09) 1.90) 2.07) 1.85) Barisal 2.42 1.80 0.86 0.72 1.52 1.24 (1.64– (1.20– (0.57– (0.47– (1.07– (0.86– 3.57)* 2.72)* 1.28) 1.09) 2.17)* 1.81) Sylhet 1.71 1.43 0.92 0.85 1.34 1.04 (1.23– (1.01– (0.62– (0.57– (0.95– (0.72– 2.38)* 2.03)* 1.35) 1.27) 1.89) 1.49) Chattogram 1.57 1.29 1.37 1.18 1.07 0.96 (1.21– (0.98– (1.03– (0.88– (0.81– (0.71– 2.04)* 1.70) 1.82)* 1.60) 1.42) 1.30) Rangpur 2.53 2.04 1.29 1.10 1.68 1.39 (1.85– (1.46– (0.93– (0.78– (1.17– (0.96– 3.46)* 2.83)* 1.79) 1.54) 2.39)* 2.02) Dhaka Ref Food security Mild to 1.71 1.56 1.69 1.57 2.29 1.99 moderate food (1.27– (1.15– (1.08– (1.00– (1.42– (1.22– insecurity 2.30)* 2.11)* 2.64)* 2.48)* 3.71)* 3.26)* Severe food 2.57 1.88 2.89 2.48 2.68 2.16 insecurity (2.11– (1.52– (2.28– (1.93– (2.07– (1.64– 3.13)* 2.34)* 3.66)* 3.19)* 3.47)* 2.83)* Food security Ref Area of residence Rural 2.03 1.27 1.43 1.12 2.11 1.66 (1.55– (0.94– (1.03– (0.78– (1.59– (1.22– 2.66)* 1.70) 2.00)* 1.60) 2.79)* 2.26)* Urban Ref Household size <5 members 1.04 — 0.92 — 1.15 — (0.86– (0.74– (0.92– 1.27) 1.14) 1.43) ≥5 members Ref Religion Muslim 0.70 0.79 0.94 — 1.06 — 28 (0.52– (0.57– (0.70– (0.80– 0.96)* 1.09) 1.27) 1.41) Others Ref Maternal education No formal and 1.29 0.94 1.53 1.22 1.71 1.48 partial primary (1.05– (0.75– (1.22– (0.96– (1.37– (1.17– 1.57)* 1.17) 1.92)* 1.55) 2.14)* 1.87)* Primary 1.38 1.21 1.25 1.14 1.39 1.18 completed (1.06– (0.92– (0.94– (0.85– (1.05– (0.89– 1.79)* 1.60) 1.65) 1.53) 1.83)* 1.58) Secondary or Ref more Asset quintile Lowest 4.81 3.34 2.29 1.70 2.80 1.82 (3.52– (2.36– (1.64– (1.19– (2.03– (1.28– 6.59)* 4.73)* 3.20)* 2.43)* 3.85)* 2.59)* Second 2.76 1.98 1.97 1.48 2.19 1.56 (2.06– (1.44– (1.42– (1.05– (1.60– (1.12– 3.69)* 2.71)* 2.73)* 2.09)* 2.99)* 2.19)* Middle 2.06 1.67 1.43 1.18 1.94 1.49 (1.55– (1.23– (1.03– (0.84– (1.41– (1.07– 2.76)* 2.27)* 1.98)* 1.65) 2.66)* 2.08)* Fourth 1.53 1.47 0.88 0.91 1.26 1.18 (1.14– (1.08– (0.64– (0.66– (0.93– (0.86– 2.07)* 2.01)* 1.20) 1.25) 1.70)* 1.60) Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI=Confidence Interval ; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh Factors associated with dietary diversity among late adolescent girls. Table 9 shows results from bivariate and multivariable analyses of factors associated with poor dietary diversity among late adolescent girls. Bivariate results indicate that poor dietary diversity was significantly associated (p < 0.05) with division of residence, food insecurity, area of residence (rural or urban), household size, religion, maternal education, and wealth quintile in at least one of the years under study (2012 to 2014). Among these variables, division of residence, food security, area of residence, household size, maternal education, and wealth quintile were significantly associated with poor dietary diversity during all three years. Multivariable analysis identified consistent associations between poor dietary diversity and division of residence, food security, household size, maternal education, and wealth quintiles over the study period. In terms of divisional differences, compared to Dhaka division, late adolescents in Khulna division had higher odds of having poor dietary diversity (AOR 1.90 in 2012; 1.47 in 2013; and 1.44 in 2014). Compared to those in food secure households, late adolescent girls living in severe food insecure or mild to moderate food insecure households had higher odds of having poor dietary diversity (AOR of 2.31 and 1.99, respectively, in 2012; AOR of 2.03 and 2.51, respectively, in 2013; AOR of 2.58 and 1.88, respectively, in 2014). Compared to girls whose households had ≥ 5 members, late adolescents living in the households with <5 members had higher odds of having poor dietary diversity. Girls whose mothers 29 did not have any formal education or had partial primary education had higher odds of having poor dietary diversity compared to those whose mothers had secondary or higher education. Finally, late adolescent girls living in poorer households had higher odds of poor dietary diversity compared to those living in wealthier households. Moreover, as seen in the analysis of early adolescents, the strength of association between poor dietary diversity and wealth quintile increased with a decrease in household wealth. Table 9: Factors associated with Poor Dietary Diversity among Late Adolescents (2012–2014) 15–19 years (Late adolescent girls) 2012 2013 2014 (n=3,693) (n=2,927) (n=3,484) Variables OR (95% AOR OR (95% AOR OR (95% AOR CI) (95% CI) CI) (95% CI) CI) (95% CI) Division Rajshahi 2.39 2.33 1.60 1.74 1.13 1.26 (1.88– (1.81– (1.21– (1.30– (0.88– (0.97– 3.03)* 2.99)* 2.11)* 2.32)* 1.45) 1.63) Khulna 1.97 1.90 1.54 1.47 1.49 1.44 (1.47– (1.39– (1.10– (1.03– (1.08– (1.03– 2.64)* 2.59)* 2.17)* 2.09)* 2.06)* 2.01)* Barisal 1.49 1.40 0.84 0.73 1.04 0.93 (1.13– (1.05– (0.60– (0.51– (0.81– (0.71– 1.96)* 1.88)* 1.18) 1.04) 1.33) 1.20) Sylhet 1.93 1.46 0.81 0.69 1.54 1.23 (1.51– (1.12– (0.60– (0.50– (1.17– (0.93– 2.47)* 1.90)* 1.10) 0.95)* 2.02)* 1.64) Chattogram 1.23 0.97 0.84 0.70 1.07 0.92 (1.01– (0.78– (0.67– (0.56– (0.87– (0.74– 1.49)* 1.19) 1.04)* 0.88)* 1.33) 1.15) Rangpur 2.40 1.84 1.76 1.46 1.31 1.04 (1.90– (1.43– (1.35– (1.11– (1.03– (0.81– 3.04)* 2.37)* 2.29)* 1.93)* 1.66)* 1.33) Dhaka Ref Food security Mild to 2.35 1.99 2.88 2.51 2.04 1.88 moderate food (1.88– (1.58– (1.89– (1.64– (1.38– (1.26– insecurity 2.94)* 2.52)* 4.38)* 3.86)* 3.03)* 2.82)* Severe food 2.97 2.31 2.25 2.03 3.18 2.58 insecurity (2.57– (1.97– (1.87– (1.67– (2.56– (2.06– 3.44)* 2.72)* 2.71)* 2.48)* 3.95)* 3.22)* Food security Ref Area of residence Rural 1.93 1.42 1.59 1.38 1.46 1.16 (1.58– (1.14– (1.25– (1.07– (1.19– (0.93– 2.34)* 1.76)* 2.02)* 1.79)* 1.79)* 1.45) Urban Ref Household size 30 <5 members 1.40 1.21 1.42 1.19 1.24 1.17 (1.22– (1.04– (1.21– (1.01– (1.07– (1.00– 1.60)* 1.41)* 1.66)* 1.41)* 1.43)* 1.36)* ≥5 members Ref Religion Muslim 0.82 0.82 0.97 — 0.91 — (0.68– (0.67– (0.78– (0.74– 0.99)* 1.00)* 1.20) 1.11) Non-Muslim Ref Maternal education No formal and 2.43 1.75 1.92 1.48 2.06 1.58 partial primary (2.01– (1.42– (1.54– (1.16– (1.67– (1.27– 2.94)* 2.17)* 2.40)* 1.87)* 2.54)* 1.98)* Primary 1.50 1.21 1.98 1.62 1.50 1.24 completed (1.22– (0.97– (1.53– (1.24– (1.19– (0.98– 1.84)* 1.51) 2.55)* 2.12)* 1.88)* 1.58) Secondary or Ref more Asset quintile Lowest 3.51 1.83 2.45 1.81 2.57 1.83 (2.80– (1.40– (1.91– (1.38– (2.03– (1.42– 4.41)* 2.37)* 3.16)* 2.38)* 3.24)* 2.37)* Second 2.80 1.63 2.32 1.73 2.61 1.97 (2.25– (1.28– (1.81– (1.33– (2.07– (1.54– 3.47)* 2.07)* 2.97)* 2.26)* 3.29)* 2.52) * Middle 2.25 1.52 1.89 1.61 1.80 1.56 (1.83– (1.21– (1.49– (1.25– (1.45– (1.24– 2.78)* 1.91)* 2.40)* 2.06)* 2.24)* 1.95) * Fourth 1.36 1.24 1.14 1.19 1.31 1.21 (1.11– (1.00– (0.90– (0.94– (1.06– (0.97– 1.67)* 1.53)* 1.43) 1.51) 1.61)* 1.51) Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI=Confidence Interval; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh 31 3.5 STATUS OF FOOD SECURITY AMONG ADOLESCENT GIRLS AND FACTORS ASSOCIATED WITH FOOD SECURITY Status of food security. Table 10 displays the status of food security, over the previous four weeks, among the households of early and late adolescent girls, classified here as food secure, mild food insecure, moderate food insecure, and severe food insecure. Estimates for the four food security groups were generated by year. It is evident from Table 10 that food insecurity in the households of both early and late adolescents improved over the three years under investigation. The severe food insecurity category among early adolescents declined in size from 54 percent in 2012 to 27 percent in 2014, and from 44 to 21 percent in the households of late adolescents. Likewise, the proportion of households with food security increased for both age groups over the three years under study. Table 10: Status of Food Security of Early and Late Adolescent Girls (2012–2014) Early adolescent girls Late adolescent girls 2012 2013 2014 2012 2013 2014 (n=2,110) (n=1,694) (n=1,832) (n=3,693) (n=2,927) (n=3,484) Food n security (%) type Food 709 966 1,228 1,612 1,899 2,610 secure (33.6) (57.02) (67.03) (43.65) (64.88) (74.91) Mild food 120 62 69 229 91 97 insecure (5.69) (3.66) (3.77) (6.2) (3.11) (2.78) Moderate 138 41 39 212 64 52 food (6.54) (2.42) (2.13) (5.74) (2.19) (1.49) insecure Severe 1,143 625 496 1640 873 725 food (54.17) (36.89) (27.07) (44.41) (29.83) (20.81) insecure Source: Food Security and Nutrition Surveillance Project, Bangladesh Factors associated with food insecurity among early adolescent girls. Table 11 shows results from bivariate and multivariable analyses of factors associated with food security among early adolescent girls. Bivariate results indicate significant associations (p < 0.05) between food insecurity and division of residence, area of residence (rural or urban), maternal education, dietary diversity, and wealth quintile over all three years under study. In multivariable analysis, only maternal education, dietary diversity, and wealth quintile were consistently associated with food insecurity over the years. Early adolescent girls living in poorer households had higher odds of food insecurity compared to those living in wealthier households. Moreover, the strength of association between food insecurity and wealth quintile was observed to increase with a decrease in household wealth. Early adolescents whose mothers had no formal or partial primary education had higher odds of food insecurity compared to girls whose mothers had secondary or more education. Finally, greater dietary diversity among early adolescent girls—consumption of five or more food groups per day—was found to decrease the odds of food insecurity. 32 Table 11: Factors associated with Food Insecurity of Early Adolescent Girls (2012– 2014) 10–14 years (Early adolescent girls) 2012 2013 2014 (n=2,110) (n=1,694) (n=1,832) Variables OR (95% AOR OR (95% AOR OR (95% AOR CI) (95% CI) CI) (95% CI) CI) (95% CI) Division Rajshahi 0.99 0.90 0.64 0.71 0.90 1.61 (0.74– (0.65– (0.44– (0.48– (0.60– (1.05– 1.32) 1.23) 0.92) 1.06) 1.35) 0.69) Khulna 1.17 1.07 1.56 1.52 2.04 2.19 (0.79– (0.70– (0.99– (0.92– (1.28– (1.34– 1.72) 1.63) 2.48) 2.51) 3.23)* 3.60)* Barisal 1.76 1.16 1.78 1.93 2.47 2.28 (1.19– (0.76– (1.20– (1.26– (1.75– (1.58– .59)* 1.78) 2.64)* 2.96)* 3.49)* 3.31)* Sylhet 1.66 1.27 1.29 1.17 2.33 1.90 (1.18– (0.87– (0.88– (0.77– (1.66– (1.32– 2.34)* 1.85) 1.87) 1.75) 3.29)* 2.74)* Chattogram 1.43 1.08 1.43 1.19 1.84 1.58 (1.10– (0.80– (1.10– (0.88– (1.36– (1.15– 1.88) 1.44) 1.88) 1.59) 2.47)* 2.18)* Rangpur 1.71 1.13 1.39 1.10 2.34 2.04 (1.25– (0.80– (1.02– (0.79– (1.67– (1.42– 2.32)* 1.58) 1.88) 1.54) 3.28)* 2.92)* Dhaka Ref Area of residence Rural 2.22 1.20 1.94 1.00 1.67 0.72 (1.69– (0.88– (1.36– (0.68– 1.21– (0.50– 2.91)* 1.62) 2.76)* 1.49) 2.30)* 1.04) Urban Ref Household size <5 members 1.03 — 0.72 0.76 1.00 — (0.90– (0.59– (0.61– (0.81– 1.17) 0.89)* 0.96) 1.24) ≥5 members Ref Religion Muslim 0.93 — 0.92 — 0.81 — (0.69– (0.70– (0.62– 1.25) 1.21) 1.07) Non-Muslim Ref Maternal education No formal and 2.29 1.64 2.26 1.87 1.94 1.59 partial primary (1.86– (1.31– (1.82– (1.48– (1.55– (1.25– 2.80)* 2.06)* 2.81)* 2.36)* 2.42)* 2.03)* Primary 1.37 1.16 1.71 1.64 1.59 1.36 completed 33 (1.06– (0.88– (1.30– (1.22– (1.20– (1.01– 1.77) 1.53) 2.25)* 2.20)* 2.09)* 1.82) Secondary or Ref more Dietary diversity 5 or more 0.42 0.56 0.38 0.44 0.38 0.47 food items (0.35– (0.45– (0.30– (0.35– (0.30– (0.37– 0.51)* 0.68)* 0.47)* 0.55)* 0.49)* 0.60)* <5 food items Ref Asset quintile Lowest 9.55 6.42 4.88 3.86 7.52 5.93 (6.81– (4.48– (3.49– (2.70– (5.19– (3.98– 13.39)* 9.19)* 6.82)* 5.50)* 10.8)* 8.83)* Second 5.65 4.35 4.54 3.89 4.76 4.05 (4.15– (3.14– (3.25– (2.74– (3.27– (2.73– 7.71)* 6.02)* 6.34)* 5.54)* 6.92)* 6.03)* Middle 3.59 2.92 2.89 2.57 3.63 3.30 (2.66– (2.15– (2.05– (1.79– (2.47– (2.21– 4.83)* 3.98)* 4.07)* 3.67)* 5.33)* 4.94)* Fourth 1.70 1.55 1.14 1.11 1.80 1.74 (1.25– (1.14– (0.80– (0.77– (1.20– (1.15– 2.29)* 2.12) 1.63) 1.61) 2.69)* 2.62) Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI=Confidence Interval; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh Factors associated with food insecurity among late adolescent girls. Table 12 shows results from bivariate and multivariable analyses of factors associated with food security among girls in the late adolescent age group. Bivariate results indicate significant associations (p < 0.05) between food security and division of residence, area of residence (rural or urban), maternal education, dietary diversity, and wealth quintile during all the years under study (2012 to 2014). In multivariable analysis, only maternal education, dietary diversity, and wealth quintiles were consistently associated with food insecurity over the period of study. Compared to late adolescent girls consuming less diverse foods, those consuming more diverse foods (five or more food groups in the last 24 hours) had lower odds of having food insecurity. Late adolescents whose mothers did not have any formal education or only had partial primary education had higher odds of experiencing food insecurity compared to girls whose mothers had secondary or higher education. Finally, late adolescents living in poorer households had higher odds of food insecurity compared to those living in wealthier households. The strength of association between food insecurity and wealth quintile was observed to increase with a decrease in household wealth. Table 12: Factors associated with Food Security of Late Adolescents (2012–2014) 15–19 years (Late adolescent girls) 2012 2013 2014 (n=2,110) (n=1,694) (n=1,832) 34 Variables OR (95% AOR OR AOR OR AOR CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Division Rajshahi 1.13 0.87 0.81 0.75 0.71 0.77 (0.90– (0.68– (0.61– (0.55– (0.51– (0.55– 1.41) 1.11) 1.08) 1.02) 0.98) 1.08) Khulna 1.43 1.24 1.53 1.45 1.51 1.48 (1.08– (0.91– (1.11– (1.03– (1.08– (1.04– 1.89) 1.68) 2.11) 2.04) 2.11) 2.11) Barisal 1.06 0.75 1.96 2.10 1.61 1.50 (0.81– (0.56– (1.39– (1.46– (1.22– (1.12– 1.39) 1.02) 2.76)* 3.01)* 2.12)* 2.01) Sylhet 2.06 1.43 1.37 1.23 2.02 1.50 (1.62– (1.08– (1.00– (0.88– (1.53– (1.12– 2.64)* 1.87) 1.89) 1.74) 2.66)* 2.01) Chattogram 1.47 1.05 2.08 2.05 1.41 1.11 (1.21– (0.84– (1.66– (1.62– (1.11– (0.86– 1.78)* 1.30) 2.60)* 2.60)* 1.80)* 1.44) Rangpur 2.03 1.12 1.56 1.21 2.29 1.98 (1.66– (0.90– (1.21– (0.92– (1.79– (1.52– 2.47)* 1.40) 2.01)* 1.58) 2.94)* 2.5)* Dhaka Ref Area of residence Rural 2.03 1.12 1.46 0.88 1.46 0.82 (1.66– (0.90– (1.13– (0.66– (1.14– (0.62– 2.47)* 1.40) 1.90)* 1.17) 1.87)* 1.07) Urban Ref Household size <5 members 1.11 1.15 1.01 1.08 (0.91– (0.99– (0.85– (0.92– 1.34) 1.34) 1.19) 1.26) ≥5 members Ref Religion Muslim 0.95 0.79 0.96 0.95 (0.79– (0.64– (0.76– (0.77– 1.14) 0.97) 1.21) 1.18) Non-Muslim Ref Maternal education No formal and 3.10 1.61 2.71 2.01 2.20 1.61 partial primary (2.56– (1.30– (2.23– (1.62– (1.82– (1.31– 3.74)* 1.98)* 3.30)* 2.49)* 2.66)* 1.98)* Primary 1.76 1.14 2.04 1.71 1.54 1.18 completed (1.44– (0.92– (1.63– (1.34– (1.22– (0.93– 2.15)* 1.43) 2.55)* 2.17)* 1.93)* 1.51) Secondary or Ref more Dietary diversity 5 or more food 0.35 0.45 0.43 0.48 0.34 0.41 item (0.31– (0.39– (0.36– (0.40– (0.28– (0.34– 0.41)* 0.53)* 0.51)* 0.58)* 0.42)* 0.50)* 35 <5 food items Ref Asset quintile Lowest 10.50 7.51 4.73 3.40 6.02 4.41 (8.17– (5.72– (3.63– (2.57– (4.51– (3.23– 13.49)* 9.84)* 6.16)* 4.51)* 8.04)* 6.02)* Second 5.67 4.46 3.81 3.01 5.32 4.07 (4.52– (3.51– (2.93– (2.27– (3.98– (3.00– 7.10)* 5.67)* 4.97)* 3.98)* 7.10)* 5.51)* Middle 3.85 3.21 2.63 2.16 2.77 2.27 (3.09– (2.56– (2.02– (1.64– (2.05– (1.67– 4.78)* 4.03)* 3.43)* 2.84)* 3.73)* 3.09)* Fourth 1.80 1.72 1.57 1.49 1.95 1.77 (1.44– (1.38– (1.19– (1.12– (1.43– (1.29– 2.23)* 2.15)* 2.07)* 1.98 2.66)* 2.44)* Highest Ref Note: AOR = Adjusted Odds Ratio; OR = Odds Ratio; CI= Confidence Interval; — = Not available. *p < 0.05. Source: Food Security and Nutrition Surveillance Project, Bangladesh 3.6 GROWTH DYNAMICS OF ADOLESCENT GIRLS To identify optimal moments to avert potential faltering, this section considers growth curves during the adolescent period. Available data from cohort studies that follow adolescents through this dynamic period of growth are relatively rare and highly specific to a particular area and population. To overcome these limitations, a pseudo cohort was constructed based on pooled repeated cross-sections of FSNSP data over the period 2011 to 2014, yielding a dataset of 4,500 adolescents. Using this pseudo cohort, changes in linear growth over the adolescent period were examined using the measure height-for-age Z scores (HAZ score), and stunting dynamics were indicated by applying the cutoff HAZ<-2 SD, which proxies chronic undernutrition. Dynamics in undernutrition and overnutrition were also explored using BMI-for-age Z score (BAZ score). Acute malnutrition over the adolescent period was indicated by using the cutoff BAZ <-2 SD. Dynamics of linear growth and stunting over the adolescent period. The graphs in Figure 11 depict the dynamics of linear growth over the adolescent years, relative to healthy standards. Height gain occurs steadily from age 10 to 14 and then levels off as full adult stature is attained. When compared to a healthy norm, a steep decline in HAZ score is apparent, which is most dramatic in the first years of the early adolescent period. The reduced slope after 15 years of age indicates a slowing in the rate of growth faltering, with predicted HAZ scores leveling off at -1.6 to -1.8 below the norm. 36 Figure 11: Dynamics of Linear Growth and Height-for-Age Z-Score of Adolescent Girls Source: Food Security and Nutrition Surveillance Project, Bangladesh Table 13 illustrates the pace of linear growth by identifying absolute differences in HAZ scores from one year to the next. It also illustrates the relative percentage change associated with each progressive year of life. Results suggest that the average rate of growth faltering is -22 percent in the early adolescent period from age 10 to 14, compared to -4 percent in the later adolescent period from 14 to 19 years of age. Table 13: Absolute and Relative Changes in Growth with Increasing Age among Adolescent Girls Source: Food Security and Nutrition Surveillance Project, Bangladesh 37 Figure 12 compares the linear growth of girls living in households from the lowest and highest wealth quintiles. While the slope and shape of the curve are similar, the curve for girls from the poorest households is a full 0.5 points below that of girls from the wealthiest households. Figure 12: Linear Growth (Height-for-Age Z-Score) comparing Adolescent Girls from the Lowest and Highest Wealth Quintiles Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 13 shows a steep gradient in the rate of stunting in the early stages of adolescence, especially between 10.0 and 11.5 years of age. The slope is less steep in the later adolescent period, although the overall level of stunting is much higher. 38 Figure 13: Dynamics of Stunting (HAZ<-2 SD) with Increasing Age among Adolescent Girls Source: Food Security and Nutrition Surveillance Project, Bangladesh Dynamics of BMI and undernutrition over the adolescent period. Figure 14 illustrates a steady increase in predicted BMI over each year of adolescence and relative to the healthy norms as expressed as BMI-for-age Z score (BAZ score). In the early years, BAZ score is well below the norm, improving slowly and then very rapidly from age 12 to 14, before leveling off and then increasing again at the end of the late adolescent period. 39 Figure 14: BMI by Age and BMI-for-Age Z Score (BAZ) over the Adolescent Period Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 15 depicts the dynamics of undernutrition (BAZ <-2 SD) over the adolescent period and differences in the BAZ score comparing adolescent girls from the poorest and wealthiest quintiles. Rates of undernutrition are initially high at the start of early adolescence, decline between ages 12 to 13, then level off and decline slowly through the late adolescent period. The curves for BAZ score comparing girls from the poorest and wealthiest households are quite different in terms of level and shape. The curve for BAZ score among girls in the poorest quintile is consistently below that of girls from wealthier households, with the lowest levels apparent in the first years of the adolescent period. Unlike the wealthier group, however, the steep slope occuring between ages 12 and 15 indicates rapid gains in BAZ score through the end of the early adolescent period. 40 Figure 15: Undernutrition (BAZ <-2 SD) Dynamics over the Adolescent Period and Differences in BAZ Score comparing Girls from the Poorest and Wealthiest Quintiles Source: Food Security and Nutrition Surveillance Project, Bangladesh 3.7 POLICIES AND STRATEGIES SUPPORTING ADOLESCENT NUTRITION AND WELL- BEING The Bangladesh Constitution and different national policies, plans, and strategies include statements supporting adolescent nutrition and the rights of women and girls to gender-responsive health services. Article 18 of the Constitution affirms that raising the level of nutrition is a duty of the state, and Article 19 ensures equal opportunity to all citizens, including the participation of women in all spheres of national life. Article 28 (1) states, “(The) State shall not discriminate against any citizen on grounds of religion, race, caste, sex, and place of birth,” and Article 28 (2) asserts the particular rights of women, which we consider inclusive of adolescent girls: “Women shall have equal rights with men in all spheres of the state and public life.” Article 28 (4) further clarifies, “Nothing in this article shall prevent the state from making special provision in favor of women or children or for the advancement of any backward section of citizens.” In addition to constitutional provisions, existing laws have been amended and new legislation has been ratified to prevent the abuse of woman and children, such as the Dowry Prevention Act, the Prevention of Marriage of Minor Girls, and the Women and Children Repression Prevention Act, 2000. Women’s abuse prevention cells and rehabilitation centers for abused women have been established to provide legal assistance and counseling. Moreover, district and session judges are allocated funds to cover the cost of legal fees. The Domestic Violence Prevention and Protection Act of 2010 supports the equal rights of women and children, as prescribed in the Constitution of Bangladesh, and ensures the protection of women and children from family violence. In addition, the Executive Magistrate has the authority to implement Section 509 of the Bangladesh Penal Code in the schedule of the Mobile Court Act to resist and prevent “eve teasing” and other forms of sexual harassment of girls and women. 41 On a global front, Bangladesh is signatory to the 1979 UN Charter on the Prevention of all Forms of Discrimination Against Women (CEDAW) and to the Child Rights Charter of 1989. In support of these commitments, the national Parliament accepted provisions for mothers to give citizenship to their children, and relevant amendments were made in 2009 to the Citizenship Act. The government of Bangladesh has also reiterated global commitments through the Perspective Plan of Bangladesh 2010–2021, which states that gender equality at all levels of education will be ensured, and all students enrolled in primary, secondary, and tertiary levels will have access to gender-responsive health, nutrition, water and sanitation, sociocultural development, and greater participation in sports, thus enabling a fruitful learning and living environment. Likewise, the Seventh Five Year Plan (2017–2020) of the government of Bangladesh looks forward to improving nutrition to children and women, along with gender- and adolescent-friendly services, which includes information for adolescents to protect themselves from health hazards. The plan also aims to facilitate greater school completion rates of girls from secondary schools and increased enrollment of girls in higher education. The National Women Development Policy of 2011 envisages a society where men and women will have equal opportunities and will enjoy all fundamental rights on an equal basis. The main objectives of the policy include improvement in nutrition (# 1) and gender equity in health services (# 9). Further supporting the goal of gender equality is the 2012 National Population Policy, which highlights commitments to gender equity and women’s empowerment and to strengthening activities to eliminate gender discrimination in family planning and maternal and child health care programs (# 4.4). Major implementation strategies include the adolescent welfare program (# 5.5). With respect to nutrition-specific policy, the National Nutrition Policy (Strategy 6.1.2.4) aims to ensure the availability of nutritious and safe food for growth and development of adolescent boys and girls. It also asserts the goal of stopping child marriage to enable healthy and productive future generations. Strategy 6.3 (direct nutrition interventions) emphasizes improved nutrition knowledge among adolescent girls and women at the family level, with the aim of healthy nutrition behavior change. Strategy 6.3.7 of this policy ensures sufficient diversified food intake of adolescent boys and girls so that their physical growth will be optimized and they attain a healthy adult height and weight. Strategy 6.4.3 (indirect nutrition interventions) emphasizes actions to improve women’s literacy rate, and empowerment and employment opportunities, and to reduce the rate of pregnancy before age 20. The Program Implementation Plan (PIP) of the Fourth Health, Population, and Nutrition Sector Investment Program (HPNSIP) 2017–2022 endorses the Adolescent Reproductive Health Strategy 2017–2030 and the related National Plan of Action of 2011–17. It recognizes that insufficient progress has been made on the prevention of child marriage, early pregnancy, unsafe abortion, violence prevention, and the promotion of healthy reproductive behaviors among adolescents. The PIP also mentions that combined action by all sectors is needed to ensure that legal, educational, social, and cultural issues that are supportive of improved adolescent nutrition are addressed. The National Nutrition Service (NNS) Operational Plan (OP) includes provisions for Behavior Change Communication (BCC) sessions targeting adolescent girls to provide them with necessary knowledge on reproductive and nutritional health through individual and group counseling. Through this OP, adolescent girls receive deworming and iron-folate tablets as they come into contact with the health system. The 42 mainstreaming of nutrition through the health system aims to ensure the delivery of nutrition services through Reproductive, Maternal, Neonatal, Adolescent and Child Health (RMNACH) platforms, and to improve their functional integration. The mainstreaming nutrition program also includes an effective BCC strategy and materials specific to nutrition to increase demand for nutrition services. Finally, the Bangladesh Essential Service Package (ESP) of the HPNSIP includes adolescent-friendly health services, including counseling on issues ranging from safe sexual behavior to substance abuse, family planning information and services, nutrition, screening and management of sexually transmitted infections (STIs), trafficking, and mental health. CHAPTER 4: DISCUSSION Trends and variations of urgent policy relevance were revealed by this analysis of the nutrition and food security status of adolescent girls and their families. Some trends were encouraging, such as positive changes in the household environments in which girls are growing up. Most notable was a secular decline in severe food insecurity, and the large proportion of girls whose mothers were educated at the secondary level or higher. Similar improvements in food security have been observed in a variety of national and area-specific studies. The high prevalence of mothers of adolescent girls reporting at least secondary education relative to levels of maternal education only a decade ago, also reflects documented improvements at a national level attributed to substantial government investments in girl-focused secondary school subsidies. On a more worrisome note, results indicated a declining yet unacceptably high prevalence of moderate to severe thinness among early adolescent girls, and a trend toward increasing rates of overweight and obesity among older adolescents in an interval of only three years. In late adolescence, high rates of stunting were also observed. Striking differences were apparent between girls living in rural and urban areas. In both early and late adolescent groups, moderate to severe thinness was more prevalent in rural areas, whereas rates of overweight and obesity were greater in urban areas. In early adolescence, stunting was more prevalent in rural versus urban areas (13 vs. 8 percent in 2015); however, this pattern was not apparent in late adolescence. Although no similar nationally representative estimates of adolescent nutritional status in Bangladesh exist, results of several smaller-scale studies provide a point of comparison. In rural northwest Bangladesh, low BMI (< 18.5 kg/m2) was documented in 36 percent of pregnant adolescents age 15 to 19 years, and 44 percent were short for their age (HAZ <-2 or height <150.1 cm if 19 years of age) (Mridha et al., 2018). The rural- focused Bangladesh Integrated Household Survey (BIHS) identified 17 percent of adolescent girls as thin (BMIZ <-2) and 4 percent as overweight or obese (Leroy et al., 2018). Lower measures of mean height (145 cm) and weight (37 kg) recorded by BIHS relative to this analysis (151 cm and 45 kg, respectively, in 2014) are probably due to its rural focus and the smaller number of older adolescents in its sample. Results from a baseline survey of almost 5,000 unmarried adolescent girls age 13 to 18 conducted by the national nutrition program identified 26 percent of adolescents as thin (BMI-for-age <15th percentile) and 0.3 percent obese (BMI-for-age >95th percentile) (Alam et al., 2010). Only one study was found that compared the nutritional status of rural and urban adolescents age 14 to 17 years. Consistent with our findings, rates of severe thinness were much 43 higher in rural versus urban areas (22 vs. 10 percent), and among girls in their younger adolescent years (Akhter et al., 2013). Of similar concern were findings indicating an increase in the proportion of early and late adolescents with poor dietary diversity over the period 2012 to 2014. Given gains in household food security and female education seen at the national level, the known positive associations between food security and dietary diversity, and recent evidence indicating the same for maternal schooling (Sinharoy et al., 2018), this finding warrants further research and policy attention. It also raises concerns about the limitations of standard definitions of food security that fail to take dietary diversity into account. The United Nation’s Committee of World Food Security defines food security as the condition in which all people, at all times, have physical, social, and economic access to sufficient safe and nutritious food that meets their dietary needs and food preferences for an active and health life (http://www.ifpri.org/topic/food-security). The diversity and nutrient contents of food are not highlighted in this definition, when both are critical determinants of diet quality and the adequacy of micronutrient intake (Nguyen et al., 2018). The counterintuitive finding that dietary diversity among adolescent girls has declined, despite improvements in food security and women’s education, may point to definitional issues or to other underlying factors that need to be identified and addressed. Multivariable analysis of the determinants of the nutritional status of girls in early adolescence revealed an association between undernutrition (moderate to severe thinness) and low levels of maternal education, and between overnutrition (overweight and obesity) and urban residence and greater household wealth. Poor dietary diversity among early adolescent girls was associated with household food insecurity and greater household poverty (lower wealth quintiles), while household food insecurity was associated with lower levels of maternal education, poor dietary diversity, and greater poverty in all years under study. In late adolescence, food insecurity was associated with undernutrition, while food security and urban residence predicted overnutrition. Poor dietary diversity was associated with low maternal education, smaller household size (<5 members), food insecurity, and rural residence. Conversely, household food insecurity was associated with low dietary diversity, maternal education, and wealth in each year under analysis. Only two studies were found that explore factors associated with undernutrition in Bangladesh, and provide a basis for comparison with findings from this analysis. One of these was conducted in rural northwest Bangladesh. The study identified low household wealth as a predictor of low BMI among adolescents, and poor household wealth and poor education as predictors of short stature (Mridha et al., 2018). Another study by Leroy et al., found a positive association between greater household wealth and higher BMI Z scores among adolescents, but no relationship with education and women’s empowerment (Leroy et al., 2018). At the time of this report no published studies examining the predictors of adolescent overnutrition in Bangladesh were available. In light of gaps in the literature on adolescent nutrition, an important contribution of this study is its nuanced examination of variations in the predictors of malnutrition (both overnutrition and undernutrition) between younger and older adolescent girls by division and rural-urban area of residence. It illustrates the value of contextualized analyses sensitive to divisional and rural-urban and age group 44 differences, and by extension, the importance of a comprehensive approach to improving adolescent nutrition that involves simultaneous investments in its multiple determinants, including maternal education, food security, household wealth, dietary diversity, etc. Examination of the growth dynamics of adolescent girls using a synthetic cohort approach, revealed a number of concerning patterns. As expected, height increased, more or less linearly until the age of 14, but was not on par with the healthy norm. Just as growth faltering is known to concentrate in early childhood in Bangladesh (6 to 24 months), results indicated that growth faltering is greatest in early adolescence. Moreover, moving from early to late adolescence, the rate of stunting more than doubles (from 11 to 22 percent), and BMI rises. While these growth patterns are similar comparing girls from the highest and lowest wealth quintiles, a stark gap is apparent with the trend line of poorer girls, which is consistently below that of the wealthiest. Analysis also revealed that adult height is attained relatively early, leading to increased rates of stunting in older adolescents. At the same time, relatively higher gains in weight compared to height are reflected in increasing BMI. These findings are of public health concern, as upward trends in BMI over the adolescent period have the potential to exacerbate already increasing rates of overweight and obesity in the country’s population at large. A survey of the policy landscape in Bangladesh revealed a host of policies and strategies that can be harnessed to improve the nutritional status of adolescent girls. The country’s constitution enshrines an obligation to improve the nutrition of all citizens. In addition, a variety of national health-related policies and strategies, such as the Adolescent Health Strategy and the Second National Plan of Action on Nutrition II, are available to guide the implementation of nationwide efforts to improve adolescent nutrition. However, implementation of these policies and strategies remain a challenge in the absence of firm political commitment and a country investment plan to improve adolescent nutrition. Some limitations to analysis must be acknowledged. One relates to sample composition and the disproportionately large number of girls in the late adolescent age group. Most probably this was a result of oversampling when more than one adolescent was present in sample households, and may explain differences in estimates provided by this study compared to the broader literature. It should also be noted that sample size was insufficient in certain divisions, making it difficult to generate reliable estimates of all nutrition-related variables under study. Despite these limitations, reasonably representative national and geographical estimates were produced for both early and late adolescent girls, and as such, specific guidance on actions to support adolescent nutrition can be generated. CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS Adolescence is a period of rapid physical, cognitive, and psychosocial growth, and a window of opportunity for lifelong health and well-being. Adequate nutrient intake during the early adolescent period is foundational to achieving full developmental potential and optimizing later adult health and productivity. In terms of physical growth, it is also a time during which gains equivalent to 50 percent of adult weight, 45 percent of skeletal mass, and 15 percent of adult height should occur. Results from this first ever nationally representative analysis of adolescent girls’ nutrition in Bangladesh indicate a situation that is improving, but of continued concern. 45 Almost 28 percent of girls in early adolescence remain severely or moderately thin, although substantial improvements have occurred since 2012. This improvement is also reflected in measures of stunting, which decreased from 17 to 11 percent among early adolescents from 2012 to 2014—with noteworthy improvements in urban areas. However, yet mirroring global trends, a steady emergence of overweight and obesity among older adolescent girls is observed, climbing from 13 to 23 percent from 2012 to 2014, and even more dramatically in urban areas (Leroy et al., 2018). Levels of stunting also worsened in the older adolescent age group, increasing from 38 to 42 percent over the same period. This situation analysis of the nutrition and food security status of adolescent girls and their families revealed trends and geographic variations. Most encouraging was evidence of secular declines in severe food insecurity across the country. At the same time, results indicated a declining yet unacceptably high prevalence of moderate to severe thinness among younger adolescent girls and a trend toward increasing rates of overweight and obesity among late adolescents. Moreover, there is an indication of a shift from undernutrition to normal BMI, and from normal BMI to overweight and obesity. The ideal programmatic goal is to increase the number of adolescents with normal BMI by reducing both undernutrition and overnutrition. Prevention of adolescent undernutrition and overnutrition will lead to increased work capacity and productivity, optimal maternal health and birth outcomes, and prevention of noncommunicable diseases. Striking differences were apparent between different divisions throughout the country and between girls living in rural and urban areas. Smaller-scale studies suggest an even higher level of undernutrition among pregnant adolescents between 15 and 19 years of age, and similar levels of severe thinness in rural areas in particular. Another worrisome finding was a consistent increase in the proportion of early and late adolescents with poor dietary diversity over the three-year period of study. There is a need to develop region-specific plans that respond to the region-specific need for adolescent nutrition. Moreover, both early and late adolescent need to be included in the plan, though most early adolescents can be reached at school. A worrisome trend in dietary diversity was apparent, with the proportion of girls categorized into the poor dietary diversity group growing to over 60 percent in both early and late adolescence. Poor dietary diversity was consistently associated with household food insecurity and lower wealth quintile in the early adolescent group, as well as rural residence, low maternal education, and small household size in the older group. Encouragingly, household food insecurity almost halved over the three years under investigation, with only a quarter of adolescents living in severe food insecure households by 2014. Given national gains in household food security and female education, and known positive associations between food security, dietary diversity, and maternal schooling, trends indicating worsening dietary diversity warrant further research and policy attention. These trends also raise concerns about the limitations of standard definitions of food security that fail to take dietary diversity into account. For adolescent girls, both undernutrition and overnutrition have long-term consequences for health in adulthood, and the well-being and productivity of future 46 generations. It follows that investments in the nutrition of adolescent girls are fundamental to Bangladesh’s pursuit of economic growth and sustained health improvement. In this respect, study findings reveal national and division-level trends and urban-rural differences useful in designing context-specific responses. Analysis also reveals critical developmental entry points that need to be leveraged. Perhaps most crucial is the need to focus on the early adolescent period when nutritional needs are greatest, and when the most rapid period of growth faltering occurs. Given that the majority of these girls are in school, school-based nutrition education, supplementation and deworming programs, and fortified/ nutritious mid-morning snacks hold particular promise. For older adolescents, efforts to keep girls in school will enable these nutritional benefits to continue. The formation or support of community-based interventions, adolescent clubs, and social media platforms hold promise as a means of engaging out- of-school adolescents more directly in their own health and nutrition needs. Potential themes include healthy dietary habits, the importance of daily physical activity, and sexual and reproductive health. Ideally these interventions are animated by adolescents themselves, as adolescent involvement in the design, delivery, and evaluation of these and other focused initiatives is crucial to their relevance and success. The role of parents in supporting adolescent nutrition is similarly important, and should be emphasized in all policy and programmatic actions. Healthy dietary habits, daily physical activity, and sexual and reproductive health are also potential areas of focus, which can be animated by youth in terms of their design, delivery, and evaluation. The role of parents in supporting youth nutrition is similarly important, and can be facilitated through the dissemination of easy, affordable, and healthy family meals, and regular exercise through radio and TV spots, and cell phone messaging. Another area of opportunity revealed in this study is the myriad existing policies and strategies relevant to adolescent health and nutrition, and the many supportive legal, educational, social, and cultural rights and protections that are in place. In addition to constitutional provisions, national acts, and UN charters ensuring the equity and safety of girls and women, these policies include the National Plan of Action on Nutrition II (NPAN2) and the Second Country Investment Plan (CIP2). Of particular importance are implementation efforts around the National Plan of Action for Adolescent Health Strategy 2017–2030, which recognizes adolescent nutrition and the imperative for multiministerial engagement. Accordingly, investments in planning, executing, and sustaining interventions that enable these policies need to be made, with sensitivity to region-specific nutrition and food security needs, and the roles of the Ministries of Trade and Agriculture in ensuring the variety and seasonal availability of local foods. This assumes that government leadership can be galvanized and that financial commitments, technical capacities, and institutional mechanisms are enabled to support such investments. In addition to public policy efforts, the powerful private sector food industry must be incentivized to produce affordable nutritious food that appeals to young people, and to help diversify diet in a healthy direction. Partnerships with the private sector through existing business networks with Scaling Up Nutrition (SUN) Business Network (SBN), facilitated by Global Alliance for Improved Nutrition (GAIN), and World Food Program (WFP) would help expedite engagement both in terms of corporate social 47 responsibilities, and the creation and promotion of healthy products for adolescent audiences. Moreover, the production of unhealthy foods should be discouraged, and there should be laws to increase taxes on unhealthy foods and drink, for example, sugar- sweetened beverages. As a follow-up to the situation analysis presented in this report, a cost-benefit study of potential adolescent nutrition interventions was recently conducted by the University of Dhaka and BRAC University with the support of SKNF, GAIN, World Bank and UNICEF. In this report, six priority interventions for adolescent nutrition programming in the Bangladesh context were identified: deworming; iron and folic acid (IFA) supplementation; Multiple Micronutrient Supplementation (MMS); prevention of child marriage; school meals or fortified snacks, and school-based nutrition education. To facilitate investments in adolescent nutrition programming, ten enabling actions were identified by key stakeholders attending a policy roundtable that also occurred in response to this report: 1. Current Government investments in adolescent nutrition should be scaled-up and additional recommended interventions included into the national budget and Eighth 5-year plan. 2. Contributions towards implementation can come from funding already available under costed NPAN2 for different Ministries. 3. At least one adolescent nutrition indicator should be included in the Eighth 5-year plan to measure progress. 4. Make secondary schools a delivery platform for adolescent nutrition interventions, for example through nutrition fairs, nutrition promotion through school clubs, school health checks, and initiatives to increase physical activity. 5. Deworming should be packaged within school-based hygiene and sanitation programs alongside menstrual hygiene management for girls. 6. Nutritious school meals provided in primary schools can be extended to secondary schools especially in divisions or districts where performance on nutritional indicators is lagging. 7. Actions to delay the first pregnancy of married adolescents are an urgent priority. 8. A phased implementation of prioritized interventions is possible as more funding is made available: • Phase 1. Deworming, IFA or MMS, nutrition education • Phase 2. Deworming, IFA or MMS, nutrition education, reducing child marriage • Phase 3. Deworming, IFA or MMS, nutrition education, reducing child marriage, school meals 9. The Government should target the delivery of a full set of interventions in areas of the country where performance on nutrition indicators is lagging. 10. Development partners and civil society must work alongside Government in supporting the execution of these interventions. 48 To support these actions, information systems that monitor the nutritional status, dietary behaviors and activity levels of both adolescent girls and boys are also needed. These data will help ascertain trends by division, and rural-urban residence, and the impact of inter-sectoral and multi-level actions involving families, schools, communities, industry, social media, as well as adolescent-led initiatives. If directed in a manner that yields high impact, investments in adolescent nutrition offer a triple dividend of benefits for Bangladesh: “now, into future adult life, and for the next generation of children” (Patton et al., 2016). By investing in adolescent girls’ nutrition interventions that yield high impact, Bangladesh will reap a triple dividend of benefits in the present, the future healthy work force, and parents nurturing the next healthy generation of children in Bangladesh. 49 ANNEX 1 Table 1A.1: Sociodemographic Characteristics of the Study Participants (Early Adolescents) 2012 2013 2014 (n=2110) (n=1694) (n=1832) Age (in years) Mean ± SD 12.44 ± 1.36 12.50 ± 1.32 12.55 ± 1.31 (All adolescents) Area of residence Rural 1,870 (88.63) 1,533 (90.5) 1,601 (87.39) Urban 240 (11.37) 161 (9.50) 231 (12.61) Division Rajshahi 320 (15.17) 189 (11.16) 198 (10.81) Khulna 142 (6.73) 87 (5.14) 98 (5.35) Barisal 166 (7.87) 129 (7.62) 217 (11.84) Dhaka 491 (23.27) 455 (26.86) 475 (25.93) Sylhet 227 (10.76) 148 (8.74) 222 (12.12) Chattogram 457 (21.66) 419 (24.73) 393 (21.45) Rangpur 307 (14.55) 267 (15.76) 229 (12.5) Wealth quintile Lowest 475 (22.51) 359 (21.19) 400 (21.83) Second 478 (22.65) 363 (21.43) 374 (20.41) Middle 460 (21.80) 317 (18.71) 339 (18.50) Fourth 371 (17.58) 338 (19.95) 368 (20.09) Highest 326 (15.45) 317 (18.71) 351 (19.16) Food security Food secure 709 (33.6) 966 (57.02) 1,228 (67.03) Mild to moderate 258 (12.23) 103 (6.08) 108 (5.9) food insecure Severe food 1,143 (54.17) 625 (36.89) 496 (27.07) insecure Religion Muslim 1,892 (89.67) 1,457 (86.01) 1,578 (86.14) Non-Muslim 218 (10.33) 237 (13.99) 254 (13.86) Maternal education Primary completed 386 (18.29) 310 (18.3) 350 (19.1) No formal and 1,017 (48.20) 679 (40.08) 760 (41.48) partial primary Secondary or more 707 (33.51) 705 (41.62) 722 (39.41) Household size <5 members 670 (31.75) 552 (32.59) 549 (29.97) ≥5 members 1,440 (68.25) 1,142 (67.41) 1,283 (70.03) Mean ± SD 5.41 ± 1.71 5.43 ± 1.80 5.50 ± 1.80 Note: SD = Standard deviation. Source: Food Security and Nutrition Surveillance Project, Bangladesh 50 Table 1A.2: Sociodemographic Characteristics of the Study Participants (Late Adolescents) 2012 2013 2014 (n=3693) (n=2927) (n=3484) Age (in years) Mean ± SD 17.04 ± 1.38 17.11 ± 1.42 17.19 ± 1.41 (All adolescents) Area of residence Rural 3,230 (87.46) 2,617 (89.41) 3,029 (86.94) Urban 463 (12.54) 310 (10.59) 455 (13.06) Division Rajshahi 519 (14.05) 394 (13.46) 415 (11.91) Khulna 264 (7.15) 225 (7.69) 235 (6.75) Barisal 292 (7.91) 180 (6.15) 415 (11.91) Dhaka 779 (21.09) 656 (22.41) 853 (24.48) Sylhet 430 (11.64) 236 (8.06) 373 (10.71) Chattogram 884 (23.94) 753 (25.73) 683 (19.6) Rangpur 525 (14.22) 483 (16.5) 510 (14.64) Wealth quintile Lowest 686 (18.58) 559 (19.1) 651 (18.69) Second 756 (20.47) 559 (19.1) 681 (19.55) Middle 781 (21.15) 600 (20.5) 721 (20.69) Fourth 771 (20.88) 597 (20.4) 703 (20.18) Highest 699 (18.93) 612 (20.91) 728 (20.9) Food security Food secure 1,612 (43.65) 1,899 (64.88) 2,610 (74.91) Mild to moderate food 441 (11.94) 155 (5.30) 149 (4.28) insecure Severe food insecure 1,640 (44.41) 873 (29.83) 725 (20.81) Religion Muslim 3,135 (84.89) 2,490 (85.07) 2,942 (84.44) Non-Muslim 558 (15.11) 437 (14.93) 542 (15.56) Maternal education Primary completed 479 (12.97) 385 (13.15) 434 (12.46) No formal and partial 702 (19.01) 518 (17.7) 629 (18.05) primary Secondary or more 2,512 (68.02) 2,024 (69.15) 2,421 (69.49) Household size <5 members 1,491 (40.37) 1,248 (42.64) 1,494 (42.88) ≥5 members 2,202 (59.63) 1,679 (57.36) 1,990 (57.12) Mean ± SD 5.18 ± 2.08 5.09 ± 2.02 5.12 ± 2.04 Note: SD = Standard deviation. Source: Food Security and Nutrition Surveillance Project, Bangladesh 51 ANNEX 2 Figure 2A.1: Trends in BMI among Early and Late Adolescent Girls in Rajshahi between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 2A.2: Trends in BMI among Early and Late Adolescent Girls in Khulna between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 52 Figure 2A.3: Trends in BMI among Early and Late Adolescent Girls in Barisal between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 2A.4: Trends in BMI among Early and Late Adolescent Girls in Dhaka between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 53 Figure 2A.5: Trends in BMI among Early and Late Adolescent Girls in Sylhet between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 2A.6: Trends in BMI among Early and Late Adolescent Girls in Chattogram between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 54 Figure 2A.7: Trends in BMI among Early and Late Adolescent Girls in Rangpur between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 55 ANNEX 3 Figure 3A.1: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Rajshahi between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 3A.2: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Khulna between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 56 Figure 3A.3: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescents Girls in Barisal between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 3A.4: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Dhaka between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 57 Figure 3A.5: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Sylhet between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh Figure 3A.6: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Chattogram between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 58 Figure 3A.7: Trends in Height-for-Age Z Score (HAZ) Categories among Early and Late Adolescent Girls in Rangpur between 2012 and 2014 Source: Food Security and Nutrition Surveillance Project, Bangladesh 59 REFERENCES Akhter, N. and F. Y. Sondhya. 2013. Nutritional status of adolescents in Bangladesh: Comparison of severe thinness status of a low-income family's adolescents between urban and rural Bangladesh. J Educ Health Promot 2: 27. doi: 10.4103/2277- 9531.114209 Akib Khan, Junior Economist, IDinsight, Zambia & World Bank Consultant Akseer, N., S. Al-Gashm, S. Mehta, et al. 2017. Global and regional trends in the nutritional status of young people: a critical and neglected age group. Ann. N. Y. Acad. Sci. 1393, 3–20. Alam, N., S. K. Roy, T. Ahmed, et al. 2010. Nutritional status, dietary intake, and relevant knowledge of adolescent girls in rural Bangladesh. J Health Popul Nutr 28(1): 86-94. Alayne M. Adams, Associate Professor, Department of International Health, Georgetown University, Washington DC, USA & UNICEF Consultant Ipsita Sutradhar, Research Fellow, BRAC JPG School of Public Health, Bangladesh Ireen Akhter Chowdhury, Nutrition Officer (Micronutrient) UNICEF, Bangladesh Leroy, J. L., M. Ruel, C. Sununtnasuk et al. 2018. Understanding the determinants of adolescent nutrition in Bangladesh. Ann N Y Acad Sci. doi: 10.1111/nyas.13530 Malay Kanti Mridha, Associate Professor & Director of the Centre of Excellence for Non- Communicable Disease and Nutrition (CNCDN), BRAC James P Grant School of Public Health, Bangladesh MD. Tanvir Hassan, Assistant Professor & Co-Director of Centre of Excellence for Urban Equity and Health (CUEH), BRAC James P Grant School of Public Health, Bangladesh Mokbul Hossain, Analyst (Statistics), BRAC James P Grant School of Public Health, Bangladesh Mridha, M. K., S. L. Matias, C. D. Arnold, et al. 2018. Factors associated with nutritional status and dietary practices of Bangladeshi adolescents in early pregnancy. Ann N Y Acad Sci. doi: 10.1111/nyas.13568 Nguyen, P. H., L. Huybregts, T. G. Sanghvi, et al. 2018. Dietary Diversity Predicts the Adequacy of Micronutrient Intake in Pregnant Adolescent Girls and Women in Bangladesh, but Use of the 5-Group Cutoff Poorly Identifies Individuals with Inadequate Intake. J Nutr 148(5): 790-797. doi: 10.1093/jn/nxy045 Nizam Uddin Ahmed, General Secretary and Executive Director, Shornokishoree Network Foundation (SKNK) 60 Patton, G.C., S.M. Sawyer, J.S. Santelli, et al. 2016. Our future: a Lancet commission on adolescent health and well-being. Lancet 387:2423–2478. Piyali Mustaphi, Chief, Nutrition Section, UNICEF, Bangladesh Rudaba Khondker, Country Director, Global Alliance for Improved Nutrition (GAIN), Bangladesh Samiun Nazrin Bente Kamal, Senior Research Associate, BRAC James P Grant School of Public Health, Bangladesh Saydah, S., K.M. Bullard, G. Imperatore, et al. 2013. Cardio metabolic risk factors among US adolescents and young adults and risk of early mortality. Pediatrics 131: e679–e686. Sinharoy, S. S., J. L. Waid, R. Haardorfer, et al. 2018. Women's dietary diversity in rural Bangladesh: Pathways through women's empowerment. Matern Child Nutr 14(1). doi: 10.1111/mcn.12489. Twig, G., G. Yaniv, H. Levine, et al. 2016. Body-mass index in 2.3 million adolescents and cardiovascular death in adulthood. N. Engl. J. Med. 374:2430–2440. United Nations Population Fund (UNFPA). 2013. Adolescent pregnancy: a review of the evidence. Accessed July 31, 2019. www.unfpa.org/sites/default/files/pub- pdf/ADOLESCENT%20PREGNANCY_UNFPA.pdf. United Nations Population Fund (UNFPA). 2014. The power of 1.8 billion: adolescents, youth and the transformation of the future. The State of the World Population 2014. New York: United Nations Population Fund. Victora, C.G. et al. 2008. Maternal and child undernutrition: consequences for adult health and human capital. The Lancet 371, 340–357. World Health Organization (WHO). 2018. Guideline: implementing effective actions for improving adolescent nutrition. Geneva: World Health Organization; 2018. Licence: CC BY-NC-SA 3.0 IGO. Ziauddin Hyder, Senior Nutrition Specialist, The World Bank Group, Washington DC, USA 61 Adolescents are among the age groups most vulnerable to malnutrition; their situation requires priority attention. However, information on adolescent nutrition in Bangladesh is limited. Using data from the Food Security and Nutrition Surveillance Project (FSNSP), we examined the nutritional situation of adolescent girls, including regional and urban-rural patterns in undernutrition and overnutrition, dietary diversity, household food security, as well as their growth dynamics. Our analysis focused on data collected from 2012 to 2014. The total sample size was 15,740 adolescent girls age 10 to 19 years, of which one-third were early adolescents (age 10 to14 years) and one-tenth lived in urban areas. We found that among younger adolescent girls (age 10 to 14), the proportion of moderate to severe thinness declined from 35 to 28 percent between 2012 and 2014, and rates of overweight and obesity were consistently low. For older adolescent girls (age 15 to 19), the proportion of moderate to severe thinness remained low across the study period, while rates of overweight and obesity increased from 13 to 23 percent between 2012 and 2014. Overall, 17 percent of younger adolescent girls were stunted in 2012, decreasing to 11 percent in 2014. Study findings also highlighted substantial regional variations in both age groups. Of particular concern was a decrease in dietary diversity. The proportion of younger adolescent girls falling into the poor dietary diversity group increased from 54 percent in 2012 to 60 percent in 2014, and for older adolescent girls, a similar pattern was evident, with rates increasing from 53 to 64 percent. The analysis of growth dynamics indicated substantial deficits relative to healthy norms in the younger adolescent period. Study findings emphasize the importance of leveraging critical developmental entry points through high-impact adolescent nutrition interventions. These investments will help ensure a future healthy workforce and a healthy next generation of children in Bangladesh. ABOUT THIS SERIES: This series is produced by the Health, Nutrition, and Population Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual author/s whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Editor Martin Lutalo (mlutalo@ worldbank.org) or HNP Advisory Service (askhnp@worldbank.org, tel 202 473-2256). For more information, see also www.worldbank.org/hnppublications. 1818 H Street, NW Washington, DC USA 20433 Telephone: 202 473 1000 Facsimile: 202 477 6391 Internet: www.worldbank.org E-mail: feedback@worldbank.org