HEALTH, NUTRITION, and POPULATION SERIES India's Undernourished Children A Call for Reform and Action Michele Gragnolati, Caryn Bredenkamp, Meera Shekar, Monica Das Gupta, Yi-Kyoung Lee India's Undernourished Children Health, Nutrition, and Population Series India's Undernourished Children A Call for Reform and Action Michele Gragnolati Caryn Bredenkamp Meera Shekar Monica Das Gupta Yi-Kyoung Lee THE WORLD BANK Washington, DC © 2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 :: 09 08 07 06 This volume is a product of the staff of the International Bank for Reconstruction and Develop- ment / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the gov- ernments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, 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. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with com- plete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6587-8 eISBN: 0-8213-6588-6 ISBN-13: 978-0-8213-6587-8 DOI: 10.1596/978-0-8213-6587-8 Library of Congress Cataloging-in-Publication Data has been applied for. Contents Figures, Tables, and Boxes vii Foreword x Acknowledgments xii Acronyms xv Overview xvii 1. Dimensions of Child Undernutrition in India 1 Why Invest in Combating Undernutrition? 4 Prevalence of Underweight 10 Prevalence of Micronutrient Deficiencies 21 Will India Meet the Nutrition MDG? 28 Conclusions 31 2. The Integrated Child Development Services Program: Are Results Meeting Expectations? 33 How ICDS Aims to Address the Causes of Persistent Undernutrition 34 Empirical Findings on the Impact of ICDS 41 Targeting of ICDS Program and Beneficiaries 42 Characteristics and Quality of ICDS Service Delivery 52 Monitoring and Evaluation 59 Lessons from Successful Innovations 64 v vi · Contents 3. Enhancing the Impact of ICDS? 71 Mismatches between Program Design and Implementation 72 How Can ICDS Reach Its Full Potential? 73 Next Steps: Rationalizing Design and Improving Implementation 87 Appendix: Additional Figures and Tables 91 Notes 99 Bibliography 105 Figures, Tables, and Boxes Figures 1.1 The weight-for-age distribution of children under 3 in India compares unfavorably with the global distribution 2 1.2 The prevalence of undernutrition in children under 3 fell modestly in India, 1992 and 1998 3 1.3 The prevalence of underweight and stunting among children under 5 in rural India fell between the mid-1970s and the late 1990s 4 1.4 In terms of underweight, India compares poorly with other countries at similar levels of economic development 12 1.5 Girls whose families are poor, belong to a scheduled tribe or caste, live in a rural area, and are at risk of being underweight 14 1.6 By the age of 2, most of the damage from undernourishment has been done 15 1.7 Demographic and socioeconomic variation in prevalence of underweight children under 3, 1992/93­1998/99 15 1.8 In 1998/99, more than half of all underweight children in India lived in just one-quarter of all villages and districts 16 1.9 Urban-rural disparities in underweight among children, by state, 1992/93­1998/99 18 1.10 Change in prevalence of underweight, by wealth tertile and state, 1992­8 20 1.11 Trends in prevalence of iron deficiency in preschool children, by world region, 1990, 1995, and 2000 23 vii viii · Figures, Tables, and Boxes 1.12 Prevalence of anemia among children 6­35 months and women of reproductive age, by demographic and socioeconomic characteristics, 1998/99 23 1.13 Changes in prevalence of subclinical Vitamin A deficiency among children under 6, by world region, 1990, 1995, and 2000 25 1.14 Proportion of children experiencing daytime and nighttime vision difficulties 26 1.15 Prevalence and number of iodine deficiency disorders in the general population, by world region and country 27 1.16 Predicted prevalence of underweight under different economic growth scenarios, 2002­15 30 1.17 Projected percentage of children under 3 in poor states who are underweight, under different intervention scenarios, 1998­2015 31 2.1 Causes of child malnutrition 35 2.2 The percentage of children 6 months to 6 years enrolled in the supplementary nutrition program, 2002, varies widely across states 44 2.3 ICDS coverage is higher in states with higher per capita net domestic product 45 2.4 In many states in which the prevalence of underweight is high, the proportion of villages with anganwadi centers is low 45 2.5 Fewer children are enrolled in ICDS in states in which the prevalence of underweight is high 46 2.6 Public expenditure by state and national governments is very low in states in which the prevalence of underweight is very high 47 2.7 Older children are more likely than younger children to attend an anganwadi center 48 2.8 The caste and tribe composition of children attending anganwadi centers varies somewhat across states 49 2.9 The percentage of children who attend anganwadi centers varies only slightly across wealth quintiles 50 2.10 Attendance at anganwadi centers varies widely both across and within states 51 2.11 Percentage of anganwadi centers with growth-monitoring equipment in place 52 Figures, Tables, and Boxes · ix Tables 1.1 Prevalence of micronutrient deficiencies in selected countries in South Asia (percent except where indicated otherwise) 5 1.2 Estimated productivity losses due to malnutrition in India 10 1.3 Percentage of children suffering from underweight, stunting, and wasting, by world region, 2000 10 1.4 Prevalence of underweight and severe underweight in children under 3, by demographic and socioeconomic group, 1992/93­1998/99 13 1.5 Prevalence of underweight, 1992/93 and 1998/99, by state 17 1.6 Classification of states by change in gender differentials in prevalence of underweight 19 1.7 Wealth disparities in the change in underweight prevalence, by state, 1992/93 and 1998/99 21 1.8 Under all likely economic growth scenarios, India will not reach the nutrition MDG without direct nutrition interventions 29 2.1 Range of services that the ICDS seeks to provide to children and women 39 2.2 Comparison of intermediate health outcomes and behaviors across children living in villages with and without an anganwadi center 43 2.3 Regularity of food supply to anganwadi centers and the availability of the take-home food program 54 2.4 Anganwadi center (AWC) infrastructure, by location 57 3.1 Menu of options for improving ICDS 74 Boxes 1.1 How is malnutrition defined? 2 1.2 The "South Asian Enigma": Why is undernutrition so much higher in South Asia than in Sub-Saharan Africa? 11 2.1 Getting things right in the Bellary district of Karnataka: A report from the field 61 Foreword India has one of the highest rates of malnutrition in the world. Nearly one in every two of India's 120 million children is underweight, almost double the prevalence in Sub-Saharan Africa. An undernour- ished child will fail to reach her human potential in her adult years-- in terms of educational attainment, health and productivity--perpetu- ating a vicious cycle of poverty and malnutrition. Progress in reducing the number of undernourished children in India over the past decade has been slower than in comparable coun- tries. The shockingly high levels of undernutrition are exacerbated by significant and increasing inequalities across states and socioeconomic groups--girls, rural areas, the poorest, and scheduled tribes and castes are the worst affected. Halving the prevalence of underweight children by 2015 is a key indicator of progress towards the Millennium Development Goal (MDG) of eradicating extreme poverty and hunger. Achieving the tar- get will require difficult choices. It cannot be met by economic growth alone, however impressive that may be at the present time. In India, until recently, food insecurity has been viewed as the pri- mary or even sole cause of child malnutrition. By contrast, research indicates that high levels of exposure to infection and inappropriate child feeding and caring practices, especially during the first two to three years of life, are salient. This misperception has resulted in resources being skewed towards ineffective food-based interventions. India's main early child development intervention, the Integrated Child Development Services program (ICDS), has been operating for about 30 years. While it has certainly had some successes, it does not x Foreword · xi appear to have made a significant dent in child malnutrition. There are two main reasons. First, it has prioritized food supplementation over nutrition and health education interventions. Second, it has focused on children above the age of three, by which time the irre- versible effects of malnutrition have already set in. Transforming the ICDS into an intervention that effectively addresses the principal causes of malnutrition will yield huge human and economic benefits for India. However, this will require substantial changes in the pro- gram's design and implementation. In particular, public investments in the ICDS should be redirected towards the younger children (0­3 years) and the most vulnerable population groups within those states and districts with a high prevalence of undernutrition. The focus should be on those ICDS components that directly address the most important causes of undernutrition in India, specifically improving child feeding and care behaviors, strengthening the referral to the health system, and providing micronutrients. The Government of India recently launched the National Rural Health Mission and the National Nutrition Mission, and has also committed itself to rapidly expand the ICDS program. A review of the characteristics of undernutrition in India and of the ICDS is therefore particularly timely. This report analyzes the successes and failures of current child nutrition policy in India and identifies effective policies and programs which could significantly reduce the current high levels of child malnutrition, and, in so doing, help break the cycle of malnu- trition and poverty. Julian Schweitzer Prafu Patel Director Vice President Human Development Department South Asia Region South Asia Region Acknowledgments Work on this report has been supported by generous funding from the Netherlands Ministry of Foreign Affairs, through the Bank- Netherlands Partnership Program. The report was authored by Michele Gragnolati (task team leader), Caryn Bredenkamp (University of North Carolina at Chapel Hill), Meera Shekar, Monica Das Gupta (World Bank), and Yi-Kyoung Lee (World Bank). A number of background papers were prepared in advance of the first draft. These include: "Who Does India's ICDS Nutrition Program Reach, and What Effect Does It Have?" by Monica Das Gupta, Michael Lokshin, and Oleksiy Ivaschenko (Development Economics Research Group [DECRG], World Bank). "Case Study on Mid-Day Meal Scheme of Tamil Nadu and Gujarat," by P. Subramaniyam. "Analysis of Public Expenditures and Impact of Public Distribution System (PDS) on Food Security," by S. Mahendra Dev. "India's Integrated Child Development Services Scheme: Meeting the Health and Nutritional Needs of Children, Adolescent Girls and Women?" by Caryn Bredenkamp and John S. Akin (University of North Carolina at Chapel Hill). "Literature Review of MDM, ICDS, and PDS (1992­2003), Includ- ing Annotated Bibliography," by New Concept Information Sys- tems, India. xii Acknowledgments · xiii "Analysis of Positive Deviance in the ICDS Program in Rajasthan and Uttar Pradesh," by Educational Resource Unit, India. "Monitoring and Evaluation in India's ICDS Programme," by Saroj K. Adhikari, Department of Women and Child Development, Gov- ernment of India. "Reviewing the Costs of Malnutrition in India," by Laveesh Bhan- dari and Lehar Zaidi, Indicus Analytics, India. "Will Asia Meet the Nutrition Millennium Development Goal? and Even If It Does, Will It Be Enough?" by Meera Shekar (HDNHE, World Bank), Mercedes de Onis, Monika Blössner, and Elaine Borghi (Department of Nutrition for Health and Development, World Health Organization). Peer reviewers were Prof. Abhijit Sen (Planning Commission, Government of India), Ruth Levine (Center for Global Develop- ment), and Harold Alderman (DECRG, World Bank). The final report was strengthened by valuable comments from the Department of Women and Child Development (DWCD), Government of India. A number of technical experts provided inputs and reviews at vari- ous stages of the report's development. Peer reviewers involved in the conceptualization of the project were Ruth Levine (Center for Global Development), John S. Akin (University of North Carolina at Chapel Hill), Harold Alderman, Meera Shekar, and Jishnu Das (World Bank). Additional analysis of various data underpinning the report was per- formed by Peter Heywood, Himani Pruthi, Jayshree Balachander, Venkatachalam Selvaraju and Julie Babinard (World Bank and con- sultants to the World Bank). Information on some of the case studies included in this report was generously shared by Deepika Chaudhery, T. Usha Kiran, and others at CARE-India. Additional inputs and comments were received from Paoli Belli, Alan Berg, Barbara Kafka (World Bank), Werner Schultnik (UNICEF, India), and Arun Gupta. The Government of India and respective State Governments pro- vided data from a baseline survey of the ICDS III program and an end- line survey of the ICDS II program. These data were collected by research teams at six research organizations, namely Agricultural Finance Corporation (AFCIndia), Indian Institute of Development xiv · Acknowledgments Management (IIDM), Indian Institute of Health Management Research (IIHMR), ORG Centre for Social Research, Rajagiri College of Sciences (RCSS), and Xavier Institute of Social Sciences (XISS). Overall project guidance and specific comments were provided by Anabela Abreu, Peter Berman, Charlie Griffin, Meera Priyadarshi, and Julian Schweitzer. Program support and administrative assistance were provided by Nira Singh and Elfreda Vincent, and editorial and publishing assistance by Rama Lakshminarayanan, Miyuki Parris, Jen- nifer Vito, Paola Scalabrin, and Mark Ingebretsen. Abbreviations ANC antenatal care ANM auxiliary nurse-midwife AWC anganwadi center AWH anganwadi helper AWW anganwadi worker BMI body mass index CDPO Child Development Project Officer DALY disability-adjusted life year DHFW Department of Health and Family Welfare DHS Demographic and Health Survey DWCD Department of Women and Child Development GDP gross domestic product HAZ height-for-age z-score ICDS Integrated Child Development Services ICN International Conference on Nutrition IDA iron deficiency anemia IDD iodine deficiency disorder IFA iron and folic acid IMR infant mortality rate LAC Latin America and the Caribbean LHW lady health-worker M&E monitoring and evaluation MDG Millennium Development Goal MoHFW Ministry of Health and Family Welfare NFHS National Family Health Survey NID National Immunization Day PEM protein energy malnutrition xv xvi · Abbreviations PPP purchasing power parity PRI panchayat raj institution RCH reproductive and child health program SAR South Asia Region SNP supplementary nutrition program TB tuberculosis VAD Vitamin A deficiency VPD vaccine preventable disease WAZ weight-for-age z-score WCD women and child development WHZ weight-for-height z-score Overview The World Bank has supported efforts to improve nutrition in India since 1980, with mixed results. This report aims at helping policymak- ers by providing information on the characteristics of child malnutri- tion in India and on the effectiveness of the Integrated Child Devel- opment Services (ICDS) program in addressing the causes and symptoms of undernutrition. The report identifies the most impor- tant mismatches between the program's intentions and its implemen- tation and presents some options for resolving the mismatches and creating a more effective, efficient, and equitable program. A short summary of each of the three chapters of the report is pre- sented below. Chapter 1: Dimensions of Child Undernutrition in India Child undernutrition has enormous consequences for child and adult morbidity and mortality. In addition, undernutrition reduces produc- tivity, so that a failure to invest in combating malnutrition effectively diminishes the potential for economic growth. In India, the situation is dire: the prevalence of underweight among children is nearly twice that of Sub-Saharan Africa, and inequalities in undernutrition between demographic, socioeconomic, and geographic groups have been increasing. More, and better, investments are needed if India is to reach the nutrition Millennium Development Goal (MDG) target. Economic growth alone will not be enough. Undernutrition--both protein-energy malnutrition and micronu- trient deficiencies--directly affects many aspects of children's devel- xvii xviii · Overview opment. It retards their physical and cognitive growth and increases susceptibility to infection and disease, further increasing the probabil- ity of being malnourished. As a result, undernutrition has been esti- mated to be associated with about half of all child deaths. More than half of child deaths from diarrhea (61 percent), malaria (57 percent), and pneumonia (52 percent) are associated with malnutrition, as well as 45 percent of deaths from measles. Child undernutrition in India is responsible for 22 percent of the country's burden of disease. Undernutrition also affects cognitive and motor development, and it undermines educational attainment. Ultimately, it affects productiv- ity at work and at home, with adverse implications for income and economic growth. Micronutrient deficiencies alone may be costing India $2.5 billion a year. Most growth retardation occurs by the age of 2--in part because about 30 percent of Indian children are born with low birth weight-- and it is largely irreversible. In 1998/99 (the latest year for which nationally representative data are available), almost three-quarters of Indian children under age 3 were below the normal weight for their age, with 47 percent underweight or severely underweight and another 26 percent mildly underweight. Levels of malnutrition declined modestly in the 1990s, with the prevalence of underweight among children under 3 falling 11 percent between 1992/93 and 1998/99. This progress lags far behind that achieved by countries with similar economic growth rates. Disaggregation of underweight statistics by socioeconomic and demographic characteristics reveals which groups are at greatest risk of malnutrition. Underweight prevalence is higher in rural areas (50 per- cent) than in urban areas (38 percent), higher among girls (49 percent) than among boys (46 percent), higher among scheduled castes (53 per- cent) and scheduled tribes (56 percent) than among other castes (44 per- cent), and although it is pervasive throughout the wealth distribution, the prevalence of underweight reaches as high as 60 percent in the low- est wealth quintile. Moreover, during the 1990s, urban-rural, intercaste, male-female, and interquintile inequalities in nutritional status widened. Interstate variation in the patterns and trends in underweight is large. In six states (Bihar, Madhya Pradesh, Maharashtra, Orissa, Rajasthan, and Uttar Pradesh), at least half of all children are underweight. Four states--Bihar, Madhya Pradesh, Rajasthan, and Uttar Pradesh-- account for more than 43 percent of all underweight children in India. Overview · xix Moreover, the prevalence of underweight is falling more slowly in high- prevalence states. The demographic and socioeconomic patterns at the state level do not necessarily mirror those at the national level. In some states, for example, inequalities in underweight are narrowing, not widening; in others, boys are more likely to be underweight than girls. Nutrition policy should take these differences into account. Undernutrition is concentrated in a relatively small number of dis- tricts and villages, with a mere 10 percent of villages and districts accounting for 27­28 percent of all underweight children and a quarter of districts and villages accounting for more than half of all underweight children. This distribution suggests that future efforts to combat mal- nutrition could give priority to a relatively small number of districts and villages. Micronutrient deficiencies are also widespread in India. More than 75 percent of preschool children suffer from iron deficiency anemia, and 57 percent have subclinical Vitamin A deficiency. Iodine defi- ciency is endemic in 85 percent of districts. Progress in reducing the prevalence of micronutrient deficiencies in India has been modest. As with underweight, the prevalence of different micronutrient deficien- cies varies widely across states. Economic growth alone is unlikely to be sufficient to significantly lower the prevalence of malnutrition--it will certainly not be suffi- cient to meet the MDG target of halving the prevalence of under- weight children between 1990 and 2015. Only by rapidly scaling up health, nutrition, education, and infrastructure interventions and improving their effectiveness can this target be met. This is especially critical in the poorest states. Chapter 2: The Integrated Child Development Services Program (ICDS) India's primary policy response to child malnutrition, the ICDS pro- gram, is well conceived and well placed to address the major causes of child undernutrition in India. But more attention has been given to increasing coverage than to improving the quality of service delivery, and the program has focused more on distributing food than on changing family-based feeding and caring behavior. As a result, impact has been limited. xx · Overview The ICDS has expanded tremendously over its 30 years of opera- tion to cover almost all development blocks in India. It offers a wide range of health, nutrition, and education services to children, women, and adolescent girls. The program is intended to target the needs of the poorest and the most undernourished, as well as the age groups that represent a "window of opportunity" for nutrition investments (that is, children under 3 and pregnant and lactating women). There is a mismatch, however, between the program's intentions and its actual implementation: · The central focus on food supplementation drains resources from other tasks envisaged in the program that are crucial for improving child nutritional outcomes. For example, not enough attention is given to educating parents about how to improve childcare behav- iors and feeding practices. · Older children (3­6) participate much more than younger ones, and many children from poorer households do not yet participate. The program fails to preferentially target girls, children from lower castes, or children from the poorest villages, all of whom are at higher risk of undernutrition. · Although expansion of the program was greater in underserved than well-served areas during the 1990s, the poorest states and those with the highest levels of undernutrition still have the lowest levels of program funding and coverage by ICDS activities. In addition to these mismatches, the program faces substantial opera- tional challenges. Inadequate worker skills, shortages of equipment, poor supervision, and weak monitoring and evaluation reduce the pro- gram's potential impact. Community workers are overburdened, because they are expected to provide preschool education to 4- to 6- year-olds as well as nutrition services to all children under 6. As a result, most children under 3--for whom nutrition interventions can have the largest impact--do not receive micronutrient supplements, and most of their parents are not reached with counseling on better feeding and childcare practices. Examples of successful ICDS interventions (in some districts) and innovations and variants of ICDS in several states (the INHP II in nine states, the Dular scheme in Bihar, and the Tamil Nadu Overview · xxi Integrated Nutrition Project) suggest that the potential for better implementation and greater impact does exist. Chapter 3: Enhancing the Impact of ICDS ICDS was designed to address the multidimensional causes of malnu- trition. As the program expands to reach more and more villages, it has tremendous potential to improve the well-being of the millions of women and children who are eligible for participation. The key con- straint on its effectiveness is that implementation deviates from the original design. Realizing ICDS' potential will require substantial commitment and resources in order to realign its implementation with its original objectives and design. Several steps need to be taken: · Ambiguity over the priority of different program objectives and interventions must be resolved. · Activities need to be refocused on the most important determinants of malnutrition. Programmatically, this means emphasizing disease control and prevention activities, education to improve domestic childcare and feeding practices, and micronutrient supplementation. Greater convergence with the health sector, in particular the Repro- ductive and Child Health Program, would help tremendously in this regard. · Activities need to better target the most vulnerable age groups (chil- dren under 3 and pregnant women). Funds and new projects need to be redirected to the states and districts with the highest prevalence of malnutrition. · Supplementary feeding activities need to better target those who need them most, and growth-monitoring activities need to be per- formed with greater regularity, with an emphasis on using this process to help parents understand how to improve their children's health and nutrition. · Communities need to be involved in implementing and monitoring ICDS, in order to bring additional resources to the anganwadi cen- xxii · Overview ters, improve the quality of service delivery, and increase accounta- bility in the system. · Monitoring and evaluation activities need to be strengthened through the collection of timely, relevant, accessible, high-quality information, and this information needs to be used to improve program function- ing by shifting the focus from inputs to results, using data to inform decisions, and creating accountability for performance. CHAPTER 1 Dimensions of Child Undernutrition in India Child undernutrition has enormous consequences for morbidity and mortality. It also affects productivity, so that failure to invest in nutrition today reduces potential economic growth tomorrow. In India, where the prevalence of under- nutrition is nearly twice that of Sub-Saharan Africa, the situation is dire. More, and better, investments are needed if undernutrition is to be reduced; growing inequalities in nutrition across demographic, socioeconomic, and geo- graphic groups diminished; and the nutrition Millennium Development Goal (MDG) target reached. Economic growth alone will not be enough. The prevalence of underweight among children in India is among the highest in the world (box 1.1). About 37 million children under the age of 3 are underweight, and many more suffer from various micronutrient deficiencies. In recent years, the prevalence of under- nutrition has declined only slightly. Dealing with malnutrition is thus an urgent policy priority (World Bank 2004a).1 As a result of undernutrition, the distribution of children's age- standardized weight is far to the left of the global reference standard (figure 1.1). In 1998/99 (the latest year for which nationally represen- tative data are available), almost three-quarters of Indian children under 3 were below the normal weight for their age. Forty-seven per- cent were underweight, of which 18 percent were severely under- weight and 26 percent were mildly underweight. About 46 percent of children were stunted, and 16 percent could be classified as wasted. Given that even mild malnutrition is linked to a twofold increase in 1 2 · India's Undernourished Children Box 1.1 How is malnutrition defined? Nutritional status is typically described in terms of anthropometric indices, such as underweight, stunting, and wasting. These terms are measures of protein-energy undernutrition and are used to describe children who have a weight-for-age, height- (or recumbent length-) for-age, and weight-for-height that is less than two standard deviations below the median value of the National Center for Health Sta- tistics­World Health Organization (WHO) reference group. These children are con- sidered to suffer from moderate malnutrition. The terms severe underweight, severe stunting, and severe wasting are used when the measurements are less than three standard deviations below the reference median; mild underweight, stunting, and wasting refer to measurements of less than one standard deviation below the refer- ence population. Underweight is generally considered a composite measure of long- and short-term nutritional status; stunting reflects long-term nutritional status, and wasting is an indicator of acute short-term undernutrition. Some indicators of micronutrient malnutrition are also used to measure malnutrition. The most com- mons forms of micronutrient malnutrition referred to in this report are Vitamin A deficiency, iodine deficiency disorders, and iron-deficiency anemia. mortality and to greatly reduced productivity levels, these levels of undernutrition significantly compromise health and productivity. The nutritional status of children improved modestly during the 1990s. Between 1992/93 and 1998/99, the prevalence of underweight fell almost 11 percent, equivalent to a 1.5 percent annual reduction (figure 1.2). But this improvement lagged far behind that achieved by countries with similar economic growth rates. Figure 1.1 The weight-for-age distribution for children under 3 in India compares unfavor- ably with the global distribution distribution curve normal distribution curve for Indian children (international reference) moderate underweight mild overweight severe underweight moderate overweight ­6.0 ­5.0 ­4.0 ­3.0 ­2.0 ­1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Source: Calculated from NFHS II (1998/99) data. Dimensions of Child Undernutrition in India · 3 Figure 1.2 The prevalence of undernutrition in children under 3 fell modestly in India, 1992 and 1998 80 ? 73 69 70 68 children 60 53 49 50 47 47 45 46 40 30 25 22 23 20 18 18 15 undernourished of 10 3 3 % 0 1992 1998 1992 1998 1992 1998 underweight stunting wasting Mild Moderate Severe Source: NFHS I (1992/93) and NFHS II (1998/99). The reduction was in line with gains made earlier (figure 1.3). According to the WHO Global Database on Child Growth and Nutrition (WHO 2004c), the prevalence of malnutrition among chil- dren under 5 in rural India fell from more than 70 percent in the late 1970s to less than 50 percent at the end of the 1990s for both under- weight and stunting measures. The prevalence of severe stunting also declined over this period, from almost 50 percent to less than 25 per- cent, while the prevalence of severe underweight declined from 37 percent to less than 20 percent. The prevalence of micronutrient deficiencies among children and women of reproductive age in India is consistently among the highest in the world (table 1.1). More than 75 percent of preschool children suffer from iron deficiency anemia. Up to 60 percent have subclinical Vitamin A deficiency, although less than 2 percent suffer from clinical Vitamin A deficiency.2 About one in four school children has goiter, a sign of severe iodine deficiency (UNICEF 2003b; WHO 2000; UNICEF and MI 2004a). Among ever-married women 15­49, 52 percent have some degree of anemia, with the prevalence of anemia among some groups of pregnant women reaching 87 percent. Clinical Vitamin A deficiency affects about 5 percent of women and subclinical Vitamin A deficiency about 12 percent of women. Iodine deficiency in pregnant women in India is estimated to have caused the congenital mental impairment of about 6.6 million children (IIPS and Orc Macro 2000; UNICEF 2003b). 4 · India's Undernourished Children Figure 1.3 The prevalence of underweight and stunting among children under 5 in rural India fell between the mid-1970s and the late 1990s 80 70 60 50 40 children 30 % undernourished20 10 0 1974­9 1988­90 1991­2 1995­6 1996­7 1974­9 1988­90 1991­2 1995­6 1996­7 underweight stunting Severe underweight Moderate underweight Source: WHO 2004a. Note: Prevalence is not strictly comparable across time periods, since each round of surveys used different sampling methodologies and calculated prevalence across different age groups. Why Invest in Combating Undernutrition? Failing to deal effectively with the undernutrition problem in India has dire consequences for children's development. It retards their physical growth and increases their susceptibility to disease in child- hood and adulthood. It also affects cognitive and motor development, limits educational attainment and productivity, and ultimately perpet- uates poverty. Moreover, in a country where undernutrition is so widespread, the consequences of undernutrition go well beyond the individual, affecting total labor-force productivity and economic growth. Effect of Undernutrition on Morbidity, Mortality, and Cognitive and Motor Development By precipitating disease and speeding its progression, malnutrition is a leading contributor to infant, child, and maternal mortality and mor- bidity. It has been estimated to play a role in about half of all child deaths (Horton 1999; Pelletier and others 1995; Pelletier and Frongillo 2003), and in more than half of child deaths from diarrhea (61 percent), malaria (57 percent), and pneumonia (52 percent). Mal- nutrition is also involved in 45 percent of all child deaths from measles (Black, Morris, and Bryce 2003; Caulfield and others 2004). Pediatric malnutrition is a risk factor for 16 percent of the global burden of dis- Table 1.1 Prevalence of micronutrient deficiencies in selected countries in South Asia (percent except where indicated otherwise) Folate Iron deficiency anemia Vitamin A deficiency Iodine deficiency deficiency Number of Number maternal Children Children Total of Number deaths under 6 under 6 goiter children of from with with Number rate born neural In In In severe subclinical clinical of child among Total mentally tube children women pregnant anemia Vitamin A Vitamin A deaths school goiter impaired defects Country under 5 15­49 women per year deficiency deficiency precipitated children rate per year per year Afghanistan 65 61 -- -- -- 53 50,000 -- 48 535,000 2,250 Bangladesh 55 36 74 2,800 0.7 28 28,000 50 18 750,000 8,400 Bhutan 81 55 68 <100 0.7 32 600 14 -- -- 150 India 75 51 87 22,000 0.7 57 330,000 19 26 6,600,000 50,000 Nepal 65 62 63 760 1.0 33 6,900 40 24 200,000 1,600 Pakistan 56 59 -- -- -- 35 56,000 -- 38 2,100,000 11,000 South Asia region -- -- -- 25,560 -- -- 471,500 -- -- 10,185,000 73,400 World -- -- -- 50,000 -- -- 1,150,000 -- -- 19,000,000 204,000 Source: UNICEF 2003b; WHO 2000; UNICEF and MI 2004a. -- Not available. 6 · India's Undernourished Children ease, but it accounts for as much as 22 percent of India's burden of dis- ease (Murray and Lopez 1997). Consequences of Protein-Energy Malnutrition Isolating the effects of protein and energy deficiencies on health and development outcomes is confounded by the fact that when food intake is low, the intake of many other nutrients is usually also inadequate (Allen 1994).3 Never- theless, it is generally accepted that children who are underweight or stunted are at greater risk for childhood morbidity and mortality, poor physical and mental development, inferior school performance, and reduced adult size and capacity for work (WHO 1995). Protein-energy malnutrition weakens immune response and exac- erbates the effects of infection (Pelletier and Frongillo 2003). As a result, children who are malnourished tend to have more severe diar- rheal episodes and are at a higher risk of pneumonia. Infections, in turn, contribute to malnutrition, through a variety of mechanisms, including loss of appetite and reduced capacity to absorb nutrients (Calder and Jackson 2000). Underweight and stunted women are also at higher risk of obstetric complications (because of smaller pelvic size) and low birth weight deliveries (ACC/SCN 1997). The result is an intergenerational cycle of malnutrition, since low birth weight infants tend to attain smaller stature as adults. Malnutrition in early infancy is also correlated with increased sus- ceptibility to chronic disease in adulthood, including coronary heart disease, diabetes, and high blood pressure (Agarwal and others 1998; Agarwal and others 2002; Barker and others 2001; Lucas, Fewtrell, and Cole 1999; Popkin and others 2001; UNICEF 1998). Although the precise mechanisms are not clear, protein-energy malnutrition during the last trimester of pregnancy and the first two years of life is also associated with poor cognitive and motor develop- ment. The magnitude of the effect depends on the severity and dura- tion of malnutrition as well as its timing: moderate protein-energy malnutrition of long-term duration has worse consequences for cog- nitive development than transient severe undernutrition. Consequences of Micronutrient Deficiencies Iron and Vitamin A deficiencies are leading risk factors for disease in developing countries, especially countries with high mortality rates (WHO 2002). Iodine deficiency also carries a mortality risk. Dimensions of Child Undernutrition in India· 7 Vitamin A deficiency. Vitamin A deficiency is a well-known cause of morbidity and mortality, especially among young children and preg- nant women. In young children, clinical Vitamin A deficiency can cause xerophthalmia (a dry, thickened, lusterless condition of the eye- ball) and keratomalacia (a softening, drying, and ulceration of the cornea), and it can lead to blindness (Vinutha, Metha, and Shanbag 2000). Subclinical Vitamin A deficiency, defined by a serum retinol concentration of less than 0.7 mol/L, can limit children's growth, weaken the immune system, exacerbate infection, and increase the risk of death (West 2002), mainly from respiratory and gastrointesti- nal infections. Often occurring concurrently among children with protein-energy malnutrition, Vitamin A deficiency is estimated to be responsible for about 1 million child deaths a year (Mason and others 2005). Pregnant women, especially in the third trimester, when micronutrient demands are at their highest, often exhibit a high prevalence of night blindness. Recent studies have shown that Vita- min A deficiency may also be associated with an increased risk of mother-to-child transmission of HIV, although Vitamin A supple- mentation fails to lower the risk of transmission (Stephenson 2003). In general, Vitamin A supplementation has proven successful in reducing the incidence and severity of illness, and it has been associ- ated with a reduction in child mortality of 25­35 percent (Beaton, Martorell, and Aronson 1993; Fawzi, Chalmers, and Herrera 1993), especially from diarrhea, measles, and malaria (Jones and others 2003). Iron deficiency anemia. Iron deficiency anemia is common across all age groups, although its incidence is highest among children and pregnant and lactating women. It affects about 2 billion people in developing countries. The consequences of iron deficiency anemia in pregnant women include increased risk of low birth weight or prema- ture delivery, perinatal and neonatal mortality, inadequate iron stores for the newborn, lowered physical activity, fatigue, and increased risk of maternal morbidity (Bentley and Griffiths 2003). It is also responsi- ble for almost a quarter of maternal deaths (Ross and Thomas 1996). Inadequate iron stores in a newborn child, coupled with insufficient iron intake during the weaning period, have been shown to impair intellectual development by adversely affecting language, cognitive, and motor development. Iron deficiency among adults contributes to low labor productivity (WHO 2004c; Seshadri 2001). 8 · India's Undernourished Children Iodine deficiency. Iodine deficiency during pregnancy is associated with low birth weight, increased likelihood of stillbirth, spontaneous abortion, and congenital abnormalities such as cretinism and irre- versible forms of mental impairment. During childhood it impairs physical growth, causes goiter, and decreases the probability of child survival. It is also the most common cause of preventable mental retardation and brain damage in the world (ACC/SCN 2000). Glob- ally, 2.2 billion people (38 percent of the world's population) live in regions where iodine deficiency is endemic (WHO 2002). Both iodine and iron deficiencies have been linked to the retarda- tion of cognitive processes in infants and young children. Maternal iodine deficiency has negative and irreversible effects on the cognitive functioning of the developing fetus. Postnatal iodine deficiency may also be associated with cognitive deficits (Black 2003): IQs of iodine- deficient children have been shown to average 13.5 points less than iodine-sufficient children (Bleichrodt and Born 1994); iron deficiency anemia has been associated with half a standard deviation reduction in IQ (Ross and Horton 1998). Effect of Undernutrition on Schooling, Adult Productivity, and Economic Growth The cognitive and physical consequences of undernutrition--both underweight and micronutrient deficiencies--undermine educational attainment and labor productivity, with adverse implications for income and economic growth. Malnutrition at any stage of childhood affects schooling and thus lifetime earnings potential (Alderman 2005). Some of the pathways through which malnutrition affects educational outcomes include the reduced capacity to learn (as a result of early cog- nitive deficits or lowered current attention span) and the reduction in the number of total years of schooling (since caregivers may invest less in malnourished children or schools may use child size as an indicator of school readiness) (Alderman 2005). In rural Pakistan, malnutrition has been found to decrease the probability of ever attending school, particularly for girls (Alderman and others 2001). In the Philippines, children with higher nutritional status during the preschool years start primary school earlier; repeat fewer grades (Glewwe, Jacoby, and King 2001); and have higher high school completion rates (Daniels and Adair 2004) than other children. In Zimbabwe, stunting, through its Dimensions of Child Undernutrition in India· 9 association with a seven-month delay in school completion and a 0.7- year loss in grade attainment, has been shown to reduce lifetime income by 7­12 percent (Alderman, Hentschel, and Sabates 2003). Measuring the productivity losses associated with undernutrition is complex, and since different studies incorporate different types of pro- ductivity gains, estimates can vary widely.4 Moreover, since a large share of productivity losses are measured in terms of forgone wages, when productivity losses are expressed in dollar terms rather than as a percentage of GDP, the productivity losses in India may appear small relative to countries with higher average wages. In general, in low- income agricultural countries in Asia, the physical impairment associ- ated with malnutrition is estimated to cost more than 2­3 percent of GDP a year--even without considering the long-term productivity losses associated with developmental and cognitive impairment (Hor- ton 1999). Iron deficiency in adults has been estimated to decrease pro- ductivity by 5­17 percent, depending on the nature of the work per- formed (Horton 1999). Data from 10 developing countries show that the median loss in reduced work capacity associated with anemia dur- ing adulthood is equivalent to 0.6 percent of GDP, while an additional 3.4 percent of GDP is lost due to the effects on cognitive development attributable to anemia during childhood (Horton and Ross 2003). The impact of iodine deficiency disorders on cognitive development alone has been associated with productivity losses of about 10 percent of GDP (Horton 1999). A few attempts have been made to estimate the productivity losses associated with malnutrition in India. As with global estimates, these estimates are intrinsically imprecise, requiring many assumptions and approximations. One study projects that in the absence of appropriate interventions, the productivity losses due to protein-energy malnutri- tion, iodine deficiency disorder, and iron deficiency anemia are likely to equal about $114 billion between 2003 and 2012 (Care India and Link- ages India 2003). Another study, examining only the productivity losses associated with forgone wage employment resulting from child malnu- trition, estimates the loss at $2.3 billion a year (Bhandari and Zaidi 2004). Other studies suggest that micronutrient deficiencies alone may cost India $2.5 billion a year (Alderman 2005) and that stunting, iodine deficiency, and iron deficiency together are responsible for a total productivity loss of almost 3 percent of GDP among manual workers alone (Horton 1999) (table 1.2). 10 · India's Undernourished Children Table 1.2 Estimated productivity losses due to malnutrition in India Disability-adjusted Estimated total life years lost due annual losses Estimated loss of to malnutrition due to malnutrition adult productivity Item in India (billions of dollars) (percent of GDP) Protein-energy malnutrition 2,939,000 8.1 1.4 (stunting) Vitamin A deficiency 404,000 0.4 -- Iodine deficiency disorder 214,000 1.5 0.3 Iron deficiency 3,672,000 6.3 1.25 Source: ASC 1998; World Bank 2004c; Horton 1999. -- Not available. Note: Productivity losses include market activities only. Prevalence of Underweight An International Perspective Undernutrition in India is among the worst in the world (table 1.3). In the late 1990s, the prevalence of underweight (47 percent) was about the same as in Bangladesh and Nepal (48 percent), but it was much higher than in all other countries in South Asia. It was also far higher than the averages for other regions of the world and nearly double that of Sub-Saharan Africa (box 1.2). High prevalence combined with Table 1.3 Percentage of children suffering from underweight, stunting, and wasting, by world region and country, 2000 Region/country Underweight Stunting Wasting Latin America and Caribbean 6 14 2 Africa 24 35 8 Asia 28 30 9 India 47 45 16 Bangladesh 48 45 10 Bhutan 19 40 3 Maldives 45 36 20 Nepal 48 51 10 Pakistan 40 36 14 Sri Lanka 33 20 13 All developing countries 22­27 28­32 7­9 Source: ACC/SCN 2004. Dimensions of Child Undernutrition in India · 11 India's large population means that of the 150 million malnourished children under the age of 5 in the world, more than a third live in India (UNICEF 2003b; ACC/SCN 2000; DWCD 2003). The decline in the prevalence of underweight during the 1990s was less rapid than in most other countries with similar socioeconomic or geographical characteristics (figure 1.4). Although per capita GDP in India rose by an average annual rate of 5.3 percent, the average annual prevalence of underweight fell just 1.5 percent a year. In some other countries, underweight prevalence fell more than 5 percent, even though annual per capita GDP growth was 2 percent or less. In China, where annual growth averaged 12 percent, the prevalence of child Box 1.2 The "South Asian enigma": Why is undernutrition so much higher in South Asia than in Sub-Saharan Africa? In 1997, when Ramalingaswami, Jonson, and Rohde wrote that "in the public imagi- nation, the home of the malnourished child is Sub-Saharan Africa . . . but . . . the worst affected region is not Africa but South Asia," their statement was met with incredulity. Today, undernutrition rates in South Asia, including and especially in India, are nearly twice those in Sub-Saharan Africa. This is not an artifact of different measurement standards or differing growth potential among ethnic groups: studies have repeatedly shown that given similar opportunities, children across most ethnic groups, including Indian children, can grow to the same levels and that the same internationally recognized growth references can be used across countries to assess the prevalence of malnutrition (Nutrition Foundation of India 1991). The phe- nomenon referred to as the "South Asian enigma" is real. The enigma can be explained by three key differences between South Asia and Sub-Saharan Africa: More than 30 percent of Indian babies are born with low birth weights, compared with about 16 percent in Sub-Saharan Africa. Low birth weight is the single most important predictor of undernutrition. Women in South Asia tend to have lower status and less decision-making power than women in Sub-Saharan Africa, limiting their ability to access the resources needed for their own and their children's health and nutrition. Low status of women can be linked to low birth weight, as well as poor child-feeding behaviors in the first 12 months of life. Hygiene and sanitation standards in South Asia are well below those in Sub-Saha- ran Africa. Poor hygiene and sanitation play a major role in causing the infections that lead to undernutrition in the first two years of life. 12 · India's Undernourished Children Figure 1.4 In terms of underweight, India compares poorly with other countries at similar levels of economic development 60 20 50 15 10 underweight 40 5 5 change 30 0 under 20 ­5 relative 10 ­10 % children 0 ­15 % 00 01 01 00 00 00 01 2000 20 20 1994­81992­8 1993­81992­9 1993­5 1994­91991­8­20001994­91996­8­20 ­20 ­20 1993­8 1993­9 1991­4­20 1993­96­20 1992­7 1993­2000 1996­2000 1995­2000 1992­ 1995 go 1992 1996 1996 1993­ 1993­ 1996 199 To India Bolivia China** Haiti d'Ivoire Ghana Vietnam SudanNicaragua PDR* Pakistan Mongolia* Gambia Thailand**Côte Lesotho Cameroon ComorosAngola Lanka** Nepal* Philippines** Lao Cambodia Bangladesh* Indonesia** MauritaniaSri Prevalence (left axis) Annual change, prevalence underweight children (right axis) Annual change, GDP per capita, purchasing power parity (right axis) Source: World Bank 2004b, e. Note: Countries in Asia with per capita GDP of less than $1,333 are denoted by *; countries in Asia with per capita GDP of more than $2,333 are denoted by **. Purchasing power parity is in constant 1995 international dollars. Criteria for inclusion in the graph were as follows: At least two household surveys were con- ducted between 1990 and 2002 in each of the countries displayed. When more than two sur- veys were available, information collected around 1992/93 and 1998/99 was used, to enhance comparability with data from India's NFHS. Countries with a prevalence of underweight among children under 5 of less than 10 percent in the first survey were dropped. Countries are either in Asia or are comparable to India in terms of per capita GDP at purchasing power parity (1995 constant international dollars), that is, have per capita GDP of $1,333­$2,333 (India's per capita GDP was $1,833 in 1995). underweight fell at an annual rate of more than 8 percent. In Bangladesh, despite economic growth that lagged behind that of India, the decline in the prevalence of underweight was greater (3.5 percent). Patterns and Trends in India The prevalence of underweight among children under 3 and recent trends in underweight vary substantially across demographic and socioeconomic groups in India (table 1.4). In 1998/99, the prevalence of underweight was much higher in rural areas (50 percent) than in urban areas (38 percent), and the differences were even larger for severe underweight, which affected 20 percent of rural children and 12 percent of urban children. Dimensions of Child Undernutrition in India · 13 Table 1.4 Prevalence of underweight and severe underweight in children under 3, by demographic and socioeconomic group, 1992/93­1998/99 Underweight Severe underweight Prevalence Prevalence Percentage Prevalence Prevalence Percentage Item 1992/93 1998/99 change 1992/93 1998/99 change Total 53 47 ­11 22 18 ­18 Urban 44 38 ­13 16 12 ­27 Rural 55 50 ­10 24 20 ­16 Quintile 1 (poorest) 61 59 ­4 30 27 ­8 Quintile 2 60 56 ­6 26 23 ­12 Quintile 3 56 52 ­6 23 21 ­7 Quintile 4 49 44 ­11 18 15 ­12 Quintile 5 (richest) 36 33 ­9 11 9 ­26 Female 52 49 ­6 21 19 ­11 Male 53 46 ­15 22 17 ­24 Scheduled castes 57 53 ­7 25 21 ­15 Scheduled tribes 57 56 ­2 29 26 ­9 Other castes 51 44 ­14 20 16 ­23 Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. Prevalence As expected, the prevalence of both underweight and severe underweight increases as household wealth falls, although at a decreas- ing rate. Underweight prevalence was as high as 60 percent in the low- est quintile, but it was so pervasive throughout the wealth distribution that even in the wealthiest fifth of the population 33 percent of chil- dren were underweight and 8 percent were severely underweight. The prevalence of both underweight and severe underweight was slightly higher among girls than boys (49 percent versus 46 percent for underweight, 19 percent versus 17 percent for severe under- weight). It was much higher among scheduled castes and scheduled tribes than among other castes. Thus children at greatest risk for underweight are girls whose fam- ilies are poor, belong to scheduled tribes or castes, and live in rural areas. Assuming independence of conditional probabilities, the chance that a girl with all these characteristics is underweight is as high as 92 percent (figure 1.5).5 14 · India's Undernourished Children Figure 1.5 Girls whose families are poor, belong to a scheduled tribe or caste, live in a rural area, and are at risk of being underweight 1 0.9 0.8 0.7 underweight 0.6 0.5 being of 0.4 0.3 0.2 probability 0.1 0 Girl Girl in Girl in Girl in (benchmark) scheduled tribe scheduled tribe, scheduled tribe, poorest quintile poorest quintile, rural area Source: Calculated from NFHS II (1998­9) data. The age pattern of undernutrition is an important dimension of the problem in India--and indeed all over the world. Growth retar- dation occurs early in life, and most of this early damage is irre- versible (ACC/SCN 2004). Most growth faltering occurs either dur- ing pregnancy (30 percent of children in India are born with low birth weight) or during the first two years of life.6 Indeed, by the age of 2, most growth retardation has already taken place (figure 1.6). Consequently, the period between pregnancy and the first two years of life represents the "window of opportunity" in which to address undernutrition. Efforts to fight undernutrition need to focus on this age group. Trends The prevalence of both underweight and severe underweight fell during the 1990s, but it fell less among segments of the population that were already more likely to be underweight in 1992/93. Conse- quently, over time, urban-rural, intercaste, male-female, and wealth inequalities in nutritional status widened (figure 1.7). The percentage reduction in severe underweight prevalence between 1992­93 and 1998­99 was dramatically higher in urban areas (26 percent) than in rural areas (16 percent). The reduction was also somewhat greater for underweight prevalence. Dimensions of Child Undernutrition in India · 15 Figure 1.6 By the age of 2, most of the damage from undernourishment has been done 0.5 0 ­0.5 scorez ­1 age for ­1.5 ­2 height ­2.5 ­3 0 3 6 9 12 15 18 21 24 27 30 33 36 age in months India Asia Africa Latin America & Caribbean Source: Regional estimates from Shrimpton and others (2001); India data from IIPS and Orc Macro (2000). Note: For the pattern of age-specific weight-for-age estimates, see figure A.1 in the appendix. Figure 1.7 Demographic and socioeconomic variation in prevalence of underweight children under 3, 1992/93 to 1998/99 70 0 3 60 ­5 50 under ­10 40 ­15 30 change ­20 % children 20 % 10 ­25 0 ­30 1 2 3 4 5 total urbanrural male tribe female caste caste quintilequintilequintilequintilequintile other scheduled scheduled Underweight 1998­9 (left axis) Underweight 1992­9 (right axis) Severe underweight 1998­9 (left axis) Severe underweight 1992­9 (right axis) Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. Note: Quintile 1 is the poorest quintile, quintile 5 the richest. By 1998/99 the percentage of underweight children in the bottom two wealth quintiles had fallen below 60 percent. However, reduc- tions in the percentage of malnourished children were lower in the lower quintiles than in the upper quintiles, indicating the growing health disparity between children of relatively low and relatively high 16 · India's Undernourished Children economic status. In fact, the greatest percentage reduction in the prevalence of underweight, especially severe underweight, accrued to children in the wealthiest quintiles. Between 1992/93 and 1998/99, underweight prevalence among boys fell 14 percent (from 53 percent to 46 percent), while under- weight prevalence among girls fell just 6 percent (from 52 percent to 49 percent). The effect was to reverse the underweight gender gap, so that, on aggregate in India, girls now lag far behind boys. The reversal was even more pronounced for severe underweight prevalence, which fell 24 percent (from 22 percent to 17 percent) for boys and 11 per- cent (from 21 percent to 19 percent) for girls. Despite the ostensible targeting of nutrition and health interven- tions to vulnerable castes, the percentage decline in underweight prevalence during the 1990s was smaller for scheduled castes and par- ticularly scheduled tribes. Among nonscheduled castes, the prevalence of underweight (and severe underweight) was reduced by 14 percent (23 percent) between 1992/93 and 1998/99. Over the same period, the prevalence of underweight (and severe underweight) among sched- uled caste groups declined just 7 percent (15 percent); among sched- uled tribes, the decline was just 2 percent (9 percent). Although underweight prevalence is widespread across India, just 10 percent of villages and districts accounted for 27­28 percent of all underweight children in the country. As few as a quarter of districts and villages accounted for more than half of all underweight children (World Bank 2004a) (figure 1.8). Figure 1.8 In 1998/99, more than half of all underweight children in India lived in just one- quarter of all villages and districts 100 90 80 70 60 underweight 50 40 30 children 20 % 10 0 0 10 20 30 40 50 60 70 80 90 100 cumulative % of villages or districts Districts Villages Source: World Bank 2004a. Note: Villages and districts are ranked by number of underweight children. Dimensions of Child Undernutrition in India · 17 The geographic concentration of underweight means that tailoring an appropriate response to malnutrition in a country as large and diverse as India requires a more richly textured picture of malnutri- tion patterns and trends than the national picture presented above. It also suggests that, where reliable data on malnutrition prevalence are available, actions to combat undernutrition could be targeted to a rel- atively small number of districts and villages. The rest of this section examines how the prevalence of and trends in underweight varied across states and across socioeconomic groups within states in 1992/93 and 1998/99. Since data from only two points in time are used, it cannot be assumed that these trends represent longer-term changes in undernutrition. Interstate and Within-State Variation Variation by State The prevalence of underweight and the extent to which it fell (or occasionally rose) during the 1990s varied widely across states (table 1.5). Underweight prevalence in Bihar and Madhya Pradesh fell from 60 percent to about 55 percent during the 1990s. As Table 1.5 Prevalence of underweight, 1992/93 and 1998/99, by state Below-average prevalence Above-average prevalence Item (less than 47 percent) (at least 47 percent) Increase in malnutrition Manipur (28; 4) Orissa (55; 4) Rajasthan (51; 14) Below-average reduction Gujarat (46; ­6) Madhya Pradesh (55; ­8) in malnutrition Haryana (35; ­2) Maharashtra (50; ­3) (0­11.6 percent) Himachal Pradesh (45; ­2) Tripura (50; ­6) Kerala (27; ­0.5) Uttar Pradesh (52; ­10) Mizoram (28; ­1) Above-average reduction Andhra Pradesh (38; ­20) Bihar (55; ­12) in malnutrition Arunachal Pradesh (25; ­35) West Bengal (49; ­14) (more than 11.6 percent) Assam (37; ­27 ) Delhi (35; ­16) Goa (29; ­16) Jammu and Kashmir (35; ­19) Karnataka (44; ­13) Meghalaya (38; ­15) Nagaland (24; ­14) Punjab (29; ­37) Tamil Nadu (37; ­22) Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. Note: First figure in parentheses refers to prevalence in 1998/99; second figure refers to the change in prevalence between 1992/93 and 1998/99. 18 · India's Undernourished Children a result, by 1998/99 no state in India had a malnutrition prevalence exceeding 60 percent. In six states--Bihar, Madhya Pradesh, Maharashtra, Orissa, Rajasthan, and Uttar Pradesh--however, at least half of children were underweight in 1998/99. A combination of large populations and high underweight prevalence means that four of these states--Bihar (11 percent), Madhya Pradesh (11 percent), Rajasthan (10 percent), and Uttar Pradesh (11 percent)--accounted for 43 percent of all under- weight children in India (World Bank 2004a). Most of these high- prevalence states also experienced the smallest reductions in the prevalence of underweight, with Orissa and Rajasthan registering sharp increases in underweight prevalence. Variation by Location In all states except Tripura, the percentage of underweight children was higher in rural areas than in urban areas (figure 1.9). The magnitude of these differentials varied. The largest percentage differences between rural and urban areas were observed in Jammu and Kashmir (81 percent), Punjab (78 percent), West Ben- gal (64 percent), and Delhi (61 percent). Although Manipur, Orissa, and Rajasthan were the only states that registered increases in total underweight prevalence between 1992/93 and 1998/99, Delhi regis- tered significant increases in the prevalence of rural malnutrition, and Figure 1.9 Urban-rural disparities in underweight among children, by state, 1992­9 70 80 60 60 40 50 20 underweight 40 0 30 change ­20 % 20 children ­40 of 10 ­60 % 0 ­80 AssamBiharGoa Kerala Orissa Nadu Delhi Pradesh Gujarat Haryana& Kashmir PradeshKamataka Pradesh Manipur Punjab Bengal MaharashtraMeghalayaNagaland Mizoram Pradesh Pradesh Tripura Rajasthan Tamilest W Uttar Andhra Himachal Jammu Madhya Arunachal Urban, 1998 (left axis) Rural, 1998 (left axis) Urban, 1992­8 (right axis) Rural, 1992­8 (right axis) Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. Dimensions of Child Undernutrition in India · 19 the northeastern states of Manipur, Meghalaya, Nagaland, and Tripura experienced increases in urban malnutrition.7 Variation by Gender At the national level, the prevalence of underweight among girls exceeded the prevalence of underweight among boys by more than 3 percentage points, and the rate of decline in the preva- lence of underweight among boys was about 2.3 times that of girls (see figure 1.7). This pattern of gender disparities did not characterize every state. Indeed, while the national trend was echoed in Assam, Bihar, Gujarat, Karnataka, Kerala, Madhya Pradesh, Meghalaya, West Bengal, and Uttar Pradesh, in other states, such as Goa, Jammu and Kashmir, Mizoram, Nagaland, and Tripura, the prevalence of under- weight fell more among girls than among boys. In the three states in which total underweight prevalence increased (Manipur, Orissa, and Rajasthan), the increase for both girls and boys was equal. In some states one gender has remained consistently disadvantaged relative to the other; in others gender disparities have worsened over time. In Delhi and Orissa, the percentage of underweight boys has been consistently higher than the percentage of underweight girls, while the reverse has been true of Punjab, Tamil Nadu, and West Bengal (table 1.6). In other states, such as Jammu and Kashmir, girls were in a worse position than boys in 1992/93 but not in 1998/99. In Assam, Bihar, Karnataka, Kerala, Madhya Pradesh, Rajasthan, and Uttar Pradesh, girls fared better than boys in 1992/93, but by 1998/99 they had lower nutritional status. Table 1. 6 Classification of states by change in gender differentials in prevalence of underweight Item States Percentage of underweight girls exceeds Andhra Pradesh, Gujarat, Haryana, percentage of underweight boys in both Manipur, Punjab, Tamil Nadu, 1992/93 and 1998/99. West Bengal Percentage of underweight boys exceeds Arunachal Pradesh, Goa, Delhi, Orissa, percentage of underweight girls in both Nagaland, Tripura 1992/93 and 1998/99. Percentage of underweight girls exceeds Assam, Bihar, Karnataka, Kerala, percentage of underweight boys in Madhya Pradesh, Meghalaya, 1998/99 but not 1992/93. Rajasthan, Uttar Pradesh Percentage of underweight boys exceeds Himachal Pradesh, Jammu and Kashmir, percentage of underweight girls in Mizoram 1998/99 but not 1992/93. Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. 20 · India's Undernourished Children Variation by Caste The national pattern in which the prevalence of underweight is highest among scheduled tribes, followed by scheduled castes and then other castes, obscures variations at the state level. Con- sistent with the national pattern, underweight prevalence in 1998/99 was higher among scheduled castes in Arunachal Pradesh, Himachal Pradesh, Jammu and Kashmir, Nagaland, and Tripura. But in Assam, Goa, and Manipur, the underweight prevalence was higher among other castes than among scheduled tribe and scheduled caste groups. Within each state, the trend in underweight prevalence varied dra- matically across castes. In Gujarat, Maharashtra, Tripura, and Uttar Pradesh, for example, the underweight prevalence of scheduled tribes increased while the underweight prevalence of other scheduled and nonscheduled castes declined. A similar pattern was observed for sched- uled castes relative to other castes in Himachal Pradesh and Kerala. Variation by Wealth With almost no exceptions, the prevalence of underweight, in both 1992/93 and 1998/99, was much higher among relatively poor households than among relatively well-off ones (figure 1.10).8 A troubling finding is that the aggregate reduction in the prevalence of underweight between 1992/93 and 1998/99 was smaller for the lowest tertile (poorest third) than for the upper tertile (richest Figure 1.10 Change in prevalence of underweight, by wealth tertile and state, 1992­8 80 120 70 100 80 60 60 50 40 underweight 40 20 change 30 0 % ­20 children 20 ­40 of 10 ­60 % 0 ­80 PradeshAssamBiharGujarat Kashmir Kerala Haryana Pradesh Manipur OrissaPunjab NaduBengal Kamataka Mizoram Nagaland Rajasthan Pradesh PradeshTripura and MaharashtraMeghalaya TamilWestUttar Andhra Madhya Jammu Arunachal Tertile 1, 1998 (left axis) Tertile 3, 1998 (left axis) Tertile 1, 1992­8 (right axis) Tertile 3, 1992­8 (right axis) Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. Note: Manipur data have very few observations in tertile 1 in 1992. Tertile 1 is the poorest, tertile 3 is the richest. Dimensions of Child Undernutrition in India· 21 third). This was true not only in aggregate but also in most states in India, indicating growing disparities in the prevalence of underweight among the well off and the not so well off. This trend was not univer- sal: in a few states (such as Assam, Tamil Nadu, and Tripura) the per- centage reduction in underweight prevalence among the lower tertile was much greater than among the upper tertile, indicating some nar- rowing of nutritional inequalities (table 1.7). Prevalence of Micronutrient Deficiencies The main micronutrient deficiencies in India are iron deficiency ane- mia, Vitamin A deficiency, and iodine deficiency disorders. Iron Deficiency Anemia Prevalence Although prevalence figures vary from study to study, there is no doubt that iron deficiency anemia is an extremely serious public health problem in India, especially among pregnant women and chil- dren. At least half of all ever-married women 15­49 and adolescent girls are believed to have some degree of iron deficiency anemia (IIPS Table 1.7 Wealth disparities in the change in underweight prevalence, by state, 1992/93 and 1998/99 Item States Growing intertertile nutritional inequalities as a result of malnutrition declined less in tertile 1 than Andhra Pradesh, Bihar, Madhya Pradesh, tertile 3 Nagaland, Punjab, West Bengal, Uttar Pradesh malnutrition increased in tertile 1 and Arunachal Pradesh, Gujarat, Jammu and declined in tertile 3 Kashmir, Maharashtra, Manipur malnutrition increased more in tertile Mizoram, Rajasthan 1 than tertile 3. Narrowing intertertile nutritional inequalities as a result of: malnutrition declined less in tertile Karnataka, Meghalaya, Tamil Nadu 3 than tertile 1 malnutrition increased in tertile 3 and Assam, Kerala, Tripura declined in tertile 1 malnutrition increased more in tertile 3 Orissa than in tertile 1. Source: Calculated from NFHS I (1992/93) and NFHS II (1998/99) data. 22 · India's Undernourished Children and Orc Macro 2000; Anand, Kant, and Kapoor 1999; Singh and Toteja 2003). One study shows that the prevalence of iron deficiency anemia among both pregnant and lactating women exceeds 75 percent and that more than half of pregnant women and a third of lactating women are moderately or severely anemic (NNMB 2002). In some states an anemia prevalence as high as 87 percent has been found among pregnant women from disadvantaged groups (IIPS and Orc Macro 2000; Seshadri 2001; Kapil and others 1999). Severe anemia from iron deficiency is believed to claim the lives of 22,000 women during pregnancy and childbirth each year (UNICEF 2003b). The prevalence of iron deficiency anemia among children is much higher than among adult women and may be partly attributable to the high prevalence of hookworm among children. The overall prevalence of anemia among children 6­35 months is 74 percent, with most suf- fering from mild (23 percent) or moderate (46 percent) anemia (IIPS and Orc Macro 2000). Prevalence among children 1­5 years is a little lower, but two-thirds of these children can be classified as anemic, with the majority suffering from moderate anemia (NNMB 2002). Trends Very little progress was made in reducing the prevalence of iron deficiency anemia between 1990 and 2000 (figure 1.11). Moreover, pop- ulation growth added 34.1 million non-pregnant and 2.3 million preg- nant anemic women during this time period (Mason, Musgrove, and Habicht 2003). Although the prevalence of iron deficiency anemia among preschool children fell somewhat, from almost 80 percent in 1990, it remained high, at about 75 percent, in 2000 (UNICEF and MI 2004b). By contrast, the prevalence of iron deficiency anemia in Bangladesh and Pakistan fell to 55 percent within the same period, and in China, the prevalence of iron deficiency anemia fell more than 60 per- cent (from more than 20 percent to the current level of 8 percent). Variation by Demographic and Socioeconomic Characteristics The prevalence of moderate iron deficiency anemia among children 6­35 months varies greatly by demographic and socioeconomic characteristics (fig- ure 1.12). It tends to be higher among children from disadvantaged groups--rural children, children living in poor households, and chil- dren from scheduled castes and tribes. The prevalence of mild anemia (about 23 percent) and severe anemia (about 5 percent) varies less with demographic and socioeconomic characteristics. There is almost no difference in the prevalence of iron deficiency anemia by gender. Dimensions of Child Undernutrition in India · 23 Figure 1.11 Trends in prevalence of iron deficiency in preschool children, by world region, 1990, 1995, and 2000 90 80 70 % 60 50 40 30 prevalence, 20 10 0 Sub-Saharan Middle East South India South China Latin Europe Total Africa & North Asia Asia America & Africa (without (without & Central India) China) Caribbean Asia 1990 1995 2000 Source: UNICEF and MI 2004b. The pattern of iron deficiency anemia among ever-married women 15­49 is similar to that among children, but the variation is larger. The total prevalence among women from scheduled tribes and the poorest 20 percent of the population, for example, was at least 10 percentage points higher than the national average of 52 percent. Iron deficiency anemia is a condition that afflicts not only the poor: more than 40 percent of women in the richest two quintiles were also anemic. Figure 1.12 Prevalence of anemia among children 6­35 months and women of reproductive age, by demographic and socioeconomic characteristics, 1998/99 90 75 % 60 45 30 prevalence, 15 0 1 2 3 4 5 1 2 3 4 5 total urbanrural femalemale castetribeother total urbanrural castetribeother quintile quintile quintile quintile quintile quintile quintile quintile quintile quintile scheduled scheduled scheduled scheduled children 6­35 months ever-married women 15­49 years Mild Moderate Severe Source: IIPS and Orc Macro 2000. 24 · India's Undernourished Children Interstate Variation The prevalence of iron deficiency anemia varies widely across states, among both children and ever-married women. While fewer than one-half of children in Kerala, Manipur, and Naga- land were anemic in 1998/99, more than 80 percent of children in Bihar, Haryana, Punjab, and Rajasthan were. The prevalence of child anemia was generally higher in states with a high prevalence of under- weight, although some states with a relatively low underweight preva- lence (such as Punjab and Sikkim, where fewer than one-third of chil- dren are underweight) had a surprisingly high prevalence of iron deficiency anemia (80 percent in Punjab and 77 percent in Sikkim). The variation in iron deficiency anemia prevalence among ever- married women was even greater, ranging from 23 percent in Kerala to 70 percent in Assam. Manipur (29 percent), Goa (36 percent), and Naga- land (38 percent) also had relatively low prevalence. By contrast, in seven states--Arunachal Pradesh, Assam, Bihar, Meghalaya, Orissa, Sikkim, and West Bengal--more than 60 percent of ever-married women were anemic. In some states, such as Arunachal Pradesh and Assam, the preva- lence of iron deficiency anemia among women was even higher than that among children under 3. (For figures on the prevalence of iron defi- ciency anemia among women and children disaggregated by state and severity of iron deficiency anemia, see appendix table A.2.) Vitamin A Deficiency Prevalence The prevalence of Vitamin A deficiency in India is one of the highest in the world, especially among preschool children, among whom 31­57 percent suffer from subclinical Vitamin A deficiency and another 1­2 percent suffer from clinical Vitamin A deficiency (UNICEF and MI 2004b; West 2002). India is home to more than one-fourth of the world's preschool children suffering from subclinical Vitamin A defi- ciency (35.4 million of 127.3 million) and one-third of preschool chil- dren with xerophthalmia (1.8 million of 4.4 million) (ACN/SCN 2004). Nationwide, Vitamin A deficiency is estimated to precipitate the deaths of more than 300,000 children a year (UNICEF and MI 2004a). Vitamin A deficiency is also prevalent among women of reproduc- tive age, among whom clinical symptoms of night blindness are extremely widespread. About 1 in every 20 pregnant women has subclinical Vitamin A deficiency, and almost 12 percent of them suf- fered from night blindness during their most recent pregnancy (West Dimensions of Child Undernutrition in India · 25 2002). An extremely high prevalence of maternal night blindness, coupled with a large number of pregnancies, means that about half of the world's pregnant woman with night blindness live in India (3 mil- lion of 6.2 million). As might be expected, the prevalence of night blindness is much higher in rural areas (14 percent) than in urban areas (6 percent) (IIPS and Orc Macro 2000). Trends Some progress has been made in reducing Vitamin A defi- ciency in India, but the prevalence of subclinical Vitamin A deficiency remains one of the highest in the world (figure 1.13). Prevalence fell rapidly in the early 1990s, to less than 60 percent among preschool children, but progress slowed in the second half of the 1990s. Recent estimates place the current prevalence at about 57 percent (UNICEF and MI 2004b; Mason and others 2003). Interstate Variation There is huge variation in the prevalence of Vita- min A deficiency among children across states. The incidence of vision problems can be used as an indicator of Vitamin A deficiency (figure 1.14).9 The number of children with vision problems is less than 10 per 1,000 children in several states and union territories, such as Gujarat and Punjab, but in many states in the North East, such as Assam, Manipur, Mizoram, Sikkim and Tripura, as well as in Goa, Jammu and Kashmir, and West Bengal, more than 30 per 1,000 chil- dren have vision problems (DWCD and UNICEF 2001). Figure 1.13 Changes in prevalence of subclinical Vitamin A deficiency among children under 6, by world region, 1990, 1995, and 2000 80 70 60 % 50 40 30 prevalence, 20 10 0 Sub-Saharan Middle East South India South China Latin Europe Total Africa & North Asia Asia America & Africa (without (without & Central India) China) Caribbean Asia 1990 1995 2000 Source: UNICEF and MI 2004b. 26 · India's Undernourished Children Figure 1.14 Proportion of children experiencing daytime and nighttime vision difficulties Daman & Diu Dadra & NagarHaveli Gujarat Punjab Uttar Pradesh Madhya Pradesh Rajasthan Nagaland Chandigarh Kamataka Andhra Pradesh Bihar Lakshadweep Himachal Pradesh Haryana Pondicherry Andaman & Nicobarlslands Delhi Maharashtra Arunachal Pradesh Tamil Nadu Kerala Orissa Meghalaya Tripura Sikkim Jammu & Kashmir West Bengal Manipur Assam Goa Mizoram 0 20 40 60 80 100 proportion of children (per 1,000) 0­4 years 5­9 years 10­14 years Source: DWCD and UNICEF 2001. Note: The variation in day-time and night-time vision difficulties across states is used as an indicator of the variation in Vitamin A deficiency. Dimensions of Child Undernutrition in India · 27 Iodine Deficiency Disorders Prevalence Although the prevalence of iodine deficiency disorders in India is lower than in most South Asian countries, the problem is ubiquitous and affects millions of people (figure 1.15). One survey shows that more than 85 percent of districts (241 of 282) are iodine deficiency disorder endemic (Ministry of Industry 2000). This places about 329 million people at risk, equivalent to one-third of India's population or one-sixth of the total global population at risk of iodine deficiency disorder. Among those who suffer from iodine deficiency disorder in India, 51 million are school-age children (6­12 years). One-third of all children in the world that are born with mental dam- age related to iodine deficiency disorder live in India (Ministry of Industry 2000; ACC/SCN 2004). Interstate Variation As with other vitamin and mineral deficiencies, the prevalence of iodine deficiency disorder varies widely across and within states. During the 1980s, 17 states and most hilly regions were identified as goiter endemic (Gopalan 1981). More recently, new endemic areas appear to have emerged in the plains (WHO 2000). According to a five-state study conducted in 2001, the prevalence of Figure 1.15 Prevalence and number of iodine deficiency disorders in the general population, by world region and country Sri Lanka Pakistan Nepal Maldives India Bhutan Bangladesh Oceania LAC Europe Asia Africa 100 80 60 40 20 0 0 300 600 900 1200 1500 prevalence of iodine deficiency disorder, % total number of people with iodine deficiency disorder, millions Source: ACC/SCN 2004. LAC = Latin America and the Caribbean. 28 · India's Undernourished Children iodine deficiency disorder ranged from 15 percent in Tamil Nadu to 46 percent in Karnataka. At the district level, the variation is even greater: for example, the East Godavari and Nellore districts of Andhra Pradesh, and the Kannur district of Kerala are effectively free of iodine deficiencies, while the prevalence is as high as 90 percent in the Shimoga district of Karnataka (WHO 2004a). Will India Meet the Nutrition MDG? The MDGs are a set of internationally agreed goals that countries and institutions have committed to reach by 2015. The first MDG is to eradicate extreme poverty and hunger. The second target of this MDG--halving the proportion of the population suffering from hunger between 1990 and 2015--uses two indicators to measure progress: the prevalence of underweight among children under 5 and the proportion of the population below a minimum level of dietary energy consumption. Several studies, using different assumptions, have considered the likelihood that India will attain the MDG target (see, for example, Wagstaff and Claeson 2004; Chhabra and Rokx 2004; World Bank 2004a).10 Although their projections differ, all of these studies con- clude that it is unlikely that the prevalence of malnutrition in India will fall from its level of 54 percent in 1990 to 27 percent by 2015 (World Bank 2004a).11 National Family Health Survey (NFHS) data show that in 1998/99, even the wealthiest quintile had a prevalence of malnutrition (33 percent) that far exceeded the MDG target. This report's projections indicate that economic growth alone is unlikely to be sufficient to lower the prevalence of malnutrition. When combined with policy interventions, the projections are rosier, but a rapid scaling-up of health, nutrition, education, and infrastructure interven- tions is needed if the MDG is to be met (World Bank 2004a). Effect of Economic Growth Alone The effect that India's economic growth in the coming decade will have on the prevalence of malnutrition in 2015 can be projected using estimates of the responsiveness (elasticity) of malnutrition to annual economic or income growth. The magnitude of these elasticities Dimensions of Child Undernutrition in India · 29 should ideally be calculated from household surveys (Haddad and oth- ers 2003), provided that they include appropriate income or expendi- ture data. In the absence of these data, an alternative is to assume a rule-of-thumb elasticity and test its sensitivity. Two assumptions are made in order to estimate the effect that eco- nomic growth will have on the prevalence of underweight. The first is that India's economy will grow at an annual rate of 3 percent, the average rate between 1990 and 2002 (World Bank 2004b). The second is that the income elasticity of underweight is 0.51 (Mkenda 2004). This means that a 1 percent increase in per capita GDP leads to a 0.51 percent reduction in the prevalence of underweight. Under these assumptions, the prevalence of underweight among children under 3 falls to 39 percent by 2015 (table 1.8 and figure 1.16). Under a more generous average annual per capita growth rate of 5 percent, prevalence falls to 36.3 percent--still short of the MDG target. Even under an unrealistically generous income elasticity assumption of 0.7, prevalence falls only to 35 percent (under the assumption of 3 percent growth) or 30 percent (under the assumption of 5 percent growth). Under the assumption that the prevalence of underweight in 2002 has fallen somewhat since 1999 (for example, by 1 percent a year to 43 percent), the change in the predicted prevalence is greater, but it still remains far in excess of the 27.4 percent mark. Only when an exceptional average annual per capita economic growth rate of 8 percent is assumed does underweight fall low enough to reach the MDG target. This sensitivity analysis shows that the con- clusion that economic growth alone will not enable India to meet the MDG target is robust to a wide range of assumptions. Table 1.8 Under all likely economic growth scenarios, India will not reach the nutrition MDG without direct nutrition interventions Estimated prevalence of Prevalence of underweight among underweight in Income elasticity children under 5, given various average 2002 (percent) of malnutrition annual per capita GDP growth rates (percent) 3 percent 5 percent 8 percent 43 0.51 35 31 47 0.51 39 36 27.3 47 0.3 41 39 47 0.7 35 30 Note: See appendix table A.1 for calculations. 30 · India's Undernourished Children Figure 1.16 Predicted prevalence of underweight under different economic growth scenarios, 2002­15 50 45 2062 40 2037 underweight 35 % 2023 30 2015 25 2002 2015 3% per capita GDP growth (0.3 elasticity) 5% per capita GDP growth (0.51 elasticity) 3% per capita GDP growth (0.51 elasticity) 8% per capita GDP growth (0.51 elasticity) MDG target Source: World Bank calculations. Note: Boxed years at the right of the graph denote the predicted date that the MDG target will be met. Effect of Economic Growth Plus an Expanded Set of Interventions Projections from a recent World Bank study (World Bank 2004a) combine economic growth assumptions and policy interventions. They show that even if poor states were brought up to the national average in terms of sanitation, road access, electricity, medical atten- tion at time of delivery, female schooling, household income (con- sumption), and public spending on nutrition per child, the cumula- tive reduction in the national prevalence of underweight would be only about 8 percentage points (or 15 percent). If the magnitude of the proposed interventions were scaled up to bring the poor states up to the average level prevailing in the nonpoor states, the cumula- tive reduction in the prevalence of underweight rate would be 21 percentage points, or 38 percent--still short of the MDG target. Only when seven specific interventions are pursued simultaneously is the prevalence of child underweight in the poor states expected to fall 25 percentage points--enough for them to reach the target fig- ure (figure 1.17).12 Dimensions of Child Undernutrition in India · 31 Figure 1.17 Projected percentage of children under 3 in poor states who are under- weight, under different intervention scenarios, 1998­2015 60 % 50 40 underweight, 30 of 20 10 prevalence 0 1998 2015 expanding medical attention at birth increasing real government expenditure on child nutrition expanding medical access to sanitation real income growth expansion in regular electricity supply expansion of female schooling increasing access to rural roads MDG Source: World Bank 2004a. Conclusions The problem of undernutrition in India is of alarming magnitude and great complexity. The prevalence of underweight is among the high- est in the world--nearly twice that in Sub-Saharan Africa--and the pace of improvement lags behind what might be expected given India's economic growth. Modest progress has been made in reducing undernutrition over the past decade, but most of this progress was driven by improvements among higher socioeconomic groups. Even if India comes close to achieving the nutrition MDG in 2015 (which it most likely will not), it will still have levels of undernutrition equiva- lent to those that exist in Sub-Saharan Africa today (Shekar and others 2004). Aggregate levels of undernutrition are extremely high, and significant inequalities across states and socioeconomic groups appear to be grow- ing. Girls, children in rural areas, children from the poorest households, and children from scheduled tribes and castes are the worst affected. In 32 · India's Undernourished Children Bihar, Madhya Pradesh, Maharashtra, Orissa, Rajasthan, and Uttar Pradesh, more than half of all children are underweight. Thus, while undernutrition is a national problem, the problem is more acute among certain groups. Immediate action needs to be taken to address malnutri- tion, using strategies that take into account local variations in nutritional status and the fact that certain demographic and socioeconomic groups are more vulnerable to malnutrition than others. CHAPTER 2 The Integrated Child Development Services Program Are Results Meeting Expectations? India's primary policy response to child malnutrition, the Integrated Child Development Services (ICDS) program, is well conceived and well placed to address the major causes of child undernutrition in India. Some important mis- matches between its policy intentions and actual implementation are preventing it from reaching its full potential, however. More attention has been given to increasing coverage than to improving the quality of service delivery; too few children under 3, for whom malnutrition prevention is most critical, are being reached; and too much emphasis is being placed on distributing food rather than changing family-based feeding and care behavior. The program also faces substantial operational challenges. The ICDS has expanded tremendously over its 30 years of operation to cover almost all development blocks in India. It offers a wide range of health, nutrition, and education services to children, women, and ado- lescent girls. The program was intended to target the poorest, the most undernourished, and the age groups that represent a significant "win- dow of opportunity" for nutrition investments (that is, children under 3 and pregnant and lactating women). There is a mismatch, however, between the program's intentions and its actual implementation: · The central focus on food supplementation drains financial and human resources from other tasks envisaged in the program that are crucial for improving child nutritional outcomes. For example, not 33 34 · India's Undernourished Children enough attention is given to improving childcare behaviors or edu- cating parents about how to improve nutrition. · Older children (3­6 years) participate much more than younger chil- dren, and many children from poorer households do not yet partici- pate. The program fails to preferentially target girls, children from lower castes, or children from the poorest villages, all of whom are at higher risk of undernutrition. · Although program growth was greater in underserved than well- served areas during the 1990s, the poorest states and those with the highest levels of undernutrition still have the lowest levels of pro- gram funding and coverage by ICDS activities. In addition to these mismatches, the program faces substantial opera- tional challenges. Inadequate worker skills, shortages of equipment, poor supervision, and weak monitoring and evaluation reduce the program's potential impact. Community workers are overburdened because they are expected to provide preschool education to 4- to 6-year-olds as well as nutrition services to all children under 6. As a result, most children under 3--the group that is most vulnerable to malnutrition--do not receive micronutrient supplements, and most of their parents are not reached with counseling on better feeding and childcare practices. Successful interventions have taken place in many districts, and innovations and variations in the ICDS have occurred in several states. These successes, described in the last section of this chapter, suggest that the potential for better implementation and greater impact exists. How ICDS Aims to Address the Causes of Persistent Undernutrition With strong government commitment and political will, the ICDS program has emerged from a small pilot in 1975 to become India's flagship nutrition program. Many of the ICDS program components are well designed to address the immediate causes of child undernutri- tion in India, although significant shifts in focus and improvements in implementation will be necessary if the program is to realize that potential. The Integrated Child Development Services Program · 35 A Conceptual Framework of the Causes of Undernutrition Child undernutrition is a consequence of complex interactions among multiple determinants. These interactions can be conceptualized using a framework that traces the causal pathways of undernutrition through different levels: the most immediate, the underlying, and the basic causes (figure 2.1). Immediate Causes of Undernutrition The most immediate causes of mal- nutrition are inadequate dietary intake and infections, which create a vicious cycle that is responsible for much of the high morbidity and mortality among children in developing countries. When children do not consume enough, their immune response is lowered, rendering them more susceptible to infectious diseases. Ill children deplete their nutritional stores and are in poor health because of reduced intake, Figure 2.1 Causes of child malnutrition Source: UNICEF 1998. 36 · India's Undernourished Children poor absorption of nutrients, and the increased demands of combating disease (Esrey and others 1990; Scrimshaw and SanGiovanni 1997; Allen and Gillespie 2001). Over the past decade, a large body of work has documented the interaction between nutrition and infection. Evidence of the malnu- trition-infection syndrome was first reported in studies conducted in Guatemala and India. These studies found that children developed diarrheal infections around the time of weaning from breastmilk and that they were subsequently more prone to infections and growth fal- tering (Gordon and others 1964; Scrimshaw and others 1968). While the weight loss associated with a single episode of infection can be made up if the diet is adequate, recurrent episodes of infection with- out sufficient food or inadequate recovery time are primary causes of poor growth among children in developing countries (Schürch and Scrimshaw 1989). Following infection, a number of weeks pass before the child's weight returns to the pre-onset level, retarding the child's growth. In the case of diarrhea, the degree of growth deficit has been shown to be proportional to the number of days ill (Martorell and others 1975). If infections are frequent, high rates of underweight prevail even when food intake is adequate. The converse is also true: if infections are less common or less severe, lower rates of child under- nutrition prevail even if average food intake is low. Thus, sufficient food intake is only one determinant of nutritional status. Underlying Causes of Undernutrition The two immediate causes of malnutrition, poor dietary intake and infection, are closely linked to the three underlying determinants of nutritional status: household- level access to food, availability of health resources (such as preventive and curative health care and clean water and sanitation), and the appropriateness of the childcare and feeding behaviors that caregivers adopt.1 Household-level food security refers to physical and economic access to foods that are socially and culturally acceptable and of suffi- cient quality and quantity. Macro-level food security (that is, sufficient food production at national or regional levels) does not necessarily ensure household-level security, which is determined by a more com- plex array of factors than agricultural production, including local prices (of food and other goods), income, and an effective trade and The Integrated Child Development Services Program · 37 transport infrastructure (Bouis and Hunt 1999). Moreover, household food security is not in itself sufficient to ensure that the nutritional needs of every child, and adult, living in a particular household will be met. Within each household, decisions are made as to the quantity and quality of food that is allocated to each household member. This decision is affected by a complex range of factors, including the rela- tive bargaining power of household members (which in turn may be related to their income, autonomy, gender, and education), as well as other characteristics, such as the health status of individual members. Consequently, the diets of individual children (or others) within the household may be deficient even though per capita caloric intake is high and the household is food secure. Overcrowding, congestion, a shortage of clean water, and inade- quate facilities for the disposal of human excreta, wastewater, and solid waste contribute to the development of gastrointestinal infec- tions such as diarrhea, and facilitate the spread of infectious disease. This explains why mortality rates in urban areas exceeded those of rural areas before the sanitation revolution but were lower than rural rates after it (Collins and Thomasson 2002). Crowding has been shown to be associated with an increased risk of infectious intestinal disease (due to rotavirus group A) in children (Sethi and others 2001) and tuberculosis infection (MacIntyre and others 1997). Poor water quality, a limited quantity of water, poor exc- reta disposal practices, and poor food hygiene are all associated with an increased prevalence of diarrhea in infants (Esrey and others 1990; Moe and others 1991). Clean water, good sanitation, and hygienic conditions at the community level generate important externalities for individual households in the community: clean water and good sanita- tion at the neighborhood level have been shown to have a positive effect on the height of children in a household, whether or not the household itself has a healthy environment (Alderman and others 2003).2 The presence of infection, particularly communicable disease, is a direct cause of malnutrition. Consequently, efforts to prevent expo- sure to infection and cure disease should be central to any strategy aimed at combating malnutrition. These efforts include regular deworming, the use of bed nets in malaria areas, oral rehydration therapy, and access to regular and affordable health check-ups. 38 · India's Undernourished Children Providing appropriate care can mitigate the impact of the malnutri- tion-infection cycle for vulnerable groups, such as children and pregnant and lactating women. Such care requires adoption of childcare and feed- ing behaviors that direct available resources toward promoting child nutritional well-being. Adequate care during pregnancy and delivery can reduce the incidence of maternal death, miscarriage, stillbirth, and low birth weight among infants. Adequate feeding of young children (initia- tion of breastfeeding within an hour of birth, exclusive breastfeeding for the first six months of life, and adequate and timely complementary feed- ing starting at six months while continuing to breastfeed) is critical for child growth. Caregivers' time, their knowledge and educational status, autonomy, control over monetary and other resources, and capacity to make appropriate decisions are often the key factors that determine whether these behaviors are adopted. Basic Causes of Undernutrition The framework shown in figure 2.1 links these underlying determinants to a set of basic determinants, includ- ing the availability of human, economic, and organizational resources with which to improve nutrition. Use of these resources is shaped by how society is organized in terms of economic structure; political and ideological expectations; and the institutions through which activities and resources within society are regulated, social values are met, and potential resources are converted into actual resources. The Design of the ICDS Program and the Underlying Causes of Child Undernutrition The ICDS program is potentially well poised to address some of the underlying causes of persistent undernutrition. The program adopts a multisectoral approach to child well-being, incorporating health, edu- cation, and nutrition interventions (table 2.1), and is implemented through a network of anganwadi centers at the community level. These centers range from open-air spaces to anganwadi workers' homes to one- or two-room stand-alone buildings. The Department of Women and Child Development's emphasis on a life-cycle approach means that malnutrition is fought through interventions tar- geted at unmarried adolescent girls, pregnant women, mothers, and children from birth to 6 years. Services provided include health Table 2.1 Range of services that the ICDS seeks to provide to children and women Children under 6 Pregnant women Lactating women Health check-ups, Health check-ups by Antenatal check-ups Postnatal check-ups and treatment AWW, ANM, LHW Treatment of diarrhea Deworming Basic treatment of minor ailments Referral for more severe illnesses Growth monitoring Monthly weighing of under-threes Quarterly weighing of 3- to 6-year-olds Weight recorded on growth cards Immunization Immunization against Tetanus toxoid poliomyelitis, immunization diphtheria, pertussis, tetanus, tuberculosis, and measles Micronutrient IFA and Vitamin A IFA supplementation supplementation supplementation for malnourished children Health and nutrition Advice includes infant Advice includes education feeding practices, infant feeding child care and practices, child development, care and utilization of health development, services, family utilization of planning, and health services, sanitation family planning, and sanitation Supplementary Hot meal or Hot meal or Hot meal or ready- nutrition ready-to-eat snack ready-to-eat snack to-eat snack providing 300 providing 500 providing 500 calories and 8g­10g calories and calories and protein 20g­25g protein 20g­25g protein Double rations for malnourished children Preschool Early Childhood Care education and Preschool Education consisting of "early stimulation" of under-threes and education "through the medium of play" for children aged 3­6 years Source: DWCD 2004a. 39 40 · India's Undernourished Children check-ups, treatment and referral for infants and children, growth monitoring, immunization, micronutrient supplementation, supple- mentary feeding, preschool education for children 3­6, and health and nutrition education for adult women. As the program has devel- oped, it has expanded its range of interventions to include compo- nents focused on adolescent girls' nutrition, health awareness, and skills development, as well as income-generating schemes for women. ICDS and the World Bank Total government expenditure on ICDS has grown significantly since the program's inception. Following expenditure of about 1,190 Indian rupees (Rs1,190) crores (1 crore equals 10 million) during its first 17 years (1975­92), the government increased spending from Rs2,271 crores under the Eighth Five-Year Plan (1992­7) to Rs4,557 crores under the Ninth Five-Year Plan (1997­2002) (DWCD 2005). The Tenth Five-Year Plan (2002­7) allocates Rs10,391 crores to the pro- gram. In addition, the program has been supported by several donors, including UNICEF, the Swedish International Development Cooper- ation Agency, the World Food Programme, Care, and the Norwegian Agency for Development Cooperation. The World Bank has supported efforts to improve nutrition in India since 1980 through six projects. With an investment of $712.3 million in the sector, India accounts for the largest share of Bank Group lending devoted specifically to nutrition programs. Support to ICDS has been provided in three overlapping phases: · In Phase I the Bank supported the Tamil Nadu Integrated Nutrition Project (TINP) as an alternative to the standard ICDS program in the state of Tamil Nadu (TINP I, 1980­9; TINP II, 1990­7). · In Phase II support was extended to the standard government ICDS programs, as well as to some additional activities (ICDS I in Orissa and Andhra Pradesh, 1991­7; ICDS II in Bihar and Madhya Pradesh, 1993­2000). · In Phase III the primary emphasis moved from expanding coverage to improving the quality of services (through an ICDS component in the Andhra Pradesh Economic Restructuring Program, 1999­2004, and the Woman and Child Development Project, 1999­2004).3 The Integrated Child Development Services Program · 41 Empirical Findings on the Impact of ICDS The ICDS program has been the subject of a large volume of research. Most evaluations have focused on the quality of infrastructure and inputs and the execution of activities. Few rigorous studies have evalu- ated the program's impact on nutritional status or health behaviors, partly because few data sources permit outcomes among program par- ticipants and non-participants to be compared. Consequently, most researchers have been unable to use the statistically rigorous method- ologies that would enable them to draw reliable conclusions about the impact of ICDS. Some studies have found that the program is associ- ated with improvements in nutritional status, while others have failed to find a positive effect. It is not clear to what extent the failure to reach consensus is the result of inadequate survey design and poor data qual- ity. In the future, to be sure of measuring the impact accurately, it will be necessary to collect data on treatment and control populations, preferably over at least two time periods. The major national-level study of program impact (NIPCCD 1992) found that the prevalence of underweight was lower among children in areas in which the ICDS program was in place, for both children under 3 and children 3­6.4 Because of the small sample sizes of the control and treatment groups, however, both these differences are statistically insignificant (Lokshin and others 2005). Three recent studies have estimated the association between having an anganwadi center in a village and the likelihood that a child is underweight. All three find little or no association between the pres- ence of a center and child nutritional status. Using multivariate analy- sis of the 1992/93 National Family Health Survey (NFHS) data, the World Bank (2004a) estimates that, for boys, having an anganwadi center is associated with a 5 percent reduction in the likelihood of being underweight but that there is no significant association for girls. Using both the 1992/93 and the 1998/99 NFHS data, Lokshin and others (2005) initially find that ICDS appears to have a significant and positive effect on nutritional outcomes, but on more rigorous explo- ration, using propensity score matching techniques, they find little significant effect when children in ICDS villages are compared with children with similar demographic, household, and village character- istics in non-ICDS villages. In a multivariate model of cross-sectional 42 · India's Undernourished Children data collected in Kerala, Rajasthan, and Uttar Pradesh between 2000 and 2002, Bredenkamp and Akin (2004) find that children in villages with anganwadi centers are not significantly less likely to be under- weight or ill than other children. Using data from Chhattisgarh, Ker- ala, Madhya Pradesh, Maharashta, Rajasthan, and Uttar Pradesh, they find that only in Kerala is actual attendance at an anganwadi center significantly associated with better nutritional status. There is little evidence that ICDS has been successful in attaining its goal of improving the coverage of specific child health interven- tions, such as deworming and Vitamin A supplementation, or encour- aging mothers to adopt appropriate childcare and feeding behaviors (including practices related to breastfeeding, weaning, and diet) that have the potential to improve child growth and health outcomes. Data from Kerala, Rajasthan, and Uttar Pradesh show no clear evidence that these behaviors were more common in ICDS areas; only in Maharashtra was an association found (Bredenkamp and Akin 2004) (table 2.2). Although communication for behavior change through the anganwadi worker is a crucial weapon in the fight against poor health and malnutrition, it appears that the information the anganwadi worker is conveying to mothers is not being communicated effectively enough to positively affect mothers' behavior. Targeting of ICDS Program and Beneficiaries Geographical Targeting: Placement of Programs across States and Villages The percentage of administrative blocks covered by ICDS has reached almost 90 percent (see appendix table A.3). The percentage of children who actually take up the services provided by the program is lower, however, and varies significantly across states (figure 2.2). By December 2002, only one-quarter of all Indian children between the ages of 6 months and 6 years were benefiting from the supplementary nutrition component of ICDS, with the figure ranging from little more than 10 percent in some states to more than 90 percent in oth- ers.5 Coverage is particularly high in the northeastern states. ICDS policy stipulates that one anganwadi center should be in place per 1,000 population, with more intensive placement of 1 per 700 The Integrated Child Development Services Program · 43 Table 2.2 Comparison of intermediate health outcomes and behaviors across children living in villages with and without an anganwadi center In villages Kerala Maharashtra Rajasthan Uttar Pradesh Percentage over 6 months No AWCs 81.2 80.5 29.8 18.0 receiving Vitamin A with AWCs 78.3*** 88.5*** 22.5*** 21.0*** supplementation Percentage older than 12 No AWCs 61.1 34.3 3.7 17.7 months ever dewormed with AWCs 66.3*** 59.7*** 4.1 13.3*** Percentage over 6 months No AWCs 78.1 78.1 27.6 36.0 consuming Vitamin A­rich with AWCs 72.0*** 90.5*** 26.9 32.5*** food within previous 3 days Percentage breastfed No AWCs 85.6 54.4 9.4 6.1 within 1 hour of delivery with AWCs 80.0*** 41.2*** 10.3 6.7 Percentage consuming No AWCs 98 8.9 74.1 53.4 colostrum with AWCs 96.9*** 28.7*** 80.4*** 37.3*** Percentage under 6 months No AWCs 67.1 21.5 38.4 99.7 who are exclusively with AWCs 58.2*** 11.3*** 43.3* 84.6*** breastfed Percentage aged 6­9 No AWCs 84.1 67.3 93.8 0.3 months consuming with AWCs 87.7 73.6 93.7 19.1*** complementary food Mean duration of No AWCs 13.4 16.3 8 23.7 breastfeeding, among with AWCs 12.5*** 17.4*** 7.1*** 22.8*** children who have been weaned (in months) Source: Calculated from ICDS III baseline/ICDS II endline survey 2000­2. Notes: * statistically significant at the 10% level; ** 5% level; *** 1% level; AWC = anganwadi center. For clarity, boldface indicates where outcomes are significantly better in villages with AWCs. population in tribal areas, where poverty tends to be more prevalent. In practice, ICDS centers are much more numerous in wealthier states (figure 2.3). States with lower per capita net state domestic product have a smaller percentage of villages covered by the ICDS program than those with higher per capita net state domestic product. The growth of program coverage from 1992 to 1998 was more rapid in the poorest villages, however (Lokshin and others 2005). Regardless of the indicator of ICDS coverage used (percentage of vil- lages with a center, number of ICDS beneficiaries, public expenditure on ICDS), access to the program appears to be worst in the poorest states and in the states with the worst nutrition indicators (figure 2.4). The five states with the highest underweight prevalence (Rajasthan, Uttar Pradesh, Bihar, Orissa, and Madhya Pradesh) rank in the bottom 10 in terms of ICDS coverage. 44 · India's Undernourished Children Figure 2.2 The percentage of children 6 months to 6 years enrolled in the supplementary nutrition program, 2002, varies widely across states Source: Department of Women and Child Development enrollment data (updated 2004); Cen- sus of India (2001). Note: Figures are calculated from Department of Women and Child Development data on the number of children between the ages of 6 months and 6 years who were beneficiaries of the Supplemental Nutrition Program in December 2002 and from population data for children under 6 in 2001. The use of different age categories may result in a slight underestimation of the per- centage of beneficiaries, while the use of population data from 2001 may result in a slight over- estimation of the percentage of beneficiaries. The magnitude and direction of the bias are hard to predict. The Integrated Child Development Services Program · 45 Figure 2.3 ICDS coverage is higher in states with higher per capita net domestic product Source: Coverage calculated from NFHS II (1998/99) data; net state domestic product data are from Indiastat.com. Note: Data are in current prices for 1998/99. Village-level data reveal that ICDS placement is less regressive within than across states. In 1998, for example, while ICDS was in place in only half of the villages in the lowest two deciles of the all- India wealth distribution, the program covered about 80 percent of the richest villages in India. The difference in program coverage between the poorest and the wealthiest villages within each state was much smaller: about 60 percent of the poorest villages in every state Figure 2.4 In many states in which the prevalence of underweight is high, the proportion of villages with anganwadi centers is low Source: Underweight prevalence calculated from NFHS II (1998­9); appendix table A.3. 46 · India's Undernourished Children were covered by the ICDS program, compared with 70 percent of the wealthiest villages (Lokshin and others 2005). The percentage of children enrolled in the ICDS program tends to be smaller in states with a higher percentage of underweight children (figure 2.5). Enrollment is lowest in Bihar (1.5 percent), where the underweight prevalence is 55 percent. At the other end of the spec- trum, Manipur, Mizoram, Nagaland, and Sikkim exhibit an under- weight prevalence that is among the lowest in India (20­30 percent) but are among the five states with the highest percentage of ICDS beneficiaries. The clear exception to this pattern is Orissa, which has a very high underweight prevalence (47 percent) but has enrolled at least 95 percent of children in the program. The states in which the prevalence of malnutrition is highest are also the states that receive the least funding from the central govern- ment and the smallest financial allocations from the state govern- ments for ICDS. Government per child expenditure in support of states' ICDS programs appears to be strongly and inversely propor- tional to the states' underweight prevalence. In addition, the (per child) amount allocated by state governments to ICDS--most of which is spent on the supplementary feeding component--is lowest in the states with the highest underweight prevalence and highest in the states with the lowest underweight prevalence. Total public expenditure figures show that four of the Figure 2.5 Fewer children are enrolled in ICDS in states in which the prevalence of underweight is high Source: Calculated from NFHS II (1998/99), DWCD (2003), and 2001 Census of India data. The Integrated Child Development Services Program · 47 states that rank in the top five for underweight prevalence (Bihar, Uttar Pradesh, Rajasthan, and Madhya Pradesh) are also the four states that receive the least funding for ICDS, on a per child basis.6 This regressive relationship holds true at the other end of the spec- trum, too, where the five largest per child allocations are made to and by the five states that have the lowest underweight prevalence (figure 2.6). Since poorer states find it difficult to mobilize resources for ICDS, the government of India has recently proposed providing addi- tional central financing to all states to cover half of the cost of the sup- plementary nutrition component. Individual Targeting: Characteristics of Beneficiaries Effective targeting restricts nutrition interventions to those individu- als or groups that are most vulnerable to malnutrition. In so doing, it maximizes the social returns and minimizes costs. However, the high generalized malnutrition prevalence in India and the administrative costs associated with excluding those who are relatively well-off means that rigorous targeting of ICDS benefits to particular socioe- conomic groups is unlikely to prove feasible. Instead, ICDS policy Figure 2.6 Public expenditure by state and national governments is very low in states in which the prevalence of underweight is very high Source: Calculated from NFHS II (1998/99) and DWCD (2003). 48 · India's Undernourished Children Figure 2.7 Older children are more likely than younger children to attend an anganwadi center Source: ICDS III baseline/ICDS II endline survey 2000­2. Note: Data show percentage of children in villages with anganwadi centers who attend a cen- ter at least once a month. follows the general guideline that a "special effort" should be made to reach children from lower-income families or scheduled tribes and castes. There is also some explicit targeting of severely malnourished children, who are supposed to receive double food rations. This section examines whether children who are most in need of the ICDS program have access to its services and use them on a regu- lar basis. It presents the findings of a survey on children's attendance at anganwadi centers in Chhattisgarh, Kerala, Madhya Pradesh, Maharashtra, Rajasthan, and Uttar Pradesh during 2000­2 (hence- forth referred to as the ICDS III baseline/ICDS II endline survey).7 The data are disaggregated by age, gender, caste, household wealth, and location. Targeting by Age Early childhood is a crucial developmental period, dur- ing which there is considerable scope to influence the growth of mal- nourished children. However, it is precisely this group of children-- infants and children under 3--that is least likely to attend the anganwadi center. Attendance is lowest among the youngest children, increasing steadily--sometimes dramatically--until the age of 3, after which it remains more or less constant (figure 2.7). In Kerala and Maharashtra, almost every child 4­6 in the sample attended the center at least once a month. Attendance rates were less than half of that in the other four The Integrated Child Development Services Program · 49 Figure 2.8 The caste and tribe composition of children attending anganwadi centers varies somewhat across states Source: ICDS III baseline/ICDS II endline survey 2000­2. Note: Data show percentage of children in villages with anganwadi centers who attend a cen- ter at least once a month. states. When daily, rather than monthly, attendance figures are exam- ined, the gap between the attendance rates of children under 3 and chil- dren 4­6 is much larger (see appendix figure A.2). Targeting by Gender Neither daily nor monthly attendance figures reveal a statistically significant difference in the participation rates of boys and girls. There appears to be no gender discrimination in the reach of ICDS services. Targeting by Caste The ICDS scheme places special emphasis on the participation of children of lower castes. Some anganwadi centers have been constructed in close proximity to scheduled caste and scheduled tribe colonies, and anganwadi workers are expected to take steps to encourage the recruitment of these children into the program. In all states, attendance rates of children from scheduled castes and tribes are in line with or slightly higher than those of children from other castes (figure 2.8). In Chhattisgarh, Madhya Pradesh, and Maharashtra, the percentage of children from scheduled tribes attend- ing a center is higher than any other caste, while in Kerala, Rajasthan, and Uttar Pradesh, the percentage of children from scheduled castes is higher than that of children from other castes. These data are sup- ported by qualitative evidence of high take-up among scheduled tribes relative to forward castes, perhaps partly because of the social stigma 50 · India's Undernourished Children associated with the receipt of benefits among the upper castes (Educa- tional Resource Unit 2004). Caste composition differs from center to center, with attendance by children of a particular caste apparently influenced by the caste of the anganwadi worker and the caste that is most dominant in the local community. Targeting by Household Wealth Among children living in villages with anganwadi centers, remarkably little variation is found in participation rates across wealth quintiles: within each state, there is not much more than a 10 percentage point difference across wealth quintiles (figure 2.9). This implies that a poor economic background does not present too formidable an obstacle to ICDS attendance. But since poorer chil- dren are more likely to be malnourished, it is desirable that ICDS attracts a larger share of lower quintile than upper quintile children. Maharashtra is the only state in which attendance declines steadily as wealth increases. In Chhattisgarh and Uttar Pradesh, attendance is slightly lower in the top quintile; in Kerala and Madhya Pradesh, atten- dance is more regressive, with higher attendance rates in the upper quintiles. A similar picture is obtained when one examines daily atten- dance figures: with the exception of Maharashtra, the percentage of upper quintile children attending centers is either as high as or higher Figure 2.9 The percentage of children who attend anganwadi centers varies only slightly across wealth quintiles Source: ICDS III baseline/ICDS II endline survey 2000­2. Note: Data show percentage of children in villages with anganwadi centers who attend a cen- ter at least once a month. Quintile 1 is the poorest quintile, quintile 5 is the richest. The Integrated Child Development Services Program · 51 than the percentage of lower quintile children (see appendix table A.4 for figures). These state-level enrollment figures may obscure low enrollment among economically disadvantaged children in specific villages. Field visits to Uttar Pradesh, for example, found that the poorest of the poor were frequently excluded from ICDS interventions and under- represented at anganwadi centers (Educational Resource Unit 2004). Targeting by Urban-Rural Location There is much heterogeneity across states in attendance rates of children living in urban, rural, and tribal areas (figure 2.10). In Chhattisgarh and Madhya Pradesh, for example, attendance rates are highest in urban areas, followed by tribal areas, while in Kerala and Uttar Pradesh attendance rates are highest in rural areas. Summary Although large proportions of vulnerable groups are indeed taking up the ICDS benefits for which they are eligible, there is sub- stantial program capture by the less needy--possibly at the expense of more vulnerable children. Attendance by lower castes is relatively high, but there is still scope to attract a greater percentage of this group. Additional effort needs to be made to reach younger children and children from poor households, who are not only underrepre- sented at anganwadi centers but also at greatest risk for malnutrition.8 Figure 2.10 Attendance at anganwadi centers varies widely both across and within states Source: ICDS III baseline/ICDS II endline survey 2000­2. 52 · India's Undernourished Children Characteristics and Quality of ICDS Service Delivery Promoting growth and providing supplementary food are central to the ICDS objective of reducing the prevalence of malnutrition. This section examines the delivery of these services, especially with respect to the availability of equipment and supplies and the frequency with which these services are delivered. It also looks at the quality of angan- wadi center infrastructure, the training and competencies of angan- wadi workers, and the coordination between the ICDS and the Repro- ductive and Child Health Program. Promoting Growth Growth-monitoring activities are hampered by poor access to appro- priate equipment, such as scales, growth cards, and wall or book charts. Equipment is often nominally present but not of sufficient quantity or quality. Anganwadi centers in Kerala and Madhya Pradesh are generally better equipped than those in Chhattisgarh, Maharash- tra, and Uttar Pradesh, although they, too, suffer equipment shortages (figure 2.11). Even in centers with working scales, many workers report that they do not weigh children under 3 every month. In all states, growth-monitoring performance appears to be superior in tribal areas, where children are weighed with greater frequency. Figure 2.11 Percentage of anganwadi centers with growth-monitoring equipment in place Source: ICDS III baseline/ICDS II endline survey (2000­2). The Integrated Child Development Services Program · 53 Anganwadi centers in urban and tribal areas are better equipped with weighing equipment than rural centers. Even with regular weighing, growth monitoring is effective only if accompanied by communication for behavior change that results in improved growth of the malnourished child. Previous studies of ICDS have noted that this does not often occur, perhaps because many anganwadi workers are not fully competent in interpreting growth cards and curves (Gopalan 1992) or because anganwadi workers fail to effectively communicate the meaning of children's growth patterns to mothers (Vasundhara and Harish 1993). Indeed, the ICDS III base- line/ICDS II endline survey reveals a very large discrepancy between the child's measured weight and the mother's subjective assessment of her child's growth status. In Kerala, all mothers think their children are experiencing normal growth; in Uttar Pradesh, where under- weight prevalence in the ICDS III baseline/ICDS II endline sample is 46 percent, 94 percent of women describe their children's nutritional status as normal. Providing Supplementary Nutrition The Supplementary Nutrition Program is one of the best-known ICDS interventions. Food is financed and procured by the states and provided to children at the center, either in the form of a ready-to-eat snack or a meal cooked by the anganwadi worker. Many children receive food at the center, with state averages ranging from about 20­80 children per center, depending on the center's location. In addition, in most states there is a take-home food component, from which about 20­25 children per center benefit. Despite the resources and energy devoted to it, the supplementary nutrition program appears to perform poorly, especially in terms of providing a regular supplementary source of nutrition to the needy while simultaneously excluding the non-needy. Irregularities in the food supply and leakage to non-targeted individuals are major prob- lems (table 2.3). The most commonly reported reasons why children do not receive supplementary food from the anganwadi center relate to inadequacies on the supply side. In decreasing order of importance, these include lack of availability of food for distribution, lack of awareness of the food program among mothers or a failure to realize that their children 54 · India's Undernourished Children Table 2.3 Regularity of food supply to AWCs and the availability of the take-home food program Uttar Madhya Kerala Maharashtra Pradesh Pradesh Chhattisgarh Percentage of AWCs with no recent irregularities in 60 41 68 27 17 food supply Percentage of AWCs with a take-home food program 15 28 42 95 75 Source: ICDS III baseline/ICDS II endline survey 2000­2. are eligible for the program, failure of the anganwadi worker to con- tact mothers or children when food is available, and the distance of the anganwadi center from children in need.9 These findings strongly suggest that ICDS needs to improve the regularity of the food supply. Indeed, in three of the five states sur- veyed in 2000­2, the majority of anganwadi centers reported irregu- larities in their food supply during the preceding three months. Another evaluation reported that 27 percent of anganwadi centers experienced disruptions in food distribution for periods of more than 90 days (NIPCCD 1992). There is also some evidence that household attitudes and behaviors are important determinants of children's access to ICDS food. Some mothers think that their children do not need the food (even though the same children have been assessed by researchers as malnourished). Other mothers fail to collect the food from the anganwadi center, sometimes because their families prohibit them from doing so. Large- scale household surveys reveal negligible complaints about food quality or quantity (Bredenkamp and Akin 2004), but field visits have shown that food is sometimes badly cooked, dry, and salty (Educational Resource Unit 2004) and should be supplemented by sugar, rice, or vegetables, perhaps procured locally, to be more palatable to children. Leakage of supplementary food to nontargeted beneficiaries appears to be widespread. In many states, attendance rates among children from relatively wealthy households are higher than those among chil- dren from relatively poor households. In practice, there appears to be little targeting of children from disadvantaged groups for supplemen- tary feeding or of malnourished children for double rations of supple- mentary food. Food is often distributed to all those who come to the The Integrated Child Development Services Program · 55 center (Educational Resource Unit 2004). Where the center is located on school premises, food is distributed to grade 1 children as well as preschool children, so that the number of beneficiaries often exceeds the number of children actually enrolled at the center. As a result, chil- dren often receive less than the recommended 300 kilocalories of food. In some instances, food is also distributed to indigent adults, and it is common for anganwadi helpers, and occasionally anganwadi workers, to take cooked food home (Educational Resource Unit 2004). There is substantial leakage in the take-home food component of ICDS, since many children share the food with siblings or elders. In Madhya Pradesh, for example, only about a third of children consume all take-home food themselves. One-third of children consume less than a quarter of the food, and 6 percent consume none of the food taken home from the center (Bredenkamp and Akin 2004). Most anganwadi workers surveyed describe the take-home food component as "not useful." The supplementary nutrition program is effective as an incentive to attract children to the centers, where they can then receive other health- and nutrition-related services; without the program, atten- dance at the centers might be much lower. Community-based moni- toring mechanisms have recently been introduced in some areas in an attempt to improve the delivery of supplementary nutrition. Providing a Safe and Hygienic Environment for ICDS Service Delivery Growth promotion, the provision of supplementary food, and the delivery of other ICDS services are sometimes performed in unsafe or unhygienic environments. Most centers in urban areas (but not those in rural areas) are located in rented buildings (table 2.4), espe- cially community buildings, such as primary schools, religious cen- ters, and panchayat buildings. While potentially improving commu- nity scrutiny of ICDS, use of these buildings may render the regular functioning of the center vulnerable to the competing purposes for which these buildings are used. Moreover, because the budgetary allocation to rent is low, anganwadi centers may be found in small or unclean locations. Some ICDS centers are run out of the homes of ICDS functionaries. About one-third of anganwadi centers in India have pucca (brick and mortar­type construction buildings), another third have semi-pucca 56 · India' s Table 2.4 Anganwadi center (AWC) infrastructure, by location Undernourished Kerala Maharashtra Uttar Pradesh Madhya Pradesh Chhattisgarh Urban Rural Tribal Urban Rural Tribal Urban Rural Urban Rural Tribal Urban Rural Tribal Percentage of AWCs with drinking water that is Children piped or pumped 69 44 50 21 44 41 54 70 100 58 83 73 83 72 open well 27 41 17 0 20 34 0 8 0 0 0 0 4 0 other 4 15 33 79 36 25 46 22 0 42 17 27 13 28 Percentage of AWCs with toilets that are flush 27 15 0 0 2 0 8 7 50 19 14 36 13 16 pit-latrine 20 26 0 0 13 10 29 10 8 15 7 9 9 4 none 53 59 100 100 85 90 63 84 42 65 79 55 78 80 Percentage of AWCs with rented building 64 41 50 96 19 41 92 15 92 46 21 82 17 44 Number of AWCs in sample 45 27 6 24 54 29 24 61 12 15 29 11 23 25 Source: ICDS III baseline/ICDS II endline survey 2000­2. The Integrated Child Development Services Program · 57 construction, fewer than one-third are in kutcha buildings (buildings constructed with low-quality materials, such as unburned brick, bam- boo, thatch, or mud roofing); a handful of centers function in open spaces, such as under trees (NCAER 2001). Cooking space is typically inadequate, as reported by 55 percent of anganwadi workers across the country (NCAER 2001). Most anganwadi centers have no toilet facili- ties, especially in rural and tribal areas.10 Among centers with toilets, flush toilets are more common in urban areas and pit-latrines are more common in rural and tribal areas. The majority of anganwadi centers obtain their drinking water from a tap or hand pump, but the water source varies substantially across state and rural-urban-tribal location. Worker Training, Workload, and Status The skills of the anganwadi worker and her capacity to mobilize the community to support ICDS and recruit participants, especially the most vulnerable, are central to good-quality service delivery and effectiveness. Too often performance is constrained by poor training and the pressure of a large and diverse workload. Skills Training Anganwadi workers tend to be well educated, but they are often poorly trained for ICDS tasks. Survey data show that almost all have at least matriculated high school, and half of those in urban areas have received some college education. Pre-service training is rare, however, with most women undergoing only short-term in-service training (Bredenkamp and Akin 2004). Recently, more resources have been directed toward strengthening capacity at the central, state, and block levels to provide high-quality support and training to functionar- ies of ICDS programs. In 2002, a new training program, Udisha ("first rays of the new dawn"), was initiated, with funding from the World Bank. This program has attempted to shift the focus of training away from the mere transfer of knowledge toward the strengthening of worker competencies. Workload, Status, and Remuneration Anganwadi workers can spend up to 40 percent of their time on supplementary nutrition­related activities and another 39 percent on preschool education (NCAER 2001). This leaves little time for other important ICDS activities, such as growth promotion, health and nutrition education, home visits, referral services, and meeting with the community. In addition, anganwadi workers must 58 · India's Undernourished Children maintain at least 10 different types of records.11 Anganwadi workers are also often given other responsibilities outside of ICDS. When anganwadi centers are located on school premises, for example, some workers have the additional responsibility of teaching class (grade) 1 (Educational Resource Unit 2004). In some communities, anganwadi workers are required to meet family planning and sterilization targets. Anganwadi workers are also called on to assist in other government programs for women and children, such as the Pulse Polio campaign. Home visits--to advise on antenatal care and promote breastfeeding, timely immuniza- tion, and regular weighing--appear to be one of the more neglected of ICDS tasks, with only 78 percent of anganwadi workers in Maharashtra, 68 percent in Chhattisgarh, 43 percent in Madhya Pradesh, 38 percent in Uttar Pradesh, and 35 percent in Kerala undertaking the equivalent of at least one home visit a day (Bredenkamp and Akin 2004). Despite the importance of their work, anganwadi workers are often held in low regard by the community (Educational Resource Unit 2004), viewed as "mere" providers of child care rather than valuable healthcare workers. There are also frequent lags in payment of honoraria. Accord- ing to ICDS III baseline/ICDS II endline survey 2000­2, as many as two-thirds of urban anganwadi workers in Uttar Pradesh report that they do not receive their honoraria regularly. The low status the community attaches to the position of anganwadi worker, and the irregularity with which workers are paid reduces workers' motivation. Collaboration between ICDS and the Reproductive and Child Health Program The objectives of the Reproductive and Child Health Program and ICDS are intertwined; the promotion of linkages between the activi- ties of the two programs would therefore be mutually beneficial. Already some of these linkages are recognized in the job descriptions of anganwadi workers and auxiliary nurse-midwives. Anganwadi work- ers are supposed to promote awareness of national immunization days and maintain immunization records, refer sick children to healthcare facilities, and encourage mothers to seek antenatal care. Auxiliary nurse-midwifes, employed by the Department of Health, are sup- posed to conduct general health check-ups of ICDS beneficiaries, give immunizations, dispense medicines and contraceptives, and provide The Integrated Child Development Services Program · 59 assistance and guidance to anganwadi workers in the discharge of their health-related duties. In practice, cooperation between the ICDS and the Reproductive and Child Health Program appears to be limited, partly because of the absence of a designated person or body to oversee the promotion of this collaboration. Site visits reveal that anganwadi workers take little interest in finding out whether mothers are registered with the auxiliary nurse- midwife and receiving antenatal care, and the ICDS III baseline/ICDS II endline survey (2000­2) shows that visits to anganwadi centers by auxil- iary nurse-midwifes are not very regular. In Kerala, for example, only 50 percent of urban centers and no rural centers had received a visit from an auxiliary nurse-midwife the previous month (Bredenkamp and Akin 2004).12 As a result, it is perhaps not surprising that some anganwadi workers, and as many as one-third of those surveyed in rural Uttar Pradesh, are inclined to believe that the auxiliary nurse-midwife does not perform significant services during her visits. The fact that the provision of health services is not consistently better in villages with anganwadi centers than in villages without them seems to suggest that there is little coordination or convergence between the two. Deworming is more fre- quent in villages with anganwadi centers in Kerala and Maharashtra but not in Rajasthan and Uttar Pradesh. More children receive Vitamin A supplementation in villages with anganwadi centers than without angan- wadi centers in Maharashtra and Uttar Pradesh but not in Kerala or Rajasthan. Although the immunization function is being performed with some regularity (at least 80 percent of anganwadi centers in Chhattis- garh, Kerala, Madhya Pradesh, and Maharashtra have immunization registers that have been regularly used), previous studies suggest that ICDS has had little to do with any improvements in immunization cov- erage (see for example, Kulkarni and Pattabhi 1988). Although many centers face problems, some are overcoming them to provide very valuable services to their communities. One example is the center in the Bellary district of Karnataka (box 2.1). Monitoring and Evaluation A strong monitoring and evaluation system helps program managers track whether project implementation is proceeding as desired and 60 · India's Undernourished Children Box 2.1 Getting things right in the Bellary District of Karnataka: A report from the field Venkatamma, an anganwadi worker, is quick to list the characteristics of a good center.* "It should be a spacious place with clean surroundings, the building should have good ventilation, enough play materials and teaching aids, a mirror for the chil- dren to come and have a look, a small garden in front of the center, and they should be received with love," she says. She pauses and then continues with a grin, "Of course, most of these things are not there in my center, but children attend regularly in good numbers." According to Venkatamma, it is the relationship with the children, a good preschool component, and food that attract children to the center. Venkatamma and her helper, Rankamma, belong to scheduled castes and live close to the center. The center has its own building, with a 12' × 20' classroom, a storeroom, and a kitchen. There are enough vessels for cooking and serving; the water tank is very close to the center, although supply is erratic and water sometimes has to be fetched from a bore-well nearby. A toilet has recently been built, although no one has yet used it. By and large, Venkatamma's pride in her center was validated by a site visit to the facility. Forty-seven children were present when the Bank team visited, unan- nounced. By about 10:30 in the morning, the children trooped in, some marching in confidently, others brought in crying by grandmothers or older siblings. The center's staff weigh the children regularly, mark their weight in registers, explain to mothers how the children's growth is progressing, and make suggestions on how to increase their growth. Sometimes, two adolescent girls from the village help run the center. Venkatamma and Rankamma work well together, and the entire community appre- ciates and respects them. Women often visit the center to informally interact with them. Mothers were able to describe pregnancy risks and how children should be breastfed. It seems as if the center has acquired a status on a par with the school, where parents send their children regularly. The center follows a program determined for the week by the state-level authori- ties. All children are made to wash their hands before they eat; in other anganwadi centers in the same village, they even use soap to do so. The children are constantly reminded not to touch the floor or dirty their hands before eating. Venkatamma reported that in her 14 years of service she had never experienced any major gaps in the supply of food, that there was always something for the children to eat. If the sup- ply of rice were delayed, there would be sprouted green-gram or energy food ready for the children. This was confirmed by mothers. Venkatamma and the health unit coordinate well with each other. She refers prob- lem health cases to the health center, and many mothers now voluntarily bring their children there. Mothers take their children to the anganwadi center on immunization days, with the result that immunization coverage is good. *Names have been changed. Source: Educational Resource Unit 2004. The Integrated Child Development Services Program · 61 make informed decisions to correct any problems. Periodically, it allows an assessment to be made of the extent to which the program is having the desired impact. In so doing, monitoring and evaluation promotes the most effective and efficient use of resources. Some notable accomplishments have occurred in monitoring and evaluation in recent years. The current system nevertheless faces many challenges. Given the size of the ICDS program, monitoring and evaluating is a daunting task. A standardized data collection procedure is employed in all states, but it is complex and for the most part relies on manual entries and compilations. Each anganwadi worker maintains as many as 10­25 dif- ferent registers into which information is entered, some of it on a daily basis.13 Once a month, the anganwadi worker compiles this information into a standardized monthly progress report that contains a number of input, process, and impact indicators. These monthly progress reports are then sent to supervisors (each of whom supervises about 20 centers), who consolidate the reports and forward them to the child development project officers, who assemble reports by project-block and remit them to the state headquarters and central ICDS monitoring cell. At the central level, some of the key indicators are analyzed, and quarterly progress reports are prepared for the World Bank­funded states.14 These reports are used by the Depart- ment of Women and Child Development, the Planning Commission, the Health and Family Welfare department, and other departments. States are ranked with respect to progress made, and detailed feedback is sent to state headquarters. However, no feedback is conveyed from the state headquarters to lower levels of program implementation, so that local action is seldom taken in response, thus rendering the feed- back system rather ineffective. In light of the important role that an effective monitoring and eval- uation system can play in improving child health, strengthening the monitoring and evaluation system is essential. There have been some significant improvements in some states, in part due to the commit- ment and effort of the government of India and in part due to the presence of bilateral and international agencies, such as Care, the World Food Programme, and the World Bank.15 The major impedi- ments that remain must be addressed. 62 · India's Undernourished Children Low Prioritization of Activities Too little emphasis is placed on monitoring and evaluation, in part due to a poor understanding of what it entails and its potential contribution to program effectiveness. The primary focus of program management (at both the central and state level) seems to be on the timely release of allocations to implementing agencies and the recording of expendi- tures; very little emphasis is placed on assessing the quality of service delivery and the impact of the program. At the local level, few angan- wadi workers are aware of the purpose and utility of data collection; they view their data collection tasks as routine, boring, and burden- some. The result is that although the ICDS program is being moni- tored--in the sense that information on inputs and outputs is regularly collected--the system is not oriented toward using that information to inform action, that is, it is not used to enhance service delivery, improve beneficiary recruitment, or, eventually, modify program design. Consequently, there have been delays and bottlenecks in the replenishment of supplies, the neediest beneficiaries are often not reached, and it is difficult to know which elements of the program are most effective. Lack of Adequate Personnel The number of qualified people assigned to monitoring and evaluat- ing ICDS is relatively small at almost all levels of program implemen- tation, and those involved usually handle other tasks as well. Overall responsibility for monitoring ICDS rests with the highest positions in the government (at the director or secretary level), but these officials oversee many other programs as well and face severe time constraints. Vacancies in monitoring and evaluation positions are also a problem, with many positions remaining unfilled for extended periods and fre- quent personnel turnover at senior levels--a phenomenon that is common throughout the Indian bureaucratic system. At the field level, positions are more stable, but vacancies and irreg- ular supervision are pervasive. In the sample of blocks included in the ICDS III baseline/ICDS II endline survey (2000­2), supervisors had been appointed to all urban anganwadi centers in the sample and were fairly active in ICDS activities (with at least 96 percent of the angan- wadi centers in five of the six states reporting that they had been vis- The Integrated Child Development Services Program · 63 ited by supervisors the preceding month).16 However, 10 percent of all rural anganwadi workers were not linked to a supervisor.17 Moreover, many supervisors did not visit regularly: at least 30 percent of the rural anganwadi centers that had supervisors in Chhattisgarh and Uttar Pradesh had not been visited by them during the previous month (Bredenkamp and Akin 2004). A monitoring and evaluation curriculum is included in the training syllabuses for field-level ICDS functionaries, but the value of monitoring and evaluation and the importance of collecting data on key project indicators are typically not adequately communicated. Inadequate Use of Information Systems and Qualitative Data The information system, which is central to keeping track and mak- ing sense of the huge quantity of data collected, is held back by insufficient use of computer networks. Almost all information col- lected by anganwadi workers, supervisors, and child development project officers and forwarded to the state level is transmitted by hand, with very limited use of computers. Software programs are seldom used to analyze the data collected at the state and central level, except in some of the states covered under World Bank ICDS projects. Lack of computer hardware remains a problem up to the district-block levels, partly due to inadequate financial allocations to monitoring and evaluation. There is also an inherent quantitative bias in the monitoring sys- tem, which comes at the expense of the collection of some qualitative information that could assist in the construction of the causal narra- tives that explain patterns in the quantitative data. Continuous social assessments, which collect qualitative information through commu- nity meetings, focus groups, and open-ended questionnaires, are cur- rently being implemented in the states supported by the World Bank, but they are not used in other states. If ICDS is to substantially reduce child malnutrition, program managers need a reliable, broad-based, and efficient monitoring and evaluation system that enables them to adjust elements of program implementation and design in order to maximize the returns to nutri- tion investments. Chapter 3 examines some ways in which the current system could be improved. 64 · India's Undernourished Children Lessons from Successful Innovations There is encouraging evidence that, with relatively small changes in project priorities and design, the impact of the ICDS program on child nutritional status could be substantially enhanced. This can be seen in studies of the successful implementation and performance of regular ICDS projects as well as in studies of projects that have exper- imented with modifications to the ICDS program (see, for example, SIDA 2000 and Johri 2004). Adapting the lessons learned from these projects and applying them to other ICDS projects can help ensure that the ICDS has the maximum impact. Achieving Synergy with the Reproductive and Child Health Program and Using Community Members as Agents of Change: Lessons from INHP II Care India's Integrated Nutrition and Health Project II (INHP II), now active in nine states, reveals the benefits of targeting behavior change interventions at children under 2 and pregnant women, that is, concentrating energies on those critical periods in the life cycle when the greatest impact on health status can be made (Care India 2004).18 The program promotes closer convergence between the ICDS pro- gram of the Department of Women and Child Development and the Reproductive and Child Health Program of the Department of Health and Family Welfare, and encourages mothers to use reproduc- tive and child health services.19 The underlying premise of conver- gence is that by working together these programs are more likely to achieve their shared objectives of reducing infant mortality, combat- ing child malnutrition, and improving the health status of women. An example of this is the facilitation of well-publicized nutrition and health days, during which the anganwadi worker (from ICDS) and auxiliary nurse-midwife (from the Reproductive and Child Health Program) provide immunizations to children under 2 and antenatal care (including check-ups, iron and folic acid supplementation, and tetanus toxoid immunization) to pregnant women at the anganwadi center. Health talks are another important element of these days; take-home rations of supplementary food (sufficient for a few weeks) are provided as an incentive for attendance. The process of setting up the nutrition and health days is facilitated by the community, by The Integrated Child Development Services Program · 65 engaging mothers groups, self-help groups, and panchayati raj institutions. Another key INHP activity is the appointment and training of "change agents" within the community. Volunteers assigned to fami- lies provide health and nutrition information, promote positive health behaviors, and encourage ICDS participation. Volunteers can be women, men, adolescent girls or boys, or traditional birth attendants, each serving 10­15 families. These agents begin their activities at the birth of the child, if not before, when they advise on appropriate new- born care. They follow up with regular home visits until the child is 2. Many of these visits are timed to coincide with critical periods in the life cycle (for example, weaning). They serve as cues to action at times when mothers should initiate new health behaviors in order to protect their children against undernutrition and disease. The INHP approach appears to be having a significant effect. Fifty-three percent of pregnant women in the intervention areas received three or more antenatal checkups, compared with 38 percent in the nonintervention areas. Other aspects of antenatal care, such as consumption of iron and folic acid tablets and receipt of tetanus tox- oid doses, were also better in the intervention areas (see appendix table A.5). Childcare practices improved substantially, with 65 percent of women in the intervention areas initiating breastfeeding within one hour of delivery, compared with 38 percent in the non-intervention areas. Higher proportions of children in the intervention areas received Vitamin A supplementation and were breastfed exclusively for six months, were introduced to complementary feeding appropri- ately, given more nutritious complementary foods, and vaccinated against measles by the age of 12 months (see appendix table A.6). There appears to be no difference in behavior by children's gender. Some of the greatest differences between intervention and noninter- vention areas are found among people of low socioeconomic status, indicating that this intervention is progressive in its reach. Using Community-Based Interventions: Lessons from the Dular Program The Dular program, undertaken by state governments in Bihar and Jharkhand, with the assistance of the United Nations Children's Fund (UNICEF), has developed several innovative approaches to improving 66 · India's Undernourished Children early childhood nutrition, care, and development.20 Active in 8 of 60 districts, it focuses on intensive upgrading of ICDS operations, includ- ing the collection of birth weight data and the monitoring of care prac- tices. The program has creatively addressed many of the past failings of the ICDS program in Bihar. As part of the strategy, the anganwadi worker in every targeted vil- lage teams up with a small group of local resource people, who are given basic training in nutrition, child care, and hygiene. Once trained, the team visits pregnant women and mothers of newborns in their homes to educate them about safe delivery, breastfeeding, immu- nization, and other essential care practices during pregnancy and early childhood. Since the team is made up of local people from the com- munity, parents respond positively. Though still young, Dular appears to be having an impact. An eval- uation of 450 households indicates that after one year of intervention there was an 8 percent decline in the prevalence of underweight among children under 3, a 20 percent increase in the use of colostrum feeding within one hour of birth, a 20 percent decline in episodes of diarrhea in children under 3 during the three months before the inter- view, and a 30 percent increase in the consumption of adequately iodized salt by participating families (Saiyed and Srivastava 2005). Setting Up Mothers Committees: Lessons from Andhra Pradesh In 1998 the state of Andhra Pradesh began establishing mothers com- mittees in ICDS villages as a means of integrating ICDS into the community and stimulating demand for improved service quality. Mothers committees are informal committees of eight village mem- bers, established in line with the guidelines of the general ICDS pro- gram, which requires the formation of a mahila mandal (women's group). The groups are registered as committees in order to allow for- mal participation in ICDS and to enhance their legitimacy and accountability. Members serve three-year terms. Currently, more than 50,000 committees have been established in 351 development blocks in Andhra Pradesh.21 Committee members are given three rounds of week-long capacity-building training courses that focus on nutrition, health, education, group formation, and economic empowerment, as well as relevant and state-specific social and legal issues. In collaboration with the state AIDS control The Integrated Child Development Services Program · 67 society, 20,000 mothers committee members and 10,000 adolescent girls have been trained to serve as "change agents" in promoting HIV awareness and healthy sexual attitudes and behaviors. The roles and responsibilities of these committees with respect to the ICDS program have evolved considerably over time. Originally, they were involved in the civil works components of the World Bank­assisted ICDS I project (selecting construction sites for angan- wadi centers, monitoring construction, and releasing funds to cover construction costs). More than 15,000 anganwadi buildings were com- pleted under the supervision of mothers committees. Today the range of responsibilities includes recruiting anganwadi workers and helpers, paying honoraria, monitoring community-based performance indica- tors for anganwadi centers, establishing local food units to prepare and distribute supplementary food to the anganwadi centers, and ensuring that potential beneficiaries receive services. Mothers committee mem- bers may also play an active role in motivating adolescent girls to join bridging courses and skills development programs; encouraging school enrollment, especially among girls who have dropped out; and moti- vating parents to send children to anganwadi preschool. Evaluation of the mothers committees indicates that the program has potential but needs reinforcing. Only 40 percent of committees are formally involved in the ICDS program, and only 31 percent of all mothers report having heard of the committees. Awareness of the committees is higher in tribal areas (49 percent of women and 34 per- cent of adolescent girls) than in rural areas (25 percent of women and 15 percent of girls) and urban areas (20 percent of girls). A survey of anganwadi workers reveals that the mothers committees are very much appreciated, with three-quarters of respondents describing the functioning of the committees as "good" and another 11 percent as "satisfactory" (World Bank 2003). To increase the impact of the committees on maternal and child health and nutrition, it has been proposed that their role as change agents be strengthened through further training. Such training would help them promote appropriate infant feeding practices and atten- dance at anganwadi centers. Another way of increasing the role of the mothers committees as change agents would be to empower the committees to manage aspects of the ICDS system rather than simply helping program staff promote healthy behaviors. Subject to the external monitoring of the 68 · India's Undernourished Children outcomes they achieve, such responsibilities could include organizing food distribution, appointing anganwadi workers, and improving anganwadi center infrastructure. This effort has sought a much more ambitious role for community participation than the INHP II and Dular programs. Those efforts hinged on involving community members as behavioral change agents. The Andhra Pradesh program tried to involve mothers com- mittees in the actual management of ICDS resources--overseeing civil works and releasing funds for construction costs, managing food preparation and distribution, and recruiting and monitoring angan- wadi workers. To carry out these tasks effectively, community mem- bers need leadership training, support, and supervision, as well as clear designation of power. These requirements need to be explicitly built into the program design. To perform their tasks effectively, program staff need to know exactly what is expected of them, and they need to be supported in the execution of their tasks by supervisors to whom they can turn for advice and who monitor their activities. In addition, it is important that the tasks that participants are expected to perform not change erratically over time. Shifting expectations, combined with lack of authority and project support, can make it difficult for community members to play an active role in program implementation. Targeting High-Risk Groups: Lessons from TINP This variation of the regular ICDS program limited itself to a rela- tively small number of interventions targeting high-risk groups (Heaver 2002; World Bank 1998). Project activities included regular growth monitoring, nutrition education, and health check-ups for all children. Therapeutic supplementary feeding was provided to moder- ately and severely malnourished children, children whose growth was faltering (especially children under 3), and high-risk pregnant and lac- tating women. The TINP also placed more emphasis than the regular ICDS on training workers, building supervision and managerial capacity, and creating an efficient management and information system. Informa- tion was analyzed and fed back into project implementation. For example, after it was discovered that families were not changing the way they fed children under 2, the project targeted more of its infor- The Integrated Child Development Services Program · 69 mation and education to parents of young children. These efforts were successful, as mothers who took part in the project knew much more about good nutrition and health practices than other mothers, they breastfed longer, and fewer of their children needed supplemen- tary feeding. Community participation was also substantially enhanced. Staff were encouraged to develop active and close collaboration with local women's and girls' groups from the community to effect behavior change in the community. Community members were taught to pro- mote birth weight recording, regular monthly weighing, and spot feeding. They were also encouraged to participate in community assessment, analysis, and problem solving. The TINP halved the prevalence of severe malnutrition in the vil- lages in which it was implemented (Heaver 2002). It showed that uni- versal feeding was not necessary to achieve substantial nutritional and health gains. The program did not, however, fully meet its objective of reducing moderate malnutrition. The project evaluation concluded that to reduce moderate malnutrition, TINP interventions must focus more on home-based actions and proactive integration of nutrition activities with the health system. CHAPTER 3 Enhancing the Impact of ICDS Urgent changes are needed to bridge the gap between the policy intentions of ICDS and its actual implementation. In particular, the three main mismatches need to be resolved, so that the program addresses the most important determi- nants of malnutrition, effectively encourages the participation of younger chil- dren and the most vulnerable segments of the population, and reaches areas in which the prevalence of undernutrition is highest. ICDS was designed to address the multidimensional causes of under- nutrition. As the program expands to reach more and more villages, it has tremendous potential to improve the nutritional and health status of millions of women and children. The key constraint on the program's effectiveness is the fact that implementation has not followed the original design. Increasing emphasis has been placed on providing supplementary feeding and preschool education to children 4­6, at the expense of other compo- nents that are crucial for combating persistent undernutrition. Because of this, most children under 3--the group that suffers most from malnutrition--are not being reached, and most of their parents are not receiving counseling on better feeding and childcare practices. Realizing the potential of ICDS will require realigning its imple- mentation with its original objectives and design. Several steps are needed: · Ambiguity over the priority of different program objectives and interventions needs to be clarified immediately. 71 72 · India's Undernourished Children · Activities need to be refocused on the most important determinants of malnutrition. This means emphasizing disease control and prevention activities, education to improve domestic childcare and feeding prac- tices, and micronutrient supplementation. Greater convergence with the health sector, in particular the Reproductive and Child Health program, would help tremendously in this regard. · Activities need to be better targeted toward the most vulnerable age groups (children under 3 and pregnant women), and funds and new projects need to be directed to the states and districts with the high- est prevalence of malnutrition. · Supplementary feeding activities need to better target those who need them most, and growth-monitoring needs to be performed with greater regularity, with an emphasis on using it to help parents understand how to improve their children's health and nutrition. · Communities need to be involved in implementing and monitoring ICDS, in order to bring additional resources into the anganwadi cen- ters, improve the quality of service delivery, and increase accounta- bility in the system. · Monitoring and evaluation activities need strengthening through the collection of timely, relevant, accessible, high-quality information-- and this information needs to be used to improve program function- ing by shifting the focus from inputs to results, informing decisions, and creating accountability for performance. Mismatches between Program Design and Implementation Studies of the ICDS program, including this one, have repeatedly raised concerns about its design and implementation. Three major mismatches in implementation undermine the potential of ICDS to address child undernutrition effectively, efficiently, and equitably. Mismatch I: Although the design of ICDS recognizes the multi- dimensional determinants of undernutrition, too much emphasis is currently given to providing food security through the supplementary nutrition program. Not enough attention is given to the most effec- tive interventions for child nutritional outcomes, such as improving Enhancing the Impact of ICDS · 73 childcare behaviors and educating parents on how to improve nutri- tion using the family food budget. Mismatch II: Service delivery is not focused enough on the youngest children (under 3), who can potentially benefit most from ICDS interventions. In addition, children from wealthier households partic- ipate much more than children from poorer ones, and ICDS is only partially succeeding in preferentially targeting girls and lower castes. Mismatch III: Although the increase in program coverage was greater in underserved than well-served areas during the 1990s, the poorest states and those with the highest levels of undernutrition still have much lower levels of program funding and coverage than other states. How Can ICDS Reach Its Full Potential? In this section a menu of options is proposed to increase the impact of ICDS on the nutritional status of priority groups (table 3.1). It draws on the findings of Millions Saved: Proven Successes in Global Health (Levine and the What Works Working Group 2004), which docu- ments 17 cases in which large-scale national, regional, and global efforts have improved health status in developing countries. In order to be labeled successful, these cases had to meet a set of rigorous selection criteria. They had to be of large scale, last at least five years, employ a cost-effective intervention, and have an impact on an impor- tant health problem. Although no single recipe emerges from the review of the successful programs, a consistent set of ingredients is found to contribute to success: predictable, adequate funding from both international and local sources; political leadership and champi- ons; technological innovation within an effective delivery system, at a sustainable price; technical consensus about the appropriate biomed- ical approach; good management on the ground; and effective use of information. In most cases, community participation was also a con- tributing factor. ICDS is assessed with respect to these elements of success. It pres- ent options that the Department of Women and Child Development could consider for realigning the design and implementation of ICDS in order to improve the program's impact.34 Particular attention is given to what can be done to fix the three mismatches. 74 Table 3.1 Menu of options for improving ICDS Positive feature Area needing improvement How to do it Overall program Designed to address the multiple Mismatch I: Wide gap between original Rationalize design and improve implementation: determinants of undernutrition intention and design and actual · Define priority objectives. (food security, health services, implementation: food supplementation · Identify cost-effective interventions to achieve those caring and feeding behaviors). dominates, at the expense of linkages with objectives. health sector and counseling of parents. · Implement activities to deliver interventions. · Monitor execution and evaluate impact. Designed to address Mismatch II: Service delivery remains focused Improve targeting of children under 3 and pregnant women: intergenerational cycle of on older children (3­6). · Strengthen nutrition and health education activities. undernutrition (that is, pregnant · Increase home visits. women and young children). · Improve targeting of poorest and most vulnerable Although initial design focus was households. on children 3­6, over the past decade, design focus shifted · Introduce mini-anganwadi centers (poriawadis). toward children 0­3. · Increase outreach activities. Designed to target poor states Mismatch III: Per child spending is higher in · Address regressive distribution of financing across and poor and vulnerable people richer states and in states with lower states by targeting future expansion to districts and within these states. prevalence of malnutrition. Some of the poorest blocks with highest prevalence of malnutrition. and most vulnerable groups are not reached. Table 3.1 (continued) Menu of options for improving ICDS Positive feature Area needing improvement How to do it Overall program Wide coverage. Quality of services is poor. Develop capacity to deliver all nutrition interventions: Strong grassroots presence. · Increase external participation in service delivery (for example, mothers groups). · Increase synergy with other programs (such as reproductive and child health and primary education). · Add a second anganwadi worker. · Contract private sector for specific activities. Optimize use of available resources: · Improve skills of anganwadi workers and helpers. · Introduce supportive supervision. · Improve supply of inputs. Strengthen focus on results and accountability: · Decentralize responsibility and management of program to state governments and panchayat raj institutions through performance-based financing. · Reform the management information system. · Reward performance at all levels of the administration. · Strengthen community ownership and enhance accountability to local communities. · Involve panchayat raj institutions in monitoring service delivery. Design is standardized and does not reflect Introduce flexibility through bottom-up planning. local needs. 75 (continued on next page) 76 Table 3.1 (continued) Menu of options for improving ICDS Positive feature Area needing improvement How to do it Food security Designed to fill the "food gap" in Food supplementation is universal and · Ensure that malnourished children are reached by the intake of young, absorbs much of the financial and time supplementary nutrition program. undernourished children. resources in the anganwadi center. · Improve efficiency of procurement and distribution of supplementary nutrition program so that resources can be freed up to strengthen other nutrition interventions. Food availability is irregular and quality · Improve procurement and distribution of food (by often poor. decentralizing procurement of food to community level or contracting with the private sector for food distribution, for example). Leakage to non-priority groups · Strengthen management information systems. · Encourage community ownership and monitoring. Health Designed to link with health Articulation with health system is weak. · Strengthen convergence with the Reproductive and services for immunization, Child Health Program. Vitamin A supplementation, and · Introduce joint bottom-up planning process with the referral of high-risk children and Reproductive and Child Health Program. pregnant women. · Provide better training of auxiliary nurse-midwifes in nutrition issues and best practices. Emphasis on counseling and behavior change · Reset priorities and redirect resources toward disease is inadequate. prevention and control. Table 3.1 (continued) Menu of options for improving ICDS Positive feature Area needing improvement How to do it Care Designed to support effective Anganwadi workers are overburdened with · Foster community support (for example, mothers nutrition counseling and growth tasks that take priority over promoting nutrition. groups). promotion linked to regular Anganwadi workers receive little training to · Increase number of workers/helpers at anganwadi growth monitoring. develop skills needed to counsel parents. centers. · Improve training. Equipment and supplies for weighing and · Strengthen management information systems and promoting growth are inadequate. improve the supply system. Emphasis on counseling and behavior change · Reset priorities and redirect resources toward is inadequate. promoting appropriate breastfeeding, home-based complementary feeding, and caring behaviors. · Provide additional training. Micronutrients Center-based interventions are Articulation with the Reproductive and Child · Strengthen convergence with the Reproductive and potentially useful for Health Program is weak. Child Health Program. supplementation of Vitamin A, iron, and folic acid. Source: World Bank recommendations. 77 78 · India's Undernourished Children Improve Service Delivery at Existing Anganwadi Centers Availability of funds has not been a major problem for ICDS, which has received extensive financing from both national and international sources. Over the years, both total spending and spending per child on various ICDS components has increased substantially (World Bank 2004d). The government of India's contribution increased from Rs329.8 crores in 1992/93 to Rs1,311.2 crores in 2001/02. Expenditure on supplementary nutrition, which is financed by state governments, increased by a factor of almost four during the same period. Funding has increased, but it is not clear that the increase has had a measurable impact on children's nutritional status. Rather than expanding coverage, it might be more beneficial to allocate funds to improving service delivery at existing anganwadi centers. Increase High-Level Commitment and Mobilize Political Leadership High-level political commitment is key to all successful public health programs. India has one of the highest proportions of underweight children in the world, and the government has often expressed its commitment to reducing malnutrition. That commitment is not ade- quately reflected in current policy discussions, however. Several factors may explain this. They include lack of awareness of the most cost-effective interventions, a tendency to view malnutrition inter- ventions as transfers to the poor and to underestimate their economic impact on the country as a whole, the multiplicity of organizational stakeholders involved, and the relatively muted voice of the poor. To build commitment and mobilize political leadership toward supporting changes in the existing array of nutrition programs in India, public and private stakeholders will have to be made aware of the size and character- istics of the undernutrition problem in India; the devastating human, social, and economic consequences of failing to address the problem; and the substantial human, social, and economic benefits associated with the implementation of available, affordable, and cost-effective nutrition interventions. Fix the Mismatches between Program Design and Implementation ICDS has not yet effectively implemented the most cost-effective nutrition interventions or reached priority groups. Substantial Enhancing the Impact of ICDS · 79 changes in program implementation need to be introduced to fix the three most important mismatches. Fix Mismatch I: Bridge the Gap between Program Design and Implementation so that the Most Important Causes of Undernutrition in India Are Addressed. Feeding and caring practices. Although exclusive breastfeeding in the first months of life is important to avoid infection, water and other supplements are frequently given in early infancy (IIPS and Orc Macro 2000). A 2003 study in 49 districts revealed that only 40 per- cent of infants were exclusively breastfed during the first six months (BPNI 2003). Other studies indicate that the quality of complemen- tary foods can be poor, due to local customs and beliefs (Roy 1997). Much needs to be done to reduce this source of nutritional depriva- tion during this crucial growth period. The situation regarding the introduction of semi-solid comple- mentary foods is even worse. According to the NFHS II (1998/99), only one-third of children in India were offered any semi-solid food between the ages of six and nine months. Along with infections, delayed introduction of semi-solid foods is an important trigger of malnutrition, which is worst between 6 months and 18­24 months. Anganwadi workers should devote much more attention to encourag- ing exclusive breastfeeding for the first six months and adding semi- solid complementary food three to four times a day in appropriate quantities thereafter (DWCD 2004b; Ghosh 2004). Another key way to improve child growth is to show women how to use their own resources to feed their children more effectively. This approach has been used in many countries, including China, the Repub- lic of Korea, and Vietnam (Whang 1981; Allen and Gillespie 2001). An intervention in Haiti taught mothers to use inexpensive local foods to prepare nutritious food for their children (King and others 1978; Berggren and others 1983; Scrimshaw 1995). The effort was highly suc- cessful in helping mothers rehabilitate their malnourished children: mortality rates of children whose mothers received demonstration-edu- cation were 68 percent of those of children whose mothers received growth-monitoring and counseling services but no demonstration-edu- cation. In households in which the mother participated in demonstra- tion-education, the younger siblings of malnourished children were also less likely to become malnourished, and they had significantly lower mortality rates than the younger siblings of malnourished children 80 · India's Undernourished Children whose mothers had not participated in demonstration-education. Simi- lar positive effects of maternal knowledge and childcaring practices have been found in Bangladesh (Karim and others 2003). Promotion of feed- ing and caring practices is a critical aspect of ICDS that needs to be strengthened. Disease control and prevention. Recognizing that child growth and health can be enhanced by improving environmental hygiene and domestic health management practices, the ICDS program includes components for deworming, iron supplementation for children, and home visits to improve childcare practices. Given the high prevalence of worm infestations and gastroenteric infections in India, these poli- cies need to be implemented much more rigorously. Anganwadi work- ers need to be given more training and encouragement to implement these interventions and work with communities to improve their sani- tary practices. Collaboration between ICDS and the health delivery system has improved in recent years. One consequence of this collaboration has been better immunization coverage. The partnership between the anganwadi worker and the auxiliary nurse-midwife has been less suc- cessful with respect to identifying high-risk pregnancies, providing antenatal and postnatal care, and conveying adequate health and nutritional messages to women. Increased collaboration would help ensure the provision of broader child and maternal health services. Strengthening the convergence of ICDS and the Reproductive and Child Health Program should be a priority. Micronutrient supplementation. ICDS can be used to facilitate chil- dren's access to national micronutrient supplementation programs for iron, Vitamin A, and iodine. These interventions have been shown to be exceptionally cost effective in a number of settings (Behrman, Alderman, and Hoddinott 2004), and their benefits for child growth, health, and cognitive development are well documented. To date, however, micronutrient interventions in India--namely, the distribu- tion of iodized salt, the administration of a semi-annual massive dose of Vitamin A to young children, and the distribution of iron folic acid tablets to vulnerable groups--appear to have had little effect (Vija- yaraghavan 2002). These programs need to be strengthened. Supplementary feeding. ICDS functionaries at all administrative levels, as well as program beneficiaries, appear to consider the supplementary Enhancing the Impact of ICDS · 81 nutrition program (food distribution) to be synonymous with the full set of nutrition interventions of ICDS, often using the two concepts interchangeably. The confusion is indicative of the pervasiveness of the food bias in the ICDS program. The food bias is also evident in the allocation of expenditure across ICDS components: the supplementary feeding program accounts for about two-thirds of the total cost of the ICDS program (Radhakrishna, Ravi, and Indrakant 1998). It is impor- tant to use supplementary feeding strategically--as an incentive for poor and malnourished children to attend anganwadi centers, where they, and their mothers, can receive health and nutrition education interventions. It is crucial that ICDS implementation emphasize the multidimensional nature of malnutrition; that food intake be under- stood as only one, and most often not the main, determinant of child nutritional status; and that resources be redirected toward improving the delivery of other ICDS services. Fix Mismatch II: Increase Impact by Reaching the Youngest Children. Because of the types of services provided and the focus on center-based activities, ICDS tends to reach mainly 3- to 6-year-olds, somewhat at the expense of pregnant women and children under 3. Young children need to be accompanied to the anganwadi center, and they require more time and attention than older children. Because fewer young children attend the center, interventions often miss this critical group. As a result, the prevalence of stunting and underweight remains very high.35 Failure to reach young children is of particular concern in light of the evidence that most growth faltering occurs during the first two years of life and that it negatively affects children's development throughout their lives (Allen and Gillespie 2001). A more concerted effort needs to be made to recruit young children into the program, perhaps by effectively reaching out to women while they are pregnant or just after they give birth. Recruiting more young children would produce a shift toward preventing malnutrition rather than treating it, often after it is too late to recover the growth trajectory. The advan- tage of some of the cost-effective measures described in table 3.1 is that unlike food supplementation, they are occasional interventions that do not require regular attendance at the anganwadi center (some can even be delivered in beneficiaries' homes). They are thus effective in reaching children under 3. 82 · India's Undernourished Children In this context, conditional cash transfers have been very successful in increasing the demand for health care for young children, educat- ing parents about adequate caring and feeding practices, and rapidly improving child nutritional and health status in Colombia (Attansio, Syed, and Vera-Hernandez 2004), Honduras (Rawlings and Rubio 2003), and Mexico (Skoufias 2001). The possibility of introducing such programs in India should be explored. Fix Mismatch III: Improve Targeting by Increasing Coverage in Poorer States and Districts. Another source of poor targeting lies in the regressive distribution of the ICDS program across states. The poorest states tend to receive the lowest government budgetary allocations per mal- nourished child. Thus the states with the highest prevalence of stunt- ing and underweight tend to have the weakest program coverage. There are some encouraging signs, though. First, the poorest states experienced the highest rate of growth of program coverage during the 1990s. Second, the program is more evenly distributed within states than across states: about 60 percent of the poorest villages in every state are covered by ICDS programs, compared with 70 percent of the wealthiest villages. Controlling for other village characteristics, within a given state, program placement is progressive. The government of India has an action plan to construct another 188,000 anganwadi centers over the next few years. Given the high degree of concentration of child malnutrition in India, any future investment in ICDS should be driven by careful targeting of high- prevalence districts, villages, and settlements across the country. Unfortunately, available data cannot yet shed light on which villages should be chosen, because the sample surveys are not large or repre- sentative enough at the village level. However, promising new methodologies, based on the merging of household survey and census data, can help identify villages that are likely to have the highest prevalence of malnutrition. Targeting resources at villages based on their need is desirable not only for equity reasons--it is also the most effective strategy to reduce the prevalence of malnutrition.36 Improve Management on Site Effective service delivery requires that trained and motivated workers are in place and have the supplies, equipment, transportation, and Enhancing the Impact of ICDS · 83 supervision to do their jobs well. This requires both adequate funding and good management. In some instances, strong management can partially compensate for budgetary restrictions. A large number of studies document the implementation difficul- ties ICDS has experienced (NIPCCD 1992; Greiner and Pyle 2000; NCAER 2001; Allen and Gillespie 2001; Educational Research Unit 2004; Bredenkamp and Akin 2004). Some of these problems are due to the rapid expansion of the program, which has been faster than the institutional capacity necessary to manage it (World Bank 1998). Rapid expansion has not allowed anganwadi workers to be trained ade- quately. As a result, many workers have been sent to their centers with little or no training and have had to learn on the job. Refresher train- ing is scarce, and adequate supervision is lacking. ICDS support serv- ices at the state level are inadequately staffed. As a result, although their job requires an understanding of nutrition, preschool education, and maternal and child health issues, anganwadi workers have very lit- tle technical or other support in providing ICDS services. Moreover, anganwadi workers are charged with a multiplicity of tasks, not all of them related to the central ICDS objectives. These responsibilities force them to divert some of their energies from the most important interventions. It is imperative that anganwadi workers be perceived and treated as the core input for ICDS service delivery and given the right tools and support to perform their tasks effectively. A second problem is the erratic supply of food in ICDS. The national evaluation conducted in 1992 (NIPCCD 1992) found that the average anganwadi center was without food 20 percent of the time, and more than one-fourth of all centers experienced shortages that lasted longer than 3 months. Widespread delays in food distribution persist today (see table 2.3). Leakages in the distribution of ICDS food are substantial at many levels, notably in the procurement of food supplies (Greiner and Pyle 2000). In the absence of localized food insecurity (such as drought or crop failure), local procurement may be a more effective means of supplying food. Local procurement would probably increase the regular- ity of the food supply, since it is easier to hold local providers account- able for delivery, and local inhabitants would have a vested interest in the well-being of the children in their community. Moreover, local procure- ment provides a source of income to local inhabitants and promotes community awareness of and involvement in ICDS activities. 84 · India's Undernourished Children A third problem is the lack of growth-monitoring equipment. Many anganwadi centers do not have weighing scales that are in work- ing condition, lack growth charts, or have insufficient numbers of growth cards. The monitoring and evaluation system fails to remedy shortfalls in supply. Growth-monitoring activities are used to educate and encourage mothers to adopt behaviors that promote the growth of their children. It is in this area that the ICDS program is most lacking. It is critical that anganwadi workers be trained to conduct growth-monitoring and growth-promotion activities. Use Information Effectively Information can improve the effectiveness of ICDS in three ways. First, information about the extent of a problem raises awareness and focuses political and technical attention on finding solutions. Second, research on health behaviors and the effectiveness of different service delivery approaches can help shape the design of a program and increase its prospects for success. Third, information creates account- ability and motivates. It is generally recognized that monitoring and evaluation activities related to ICDS need strengthening, and a concerted effort is cur- rently being made to do so. Toward this end, the Department of Women and Child Development might consider applying the moni- toring and evaluation framework it uses for World Bank­funded ICDS projects to all ICDS projects. High-quality information needs to be collected that is relevant, in the sense that the data clearly reveal something about the functioning of important aspects of the program. The quantity of data collected must be manageable, since large volumes of information are unlikely to be used to inform decisions. In this regard, it may be helpful to revisit the guidelines and instructions issued for the monitoring and evaluation of ICDS and to streamline and fine-tune them in an effort to reduce the volume of superfluous information and the time needed to process it. The number of registers currently collected by angan- wadi workers, for example, far exceeds the capacity to use this infor- mation for program management. Simultaneous with an effort to streamline and standardize the indi- cators collected across states should be the development of a standard Enhancing the Impact of ICDS · 85 template with which to display information. Such a template would make ICDS data more accessible at more levels and to more people in the project management system. Standardization would also facilitate comparisons across states, highlighting the states from which lessons can be learned in key areas of implementation. It would also promote the analysis of trends within states and the aggregation of data at the national level. Computerization and electronic processing of information would greatly facilitate monitoring and evaluation. The challenge is to find a way of processing the data into a form that is usable, so that a program manager or other interested party can determine the status of activities--the percentage of a target group receiving benefits, the percentage of centers with weighing scales, whether food was received the previous month--at any point in time, past or present. Ideally, users should have easy access not only to aggregate indicators, but also to block- and district-level information. Periodically, quality control checks on monitoring data should be undertaken to uncover any sys- tematic errors in reporting and identify the sources of any discrepan- cies. These changes would help transform the data collected by angan- wadi centers into information that can be used to identify problems and to take the action needed to resolve them. More human resources need to be devoted to monitoring and eval- uation. One way to do so would be to increase awareness of the importance of monitoring at all levels of implementation, so that functionaries give these activities the attention they deserve. Creating awareness is challenging and requires a substantial mind shift for functionaries toward outcomes, results, and performance rather than inputs. Strengthening of community monitoring is also desirable, through existing community institutions or, more informally, by encouraging community members to be alert to anganwadi center opening hours and attendance and demand improvements where needed. Increase Decentralization and Community Participation With few exceptions, ICDS remains a highly standardized interven- tion that follows rules and regulations set centrally. Given the hetero- geneity of malnutrition patterns in India, state governments should be encouraged to tailor the basic model to local needs and assume 86 · India's Undernourished Children responsibility for managing the overall program rather than focusing almost exclusively on the procurement and distribution of supplemen- tary food (the only activity in the program they finance directly). A budget line that is specific to the financing of ICDS should be introduced in state budgets, so that the planning and monitoring of investments in ICDS becomes an explicit state-level activity. ICDS is run in a very top-down fashion, with all the logistical and implementation inefficiencies and rigidities that such an approach entails. A program to provide daily services to young children and pregnant women requires strong participation and supervision by the community. There appears to be some empirical association between the strength of community support for ICDS, in the form of financial contributions from the panchayat, and the performance of anganwadi centers (Bredenkamp and Akin 2004). However, countrywide, only about 25 percent of states receive support from panchayat leaders, and this support has been mainly in the form of providing space for the anganwadi center and recruiting beneficiaries (NCAER 2001). Despite statements of intent to involve communities in the process, there is little sense of community ownership (Greiner and Pyle 2000). This impression is reinforced by the fact that in most communities the anganwadi worker is hired and paid by the government and is not accountable to the community in which he or she works. Equipment, food, and other supplies are provided directly by the government. Because of their daily presence in the village, anganwadi workers are asked to take on many additional duties to support the field outreach staff of other government agencies (education, health, and rural devel- opment, in particular); they are not encouraged to work closely with community organizations, such as the gram panchayat or mahila man- dal. Given the extensive decentralization that has been under way in India over the past decade, there is considerable scope for involving locally elected village committees much more actively in implement- ing ICDS. The experience of the mothers committees in Andhra Pradesh (see chapter 2) could be replicated in other states. One important way to enhance the responsiveness of the ICDS pro- gram and cultivate a sense of local ownership is to always select the anganwadi worker from the community in which he or she works. Although included as a recommendation in the Department of Women and Child Development's guidelines, this does not always occur in Enhancing the Impact of ICDS · 87 practice: appointments are sometimes political or compassionate (made to people in difficult circumstances); sometimes they are even for sale. In many cases, the anganwadi worker is from a forward caste, which may affect the access of children from scheduled castes or tribes since, by their own admission, some anganwadi workers from forward castes make only infrequent home visits to scheduled caste hamlets (Educational Resource Unit 2004). Next Steps: Rationalizing Design and Improving Implementation ICDS has enormous potential to improve the nutritional status of India's children, but it needs to meet some challenges if this potential is to be realized. One challenge is the large and ever-increasing range of duties that anganwadi workers are expected to fulfill. Since, unlike most government workers, their workplace is located at the grass roots, they are asked to help implement a multiplicity of government programs in addition to ICDS. This diverts attention away from their core duties, which are already onerous and rarely can be performed satisfactorily. A second challenge is the fact that the changing scope of the ICDS has resulted in considerable ambiguity among higher-level officials as to the program's objectives, and the capacity of both the central and state units to manage and deliver the program is being stretched. A third challenge is the need to address the mismatches between what an effective nutrition intervention should do and what ICDS is currently doing. Failure to meet these challenges is preventing ICDS from doing as much as it could to reduce the prevalence of malnutrition. It may be time to consider a new approach. One option would be to retain the present structure, in which a preschool function for older children (4­6 years), on the one hand, and maternal and child health and nutrition interventions with special emphasis on younger children (0­3 years), on the other, are offered within the same program. If this option is pursued, the difficulties in simultaneously carrying out these disparate tasks need to be resolved. Under the current program, anganwadi workers devote most of their time to preschool education and older children, squeezing out atten- dance by younger children. Since anganwadi workers spend most of 88 · India's Undernourished Children their remaining time preparing food, they have little time for health interventions or counseling parents about feeding and caring prac- tices. If the present structure is maintained, introducing a system of two workers--one charged with health and nutrition functions, the other charged with the preschool function--may make sense. The National Rural Health Mission launched in fiscal 2005­6 plans to introduce an additional village health worker to focus on maternal and neonatal health issues. Such a worker could attend to the needs of children under 3, including nutrition. The anganwadi worker could focus on preschool education for older children and continue to pre- pare food. Coordination with the auxiliary nurse-midwife of the Reproductive and Child Health Program also needs to be carefully studied, defined, and monitored. A more radical alternative would be to separate services provided to children 4­6 from those provided to younger children and pregnant and lactating women. The demand for preschool education and for feeding older children could be met by devolving these responsibili- ties to the Department of Education or to local authorities. The Dis- trict Primary Education Program already delivers preschool education services in some districts; the feeding of children 4­6 could become part of the National Mid-Day Meals Program (Measham and Chat- terjee 1999). In this manner, more of the anganwadi worker's time could be freed up for nutrition and health education and for growth promotion, increasing the prospect of achieving better nutrition out- comes. Coordination between the anganwadi worker, the auxiliary nurse-midwife, and the accredited social health activists (in the event that the proposal by the National Health Mission is implemented) will be crucial to the success of this effort. Bridging the gap between the policy intentions of ICDS and its actual implementation probably represents the single greatest chal- lenge in international nutrition. Meeting this challenge would have an enormous long-term impact on human development and economic growth. Greater clarity and focus are needed if ICDS is to make a substantial dent in India's persistent undernutrition. In particular, the three mis- matches identified in this report need to be resolved. Only by doing so can the program address the most important determinants of malnutri- tion, reach younger children and the most vulnerable segments of the Enhancing the Impact of ICDS · 89 population, and target areas in which the prevalence of undernutrition is highest. Leadership and commitment are required to address some of the structural inefficiencies of ICDS, including weak information systems, limited orientation toward results, and a lack of accountability for performance at all levels, that are hindering the program from achieving greater results. Appendix Table A.1 Responsiveness of prevalence of underweight to rising per capita GDP, 2002­15 Prevalence of underweight among children Year GDP per capita (billions) under 5 (percent) 2002 487.0 47.0 2003 501.6 46.3 2004 516.7 45.6 2005 532.2 44.9 2006 548.1 44.2 2007 564.6 43.5 2008 581.5 42.9 2009 599.0 42.2 2010 616.9 41.6 2011 635.4 40.9 2012 654.5 40.3 2013 674.1 39.7 2014 694.4 39.1 2015 715.2 38.5 Source: World Bank calculations. Note: Calculations assume annual economic growth of 3 percent, exogenous income elasticity of malnutrition of 0.51, and percentage change in the prevalence of malnutrition of 2 percent. 91 92 · India's Undernourished Children Figure A.1 Weight-for-age estimates of change in nutritional status, in selected regions 0.5 0.0 ­0.5 age ­1.0 for ­1.5 weight ­2.0 ­2.5 ­3.0 0 3 6 9 12 15 18 21 24 27 30 33 36 age in months Africa Asia Latin America and Caribbean India Source: Regional estimates from Shrimpton and others 2001; India data from IIPS and Orc Macro (2000). Appendix · 93 Table A.2 Prevalence of anemia among children and women in India, by state, 1998­9 (percent) Children under 3 Ever-married women 15­49 State Mild Moderate Severe Total Mild Moderate Severe Total Andhra Pradesh 23.0 44.9 4.4 72.3 32.5 14.9 2.4 49.8 Arunachal Pradesh 29.1 24.7 0.7 54.5 50.6 11.3 0.6 62.5 Assam 31.0 32.2 0.0 63.2 43.2 25.6 0.9 69.7 Bihar 26.9 50.3 4.1 81.3 42.9 19.0 1.5 63.4 Delhi 22.2 42.9 3.9 69.0 29.6 9.6 1.3 40.5 Goa 23.5 27.9 2.0 53.4 27.3 8.1 1.0 36.4 Gujarat 24.2 43.7 6.7 74.5 29.5 14.4 2.5 46.3 Haryana 18.0 58.8 7.1 83.9 30.9 14.5 1.6 47.0 Himachal Pradesh 28.7 39.0 2.2 69.9 31.4 8.4 0.7 40.5 Jammu and Kashmir 29.1 38.5 3.5 71.1 39.3 17.6 1.9 58.7 Karnataka 19.6 43.3 7.6 70.6 26.7 13.4 2.3 42.4 Kerala 24.4 18.9 0.5 43.9 19.5 2.7 0.5 22.7 Madhya Pradesh 22.0 48.1 4.9 75.0 37.6 15.6 1.0 54.3 Maharashtra 24.1 47.4 4.4 76.0 31.5 14.1 2.9 48.5 Manipur 22.6 21.7 0.9 45.2 21.7 6.3 0.8 28.9 Meghalaya 23.4 39.8 4.3 67.6 33.4 27.5 2.4 63.3 Mizoram 32.2 22.7 2.3 57.2 35.2 12.1 0.7 48.0 Nagaland 22.0 18.7 3.0 43.7 27.8 9.6 1.0 38.4 Orissa 26.2 43.2 2.9 72.3 45.1 16.4 1.6 63.0 Punjab 17.4 56.7 5.9 80.0 28.4 12.3 0.7 41.4 Rajasthan 20.1 52.7 9.5 82.3 32.3 14.1 2.1 48.5 Sikkim 28.4 40.7 7.5 76.5 37.3 21.4 2.4 61.1 Tamil Nadu 21.9 40.2 6.9 69.0 36.7 15.9 3.9 56.5 Uttar Pradesh 19.4 47.8 6.7 73.9 33.5 13.7 1.5 48.7 West Bengal 26.9 46.3 5.2 78.3 45.3 15.9 1.5 62.7 India 22.9 45.9 5.4 74.3 35.0 14.8 1.9 51.7 Source: IIPS and Orc Macro 2000. Table A.3 Percentage of villages covered by ICDS, by state, 1992/93­98/99 State 1992/93 1998/99 Andhra Pradesh 30 65 Arunachal Pradesh 65 82 Assam 39 30 Bihar 14 32 Delhi 53 55 Goa 85 95 Gujarat 61 84 Haryana 64 92 Himachal Pradesh 39 52 Jammu 44 70 Karnataka 63 86 Kerala 100 97 Madhya Pradesh 27 53 Maharashtra 66 81 Manipur 60 83 Meghalaya 07 22 Mizoram 97 73 Nagaland 54 84 Orissa 42 47 Punjab 39 70 Rajasthan 36 52 Sikkim -- 27 Tamil Nadu 77 43 Tripura 76 83 Uttar Pradesh 20 33 West Bengal 45 58 Total 35 52 Source: Calculated from NFHS (1992/93) and NFHS (1998/99) data by Lokshin and others (2005). -- Not available. Figure A.2 Percentage of children attending anganwadi centers on daily basis, by age and state 100 90 80 70 60 50 children 40 % 30 20 10 0 <12 12­23 24­35 36­47 48­59 60­71 age in months Kerala Maharashtra Rajasthan Uttar Pradesh Madhya Pradesh Chhattisgarh Source: ICDS III baseline/ICDS II endline survey 2000­2. 94 Table A.4 Percentage of children attending anganwadi centers, in villages with centers Kerala Maharashtra Rajasthan Uttar Pradesh Madhya Pradesh Chhattisgarh More than More than More than More than More than More than Item once a month Daily once a month Daily once a month Daily once a month Daily once a month Daily once a month Daily Total 54 49 75 62 10 6 24 6 35 12 47 10 Quintile Quintile 1 (poorest) 52 47 79 63 11 6 a a 31 9 49 6 Quintile 2 53 48 77 60 10 7 24 6 34 10 48 9 Quintile 3 52 48 75 62 9 4 28 6 36 11 52 10 Quintile 4 54 49 73 69 11 5 25 7 34 11 47 11 Quintile 5 (richest) 56 51 66 52 11 6 18 4 40 24 41 13 Age 3 and under 30 22 67 50 10 6 22 3 30 8 46 5 4­6 91 91 94 90 14 7 29 12 42 19 49 18 Gender Boys 54 49 75 62 10 5 24 6 34 12 47 10 Girls 53 48 75 62 11 6 24 6 36 12 48 10 Caste Scheduled caste 55 50 74 64 15 9 30 6 36 13 47 8 Scheduled tribe 49 45 79 65 11 7 19 4 41 14 50 11 Other backward groups 54 49 76 58 8 4 22 5 31 11 46 9 Other castes 53 48 71 60 9 5 20 6 34 15 44 11 Locality Urban 51 48 72 64 12 7 16 3 54 37 50 21 Rural 58 51 75 62 9 4 26 6 29 7 44 5 Tribal 45 39 79 60 11 7 -- -- 43 15 51 10 95 Source: ICDS III Baseline/ICDS II endline survey 2000­2. -- Not available. a. Too few observations. 96 Table A.5 Receipt of health interventions during pregnancy under Care India's Integrated Nutrition and Health Project II (percent) All Low SES High SES Intervention Nonintervention Intervention Nonintervention Intervention Nonintervention Intervention areas areas areas areas areas areas Consumption of 90+IFA 60 41* 62 40* 58 43 Tetanus toxoid (2+) 87 74* 91 70* 84 78 Antenatal checkups (3+) 53 38* 54 29* 53 53 Number of observations 189 151 69 83 120 68 Source: Personal communication with Care India. * Statistically significant differences between intervention and nonintervention areas. IFA = iron and folic acid supplement. Table A.6 Adoption of appropriate infant feeding behaviors under Care India's Integrated Nutrition and Health Project II (percent) All Low socioeconomic status High socioeconomic status Intervention Nonintervention Intervention Nonintervention Intervention Nonintervention Feeding behavior areas areas areas areas areas areas Initiation of breastfeeding 65.2 38.3* 75.4 42.0* 59.5 33.8* within 1 hour of birth n = 181 n = 149 n = 65 n = 81 n = 116 n = 68 Exclusive breastfeeding 69.3 57.6* 69.6 63.9 69.2 50.0* for at least 6 months n = 189 n = 151 n = 69 n = 83 n = 120 n = 68 Complementary feeding initiated 65.3 43.6* 66.1 36.5* 63.6 50 (among 6- to 9-month-olds) n = 121 n = 110 n = 55 n = 52 n = 66 n = 58 Among those who initiated complementary feeding, dietary diversity in complementary feeding vegetables given 68.0 43.6* 62.9 50.7 71.9 37.0* oil added to food 41.9 20.5* 38.2 22.7* 44.7 18.5* dal or animal foods given 79.8 55.8* 73.0 58.7 85.1 53.1* n = 203 n = 156 n = 89 n = 75 n = 114 n = 81 Appropriate quantity, frequency, 6.1 0.5* 2.8 0 8.7 0.9* and diversity in feeding for age n = 244 n = 218 n = 106 n = 109 n = 138 n = 109 Measles immunization by 12 months 55.4 35.1* 47.3 25.0* 62.1 44.1 n = 121 n = 111 n = 55 n = 52 n = 66 n = 59 Vitamin A (one dose) among 59.5 43.2* 49.1 44.2* 68.2 42.4* children 9­11 months n = 121 n = 111 n = 55 n = 52 n = 66 n = 59 Source: Personal communication with Care India. * Statistically significant differences between intervention and nonintervention areas. 97 Notes Chapter 1 1. The term malnutrition refers to both under- and overnutrition. India does have a small but increasing percentage of overweight children who are at risk for noncom- municable diseases such as diabetes and cardiovascular heart disease later in life. However, in view of the size and urgency of the undernutrition problem in India and its links to human development, this analysis deals only with undernutrition. 2. Clinical Vitamin A deficiency is a severe form of Vitamin A deficiency, which may result in xerophthalmia, a condition caused by inadequate functioning of the glands that produce tears. Symptoms include night blindness, Bitot's spots, xerosis, and keratomala- cia. If not treated early enough, xerophthalmia can eventually lead to blindness. Sub- clinical Vitamin A deficiency is associated with increased vulnerability to a variety of infectious diseases and, therefore, an increased risk of mortality and morbidity. 3. Protein-energy malnutrition develops in children and adults whose consump- tion of protein and energy is insufficient. In most cases, both protein and energy defi- ciencies occur simultaneously. If protein deficiencies predominate, protein-energy malnutrition may manifest as kwashiorkor, which usually appears around the age of 12 months when breastfeeding ceases, but can also occur later in childhood. Kwashiorkor is characterized by edema, hair discoloration, and peeling skin. If energy deficiencies predominate, protein-energy malnutrition may manifest as marasmus, which usually develops in children 6­12 months who have been weaned from breastmilk or suffer from weakening infections, such as diarrhea. It is characterized by stunted growth and wasting. 4. Estimating the economic costs of malnutrition typically takes into account the prevalence of a particular macro- or micronutrient deficiency among men and women and their average levels of participation in market economic activity and heavy labor. Economic calculations are based only on market activities; they exclude non-market losses, even though they may be socially valuable. The calculations also require estimat- ing the degree to which different nutritional conditions may coexist. 99 100 · Notes 5. This estimate represents an upper bound, since the economic status of the child, for example, is unlikely to be completely independent of urban-rural location or caste. 6. Measuring the incidence of low birth weight in developing countries is chal- lenging because of measurement error (as suggested by the heaping of data at the low birth weight cut-off of 2,500 grams) and because relatively few babies are weighed at birth. 7. The rural population of Delhi is not strictly comparable to the rural popula- tions of other states, however, as most of Delhi's "rural" population consists of poor urban populations on the periphery of the city. 8. Principal component analysis, conducted on a set of variables including house- hold assets and housing characteristics, was used to generate the cut-off points for the wealth tertiles, which divide the population of each state into three categories based on the individual's position in the India wealth distribution. Tertiles are used rather than quintiles because in some states there are too few observations available in some quintiles. 9. In the source data (DWCD and UNICEF 2001), reports of day and night-time vision problems were used as indicators of Vitamin A deficiency. However, it is likely that not all vision problems are Vitamin A­related and that there may be some under- reporting in disadvantaged areas due to poorer availability of diagnostic services. 10. Using 1990­2 data from rural areas, as well as the NFHS I (1992/93) and NFHS II (1998/99) data, Wagstaff and Claeson (2004) obtain an average annual reduction of 3.9 percent. Using a constant rate of change and data from NFHS I and NFHS II, Chhabra and Rokx (2004) and World Bank (2004a) obtain similar estimates (1.7 percent and 1.9 percent, with the difference attributable to rounding). 11. The rate shown for 1990 is projected from the change observed between the NFHS surveys conducted in 1992/93 and 1998/99. This MDG target is calculated for children under 3 and therefore differs from the WHO target, which focuses on chil- dren under 5. 12. The World Bank (2004a) estimates that reaching the 2015 MDG target is feasi- ble under the following combination of economic growth and policy interventions: a 0.3 percent increase in average years of female schooling, a 4 percent increase in per child government expenditure on nutrition programs, a 3 percent increase in per capita consumption expenditure, a 1 percentage point increase in the coverage of reg- ular electricity supply, a 1.5 percentage point increase in the population coverage of professionally assisted deliveries, a 1 percentage point increase in village access to pucca (blacktop) roads, and a 2 percentage point decrease in the population with no access to toilets since 1998/99. Notes · 101 Chapter 2 13. Together these factors constitute the concept of "nutrition security," which is viewed as the outcome of good health, a healthy environment, and good caring prac- tices, combined with household-level food security. 14. For evidence from Peru, see Alderman, Hentschel, and Sabates (2003). For evi- dence from Andhra Pradesh, see Alderman, Hentschel, and Sabates (2003) and Gor- don and Dunleavy (2001). 15. The Woman and Child Development Project supports ICDS service delivery in 11 states (Bihar, Chhattisgarh, Jharkhand, Kerala, Madhya Pradesh, Maharashtra, Orissa, Rajasthan, Tamil Nadu, Uttaranchal, and Uttar Pradesh). It includes a component that supports training for ICDS officials across India. 16. Among children under 3, the prevalence of underweight was 29.2 percent where the program was in place and 32.3 percent where it was not. Among children 3­6, the prevalence of underweight was 25.3 percent where the program was in place and 30.2 percent where it was not. 17. The percentage of children who receive any of ICDS's many services is difficult to estimate. The percentage of beneficiaries of the Supplementary Nutrition Pro- gram, one of the main ICDS services, is used as an indicator of the number of ICDS beneficiaries because data on this service is more readily available than other data. 18. This public expenditure estimate combines government expenditure on ICDS with state allocations to ICDS. It excludes any expenditure on ICDS by local govern- ment institutions. 19. Unless otherwise stated, "attendance" refers to visiting the anganwadi center at least once a month, conditional on there being a center in the village. (For figures on children's attendance, see appendix table A.4.) Since the villages and blocks in which households are located were not sampled randomly, the absolute levels of participa- tion cannot be generalized to the entire state but only to the sampled blocks. The dif- ferentials in access by subgroup are likely to be more representative. 20. For disaggregated attendance rates by state, subgroup, and frequency of atten- dance, see appendix table A.4. 21. Anganwadi centers are located an average of 100­200 meters away from benefi- ciary households, with an average travel time of 5­10 minutes (NCAER 2001). 22. Similar findings were obtained in a countrywide study (NCAER 2001), which showed that just 17 percent of centers had toilets. 102 · Notes 23. These include records for daily attendance, preschool education, supplies, the sup- plementary nutrition program, births, deaths, immunization, weight, pregnancy, health referral, a daily dairy, a monthly progress report, and a survey of households in the area covered by the center. 24. In some states, performance is better. In Chhattisgarh, for example, 95 percent of anganwadi centers report being visited by an auxiliary nurse-midwife every month. 25. Most of these registers contain information on the take-up of different ICDS services, but anganwadi workers are also frequently charged with collecting informa- tion for other government programs, such as old-age schemes. 26. Key indicators include figures on personnel, operationalization of blocks and anganwadi centers, supply of supplementary nutrition, preschool education, births and deaths, and malnutrition status using the IAP (Gomez) classification. 27. The government has issued clearer monitoring and evaluation guidelines to the states, held annual and periodic review meetings at the central level, provided small supplementary financial allocations to monitoring and evaluation activities at the local level, and plans to revise the monitoring formats and the number of anganwadi center registers. In World Bank project states, ICDS input, process, and impact indicators that are compatible with the project's development objectives were defined at the out- set of the project, and adequate financial allocations were made to the monitoring and evaluation component of ICDS. Monitoring and evaluation activities include field visits, periodic reviews, operations research, continuous social assessments, and base- line and endline surveys, in addition to the standard ICDS monitoring activities. 28. In Madhya Pradesh only 58 percent of urban anganwadi centers had been vis- ited by supervisors in the previous month. 29. In Chhattisgarh 43 percent of anganwadi workers were not linked to supervisors. 30. This project is implemented in partnership with the Department of Women and Child Development and the Department of Health and Family Welfare of the Government of India, nongovernmental organizations, and community-based organ- izations, with support from USAID and its BASICS II project for child survival. It is being implemented in Andhra Pradesh, Bihar, Chhattisgarh, Jharkhand, Madhya Pradesh, Orissa, Rajasthan, Uttar Pradesh, and West Bengal. 31. Interventions include antenatal care, nutrition counseling, and birth prepared- ness; home-based newborn care; maternal and child immunization; child feeding advice; vitamin A supplementation for children; and supplementary nutrition. 32. This section draws on research by the International Food Policy Research Institute (IFPRI 2003). Notes · 103 33. The target number of committees is 53,144, which will cover all anganwadi centers in Andhra Pradesh. Chapter 3 34. Technological innovation is not considered, because it is not likely to play a key role in this type of nutrition intervention. 35. Many of these problems were addressed in Tamil Nadu's modification of the ICDS program (TINP), which halved the prevalence of severe malnutrition in the villages in which it was implemented by targeting food to the needy and requiring them to eat it on the premises instead of taking it home to share with others (Heaver 2002; Greiner and Pyle 2000). 36. See World Bank (2004a) for an explanation of the concentration of child malnutri- tion and possible methodologies for improving targeting. Bibliography ACC/SCN (United Nations Administrative Committee on Coordination/Standing Committee on Nutrition). 1997. Third Report on the World Nutrition Situation. Geneva: ACC/SCN. ------. 2000. Fourth Report on the World Nutrition Situation: Nutrition throughout the Life Cycle. Geneva: ACC/SCN. ------. 2004. Fifth Report on the World Nutrition Situation: Nutrition for Improved Devel- opment Outcomes. Geneva: ACC/SCN. Adhikari, S. 2004. "Monitoring and Evaluation in India's ICDS Programme." Back- ground paper for this report. Agarwal, D.K., K.N. Agarwal, K. Satya, and S. Agarwal. 1998. "Weight Gain During Pregnancy: A Key Factor in Perinatal and Infant Mortality." Indian Pediatrics 35 (8): 733­44. Agarwal, S., A. Agarwal, A.K. Bansal, D.K. Agarwal, and K.N. Agarwal. 2002. "Birth Weight Patterns in Rural Undernourished Pregnant Women." Indian Pediatrics 39 (3): 244­53. Alderman, H. 2005. "Linkages between Poverty Reduction Strategies and Child Nutrition: An Asian Perspective." Economic and Political Weekly 40 (46): 4837­42. Alderman, H., J. Hentschel, and R. Sabates. 2003. "With the Help of One's Neighbors: Externalities in the Production of Nutrition in Peru." Social Science and Medicine 56 (10): 2019­31. Alderman, H., J. Behrman, V. Lavy, and R. Menon. 2001. "Child Health and School Enrollment: A Longitudinal Analysis." Journal of Human Resources 36 (1): 185­205. Allen, H.E., D.W.T. Crompton, N. De Silva, P. T. Loverde, and G.R. Olds. 2002. "New Policies for Using Antihelmintics in High-Risk Groups. Trends in Parasitol- ogy 18: 381­82. 105 106 · Bibliography Allen, L.H. 1994. "Nutritional Influences on Linear Growth: A General Review." European Journal of Clinical Nutrition 48, Suppl. 1: S75­S89. Allen, L.H., and S.R. Gillespie. 2001. What Works? A Review of the Efficacy and Effective- ness of Nutrition Interventions. Manila: Asian Development Bank, in collaboration with ACC/SCN. Anand, K., S. Kant, and S.K. Kapoor. 1999. "Nutritional Status of Adolescent School Children in Rural North India." Indian Pediatrics 36 (8): 810­16. ASC (Administrative Staff College of India). 1998. National Strategy to Reduce Child Malnutrition. Hyderabad. Attanasio, O., M. Syed, and M. Vera-Hernandez. 2004. "Early Evaluation of a New Nutrition and Education Programme in Colombia." Briefing Note No. 44. Insti- tute for Fiscal Studies, London. Barker, D.J.P., T. Forsčn, A. Uutela, C. Osmond, and J.G. Eriksson. 2001. "Size at Birth and Resilience to Effects of Poor Living Conditions in Adult Life: Longitu- dinal Study." British Medical Journal 323 (7324): 1273­76. Beaton, G., R. Martorell, and K. Aronson. 1993. Effectiveness of Intermittent Iron Supple- mentation in the Control of Iron Deficiency Anemia in Developing Countries: An Analy- sis of Experience. Final Report to the Micronutrient Initiative, Ottawa. Behrman, J.R., H. Alderman, and J. Hoddinott. 2004. "Copenhagen Consensus: Chal- lenges and Opportunities: Hunger and Malnutrition." Copenhagen Consensus Chal- lenge Papers. May 7. Bentley, M.E., and P.L. Griffiths. 2003. "The Burden of Anemia among Women in India." European Journal of Clinical Nutrition 57 (1): 52­60. Berggren G., W. Berggren, A. Verly, N. Gamier, and W. Peterson. 1983. "Traditional Midwives, Tetanus Immunization, and Infant Mortality in Rural Haiti." Tropical Doctor 13 (2): 79­87. Bhandari, L., and L. Zaidi. 2004. "Reviewing the Costs of Malnutrition in India." Background paper for this report. Black, M.M. 2003. "Micronutrients Deficiencies and Cognitive Functioning." Journal of Nutrition 133 (11): S3927­31. Black, R.E., S.S. Morris, and J. Bryce. 2003. "Where and Why Are 10 Million Chil- dren Dying Every Year?" Lancet 361 (9376): 2226­34. Bleichrodt, N., and M.P. Born. 1994. "A Meta-Analysis of Research on Iodine and its Relationship to Cognitive Development." In The Damaged Brain of Iodine Deficien- cy: Cognitive, Behavioral, Servomotor and Educative Aspects, ed. J.B. Stanbury, 279­85. New York: Cognizant Communication Corporation. Bibliography · 107 Bouis, H., and J. Hunt. 1999. "Linking Food and Nutrition Security: Past Lessons and Future Opportunities." Asian Development Review 17 (1, 2): 168­213. BPNI (Breastfeeding Promotion Network of India). 2003. Status of Infant and Young Child Feeding in 49 Districts, 98 Blocks. New Delhi: BPNI. Bredenkamp, C., and J.S. Akin. 2004. "India's Integrated Child Development Services Scheme: Meeting the Health and Nutritional Needs of Children, Adolescent Girls and Women?" Background paper for this report. Calder, P.C., and A.A. Jackson. 2000. "Undernutrition, Infection and Immune Func- tion." Nutritional Research Reviews 13 (1): 3­29. Care India. 2004. Implementing an Integrated Package of Nutrition and Health Interven- tions. Lessons from the Early Learning Phase of INHP II, May 2002 to August 2003. New Delhi. Care India and Linkages India. 2003. Profiles for India 2003. New Delhi: Care India. Caulfied, L., M. De Onis, M. Blössner, and R.E. Black. 2004. "Under-Nutrition as an Underlying Cause of Child Deaths Associated with Diarrhea, Pneumonia, Malar- ia, and Measles." American Journal of Clinical Nutrition 80 (1): 193­98. Census of India. 2001. Basic Population Data. http://www.censusindia.net/t_ 00_005.html. Chhabra, R., and C. Rokx. 2004. "The Nutrition MDG Indicator: Interpreting Progress." HNP Discussion Paper, World Bank, Washington, DC. Collins, W.J., and M.A. Thomasson. 2002. "Exploring the Racial Gap in Infant Mor- tality Rates, 1920­1970." Working Paper No. 02­W01, Vanderbilt University, Department of Economics, Nashville, TN. Daniels, M., and L. Adair. 2004. "Growth in Young Filipino Children Predicts School- ing Trajectories through High School." Journal of Nutrition 134 (6): 1439­46. De Onis, M., M. Blössner, E. Borghi, E.A. Frongillom, and R. Morris. 2004. "Esti- mates of Global Prevalence of Childhood Underweight in 1990 and 2015." Jour- nal of the American Medical Association 291 (21): 2600­06. Dev, S. Mahendra. 2004. "Analysis of Public Expenditures and Impact of Public Distri- bution System (PDS) on Food Security." Background paper for this report. DWCD (Department of Women and Child Development). 2003. The Indian Child: A Profile. http://wcd.nic.in/indianchild/index.htm. ------. 2004a. UDISHA. http://wcd.nic.in/udisha/htm/services.htm. 108 · Bibliography ------. 2004b. National Guidelines on Infant and Young Child Feeding. Department of Women and Child Development, New Delhi. ------. 2005. Annual Report 2004­2005. New Delhi: Department of Women and Child Development. DWCD (Department of Women and Child Development), and UNICEF. 2001. Mul- tiple Indicator Survey 2000 (MICS­2000): India Summary Report. November. Department of Women and Child Development, New Delhi. Educational Resource Unit. 2004. "Analysis of Positive Deviance in the ICDS Pro- gram in Rajasthan and Uttar Pradesh." Background paper for this report. Esrey, S.A., J.B. Potash, L. Roberts, and C. Shiff. 1990. Health Benefits from Improve- ments in Water Supply and Sanitation: Survey and Analysis of the Literature on Selected Diseases. WASH Technical Report 66, Water and Sanitation for Health Project, Arlington, VA. Fawzi, W.W., T.C. Chalmers, and M.G. Herrera. 1993. "Vitamin A Supplementation and Child Mortality: A Meta-Analysis." Journal of the American Medical Association 269 (7): 898­903. Ghosh, S. 2004. "Child Malnutrition." Economic and Political Weekly 39 (40): 4412­14. Glewwe, P., H. Jacoby, and E. King. 2001. "Early Childhood Nutrition and Academic Achievement: A Longitudinal Analysis." Journal of Public Economics 81 (3): 345­68. Gopalan, C. 1981. The National Goitre Control Programme: A Sad Story. Nutrition Foundation of India, New Delhi. http://nutritionfoundationofindia.res.in/ archives.asp?archiveid=35&back=byauthor.asp. ------. 1992. "Growth Charts in Primary Child-Health Care: Time for Reassess- ment." Indian Journal of Maternal and Child Health 3 (4): 98­103. Gordon, J.E., M.A. Guzman, W. Ascoli, and N.S. Scrimshaw. 1964. "Acute Diarrhoeal Disease in Less Developed Countries: Patterns of Epidemiological Behaviour in Rural Guatemalan Villages." Bulletin of the World Health Organization 31: 9­20. Gordon, K.L., and M. Dunleavy. 2001. Environmental Health in India: Priorities in Andhra Pradesh. World Bank, South Asia Region, Environment and Social Devel- opment Unit, Washington, DC. Grantham-Mcgregor, S.M., and C.C. Ani. 2001. "Undernutrition and Mental Devel- opment." In Nestlé Nutrition Workshop Series Clinical and Performance Program 5, ed. J.D. Fernstrom, R. Uauy, and P. Arroyo, 1­18. Basel: Nestlé Ltd. Greiner, T., and D.F. Pyle. 2000. India: Nutrition Assessment. World Bank, Washing- ton, DC. Bibliography · 109 Haddad, L., H. Alderman, S. Appleton, L. Song, and Y. Yohanes. 2003. "Reducing Malnutrition: How Far Does Income Growth Take Us?" World Bank Economic Review 17 (1): 107­13. Heaver, R. 2002. "India's Tamil Nadu Nutrition Program: Lessons and Issues in Man- agement and Capacity." HNP Discussion Paper. World Bank, Washington, DC. Heston, A., R. Summers, and B. Aten. 2002. Penn World Tables Version 6.1. University of Pennsylvania, Center for International Comparisons, Philadelphia. Horton, S. 1999. "Opportunities for Investments in Nutrition in Low-Income Asia." Asian Development Review 17 (1, 2): 246­73 Horton, S., and J. Ross. 2003. "The Economics of Iron Deficiency." Food Policy 28 (1): 51­75. Horton, S., T. Sanghvi, M. Philipps, J. Fiedler, R. Perez-Escamilla, C. Lutter, A. Rivera, and A.M. Segall-Correa. 1996. "Breastfeeding Promotion and Priority Setting in Health." Health Policy and Planning 11 (2): 156­58. Hunt, J.M. 2004. "Costs of Halving Global Hunger: Reducing Micronutrient Defi- ciencies and Child Underweight Prevalence in Accord with the Millennium Development Goals #1 and #4." Background paper prepared under management of the International Food Policy Research Institute for the United Nations Hunger Task Force. IFPRI (International Food Policy Research Institute). 2003. New and Noteworthy in Nutrition 40. Washington, DC. IIPS (International Institute for Population Sciences). 1995. National Family Health Survey (NFHS ­1), 1992­93, India. Mumbai. IIPS (International Institute for Population Sciences), and Orc Macro. 2000. National Family Health Survey (NFHS­2), 1998­99, India. Mumbai. Johri, N. 2004. Effect of Integrated Program Design on Child Health Inputs and Outcomes: Estimates from a Nutrition and Health Program in India. PhD dissertation, University of North Carolina, Department of Health Policy and Administration, Chapel Hill. Jones, G., R.W. Steketee, R.E. Black, Z.A. Bhutta, S.S. Morris, and the Bellagio Child Survival Study Group. 2003. "How Many Child Deaths Can We Prevent this Year?" Lancet 362 (9377): 65­71. Jonsson, U. 1993. "Integrating Political and Economic Factors within Nutrition- Related Policy Research: An Economic Perspective." In The Political Economy of Food and Nutrition Policies, ed. P. Pinstrup-Andersen, 193­205. Baltimore, MD: Johns Hopkins University Press. 110 · Bibliography Kapil, U., P. Pathak, M. Tandon, C. Singh, R. Pradhan, and S.N. Dwivedi. 1999. "Micronutrient Deficiency Disorders amongst Women in Three Urban Slum Communities of Delhi." Indian Pediatrics 36 (10): 983­89. Karim, R,, S.A. Lamstein, M. Akhtaruzzaman, K.M. Rahman, and N. Alam. 2003. The Bangladesh Integrated Nutrition Project: Endline Evaluation of the Community-Based Nutrition Component. Institute of Nutrition and Food Science, University of Dhaka, Dhaka. King, K.W., W. Fougere, A. Hilaire, R.E. Webb, W. Berggren, and G. Berggren. 1978. "Preventive and Therapeutic Benefits in Relation to Cost: Performance over Ten Years of Mothercraft Centers in Haiti." American Journal of Clinical Nutrition 31 (4): 679­90. Kulkarni, M.N., and Y.N. Patthabi. 1988. "Evaluation of the Effectiveness of ICDS in 7 Anganwadi Centres on the Health Status of Pre-School Children." Indian Jour- nal of Community Medicine 13 (2): 86­90. Levine, R., and the What Works Working Group. 2004. Millions Saved: Proven Success- es in Global Health. Center for Global Development, Washington, DC. Lokshin, M., M. Das Gupta, M. Gragnolati, and O. Ivaschenko. 2005. "Improving Child Nutrition: The Integrated Child Development Services in India." Develop- ment and Change 36 (4): 613­40. Lucas, A., M.S. Fewtrell, and T.J. Cole. 1999. "Fetal Origins of Adult Disease: The Hypothesis Revisited." British Medical Journal 319 (7204): 245­49. Macintyre, C.R., N. Kendig, L. Kummer, S. Birago, and N.M. Graham. 1997. "Impact of Tuberculosis Control Measures and Crowding on the Incidence of Tuberculous Infection in Maryland Prisons." Clinical Infectious Diseases 24 (6): 1060­67. Martorell, R., J.P. Habicht, C. Yarbrough, A. Lechtig, R.E. Klein, and K.A. Western. 1975. "Acute Morbidity and Physical Growth in Rural Guatemalan Children." American Journal of Diseases in Children 129 (11): 1296­1301. Mason, J.B., P. Musgrove, and J.P. Habicht. 2003. "At Least One-Third of Poor Coun- tries' Disease Burden Is Due to Malnutrition." Working Paper 1, National Insti- tutes of Health, Fogarty International Center, Disease Control Priorities Project, Bethesda, MD. Mason, J., A. Bailes, M. Beda-Andourou, N. Copeland, T. Curtis, M. Deitchle, L. Fos- ter, M. Hensley, P. Horjus, C. Johnson, T. Lloren, A. Mendez, M. Munoz, J. Rivers, and G. Vance. 2005. "Recent Trends in Malnutrition in Developing Regions: Vitamin A Deficiency, Anemia, Iodine Deficiency, and Child Under- weight." Food and Nutrition Bulletin 26 (1): 59­108. Bibliography · 111 Mata, L.J., R.A. Kromal, J.J. Urrutia, and B. Garcia. 1977. "Effect of Infection on Food Intake and the Nutritional State: Perspectives as Viewed from the Village." American Journal of Clinical Nutrition 30 (8): 1215­27. Measham, A.R., and M. Chatterjee. 1999. Wasting Away: The Crisis of Malnutrition in India. Washington, DC: World Bank. Ministry of Agriculture. 2002. Agricultural Statistics at a Glance 2003. Government of India, New Delhi. Ministry of Industry. 2000. Sustaining Elimination of Iodine Deficiency Disorders: Univer- sal Salt Iodization in India. Historical Overview and Future Strategies. Government of India, New Delhi. Mkenda, A.F. 2004. "The Benefits of Malnutrition Interventions: Empirical Evidence and Lessons from Tanzania." Unpublished manuscript. Moe, C.L., M.D. Sobsey, G.P. Samsa, and V. Mesolo. 1991. "Bacterial Indicators of Risk of Diarrhoeal Disease from Drinking Water in the Philippines." Bulletin of the World Health Organization 69 (3): 305­17. Murray, C.J.L., and A.D. Lopez. 1997. "Global Mortality, Disability and the Contribu- tion of Risk Factors: Global Burden of Disease Study." Lancet 349 (9063): 1436­42. NCAER (National Council of Applied Economic Research). 2001. Concurrent Evalua- tion of Integrated Child Development Services. New Delhi. New Concept Information Systems. 2004. "Literature Review of MDM, ICDS, and PDS (1992­2003), Including Annotated Bibliography." Background paper for this report. NFHS (National Family Health Survey) II database. Mumbai: International Institute for Population Sciences (IIPS). NIPCCD (National Institute of Public Cooperation and Child Development). 1992. National Evaluation of Integrated Child Development Services. New Delhi. NNMB (National Nutrition Monitoring Bureau). 2002. Diet and Nutritional Status of Rural Population. Hyderabad: NNMB. NSS (National Sample Survey Organisation). 1997. NSS 53rd Round 1997. New Delhi. ------. 2001. NSS 56th Round 2000­2001. New Delhi. ------. 2002. NSS 57th Round 2001­2002. New Delhi. Nutrition Foundation of India. 1991. Growth Performance of Affluent Indian Children. Scientific Report No. 11. New Delhi. 112 · Bibliography Orc Macro. 2004. Measure DHS Statcompiler. http://www.measuredhs.com/statcompiler/ start.cfm?action=new_table&userid=170232&usertabid=186282&CFID=659299& CFTOKEN=41841064. Pandav, C.S., K. Anand, S. Gupta, and G.V. Murthy. 1998. "Cost of Vitamin A and Iron Supplementation to `At Risk' Population." Indian Journal of Pediatrics 65 (6): 849­56. Pelletier, D.L., and E.A. Frongillo. 2003. "Changes in Child Survival Are Strongly Associated with Changes in Malnutrition in Developing Countries." Journal of Nutrition 113 (1): 107­19. Pelletier, D.L., E.A. Frongillo, Jr., and J.P. Habicht. 1993. "Epidemiologic Evidence for a Potentiating Effect of Malnutrition on Child Mortality." American Journal of Public Health 83 (8): 1130­33. Pelletier, D.L., E.A. Frongillo, D.G. Schroeder, and J.P. Habicht. 1995. "The Effects of Malnutrition on Child Mortality in Developing Countries." Bulletin of the World Health Organization 73 (4): 443­48. Planning Commission. 2003. Annual Report 2002­2003. Government of India, New Delhi. Popkin, B.M., S. Horton, S.W. Kim, A. Mahal, and J. Shuigao. 2001. "Trends in Diet, Nutritional Status, and Diet-Related Non-Communicable Diseases in China and India: The Economic Costs of the Nutrition Transition." Nutrition Reviews 59 (12): 379­90. Radhakrishna, R., C. Ravi, and S. Indrakant. 1998. Cost Effectiveness and Efficiency of the Nutrition Programmes in India. Center for Economics and Social Studies, Hyderabad. Ramalingaswami V., U. Jonson, and J. Rohde. 1997. "The Asian Enigma." In The Progress of Nations. New York: UNICEF. Ramaswami, B. 2002. "Efficiency and Equity of Food Market Interventions." Economic and Political Weekly 37 (12): 1129­35. Rawlings, L., and G. Rubio. 2003. "Evaluating the Potential of Conditional Cash Transfer Programs." World Bank Research Observer 20 (1): 29­55. Ross, J., and S. Horton. 1998. Economic Consequences of Iron Deficiency. Micronutrient Initiative, Ottawa. Ross, J.S., and E.L. Thomas. 1996. Iron Deficiency Anemia and Maternal Mortality. Pro- files 3, Working Notes Series No. 3. Academy for Education Development, Washington, DC. Roy, S.K. 1997. "Complementary Feeding in South Asia." In Malnutrition in South Asia: A Regional Profile, ed. S. Gillespie, 51­73. Kathmandu: UNICEF Regional Office for South Asia. Bibliography · 113 Saiyed, F. and A. Srivastava. 2005. "Project Dular (Love) Goes to More Children in Jharkhand." UNICEF India. http://www.unicef.org/india/health_963.htm. Schürch, B., and N.S. Scrimshaw. 1989. Activity, Energy Expenditure and Energy Require- ments of Infants and Children. Proceedings of an International Dietary Energy Con- sultancy Group Workshop held in Cambridge, MA, November 14­17. Scrimshaw, N.S., ed. 1995. Community-Based Longitudinal Nutrition and Health Studies: Classical Examples from Guatemala, Haiti and Mexico. Boston: International Nutri- tion Foundation for Developing Countries. Scrimshaw, N.S. and J.P. SanGiovanni. 1997. "Synergism of Nutrition, Infection, and Immunity." American Journal of Clinical Nutrition 66 (2): 464S­77S. Scrimshaw, N.S., M.A. Guzmán, M. Flore, and J.E. Gordon. 1968. "Nutrition and Infection Field Study in Guatemalan Villages, 1959­1964: V. Disease Incidence among Preschool Children under Natural Village Conditions, with Improved Diet and Medical and Public Health Services." Archives of Environmental Health 16: 223­34. Seshadri, S. 2001. "Prevalence of Micronutrient Deficiency Particularly of Iron, Zinc and Folic Acid in Pregnant Women in South East Asia." British Journal of Nutri- tion 85, Suppl. 2: S87­S92. Sethi, D., P. Cumberland, M.J. Hudson, L.C. Rodrigues, J.G. Wheeler, J.A.Roberts, D.S. Tompkins, J.M. Cowden, and P.J. Roderick. 2001. "A Study of Infectious Intestinal Disease in England: Risk Factors Associated with Group A Rotavirus in Children." Epidemiology and Infection 126 (1): 63­70. Shekar, M., M. de Onis, M., M. Blössner, and E. Borghi. 2004. "Will Asia Meet the Nutrition Millennium Development Goal? And Even if it Does, Will it Be Enough?" Background paper for this report. Shrimpton, R., C.G. Victora, M. De Onis, R.C. Lima, M. Blossner, and G. Clugston. 2001. "Worldwide Timing of Growth Faltering: Implications for Nutritional Interventions." Pediatrics 107 (5): E75. SIDA (Swedish International Development Cooperation Agency). 2000. Reaching Out to Children in Poverty: The Integrated Child Development Services in Tamil Nadu, India. Department of Democracy and Social Development, Stockholm. Singh, P., and G.S. Toteja. 2003. "Micronutrient Profile of Indian Children and Women: Summary of Available Data for Iron and Vitamin A." Indian Pediatrics 40 (5): 477­99 Skoufias, E. 2001. Progresa and its Impacts on the Human Capital and Welfare of Households in Rural Mexico: A Synthesis of the Results of an Evaluation by IFPRI. International Food Policy Research Institute, Washington, DC. 114 · Bibliography Smith, L., and L. Haddad. 2000. Explaining Child Malnutrition in Developing Countries: A Cross-Country Analysis. Research Report 111, International Food Policy Research Institute, Washington, DC. Stephensen, C.B. 2003. "Vitamin A, (Beta)-Carotene, and Mother-to-Child Transmis- sion of HIV." Nutrition Reviews 61 (8): 280­84. Subramaniyam, P. 2004. "Case Study on Mid-Day Meal Scheme of Tamil Nadu and Gujarat." Background paper for this report. UNICEF (United Nations Children's Fund). 1990. Strategy for Improved Nutrition of Children and Women in Developing Countries. UNICEF Policy Review 1990­1 (E/Icef/1990/L.6). New York: UNICEF. ------. 1998. The State of the World's Children 1998: Focus on Nutrition. New York: Oxford University Press. ------. 2003a. The State of the World's Children. 2003. New York: UNICEF. ------. 2003b. UNICEF Statistics: Malnutrition. New York: UNICEF. ------. 2004. A Glance: India. http://www.unicef.org/infobycountry/india.html. UNICEF (United Nations Children's Fund) and MI (Micronutrient Initiative). 2004a. Vitamin and Mineral Deficiency: A Global Damage Assessment Report. http://www. micronutrient.org/resources/publications.asp. ------. 2004b. Vitamin and Mineral Deficiency: A Global Progress Report. Vasundhara, M.K., and B.N. Harish. 1993. "Nutrition and Health Education through ICDS." Indian Journal of Maternal and Child Health 4 (1): 25­26. Vijayaraghavan, K. 2002. "Control of Micronutrient Deficiencies in India: Obstacles and Strategies." Nutrition Reviews 60 (5) Suppl.1: 73­76. Vijayaraghavan, K., and H.D. Rao. 1998. "Diet and Nutrition Situation in Rural India." Indian Journal of Medical Research 108: 243­45. Vinutha, B., M.N. Mehta, and P. Shanbag. 2000. "Vitamin A Status of Pregnant Women and Effect of Post-Partum Vitamin A Supplementation." Indian Pediatrics 37 (11): 1188­93. Wagstaff, A., and M. Claeson. 2004. Rising to the Challenges: The Millennium Develop- ment Goals for Health. Washington, DC: World Bank. West, K.P. 2002. "Extent of Vitamin D Deficiency among Preschool Children and Women of Reproductive Age." Journal of Nutrition 132 (9): 2857s­66s. Bibliography · 115 Whang, I. 1981. Management of Rural Change in Korea. Seoul: Seoul National Universi- ty Press. WHO (World Health Organization). 1995. Physical Status: The Use and Interpretation of Anthropometry. Report of a WHO Expert Committee. WHO Technical Report Series 854, Geneva. ------. 2000. Nutrition Profile of the WHO: South-East Asia Region. Regional Office for South-East Asia, New Delhi. ------. 2002. World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva. ------. 2004a. Global Database on Iodine Deficiency Disorders: India. Geneva. ------. 2004b. Micronutrient Deficiency. Battling Iron Deficiency Anaemia: The Challenge. Geneva. ------. 2004c. WHO Global Database on Child Growth and Malnutrition. Geneva. World Bank. 1998. Implementation Completion Report on the Second Tamil Nadu Integrated Nutrition Project. Operations Evaluation Department. Washington, DC. ------. 2003. Mid-Term Survey for the Andhra Pradesh Economic Restructuring Project (APERP)-Nutrition/ICDS Component. Washington, DC. ------. 2004a. Attaining the Millennium Development Goals in India: How Likely and What Will It Take to Reduce Infant Mortality, Child Malnutrition, Gender Disparities and Hunger-Poverty and to Increase School Enrollment and Completion. Washington, DC. ------. 2004b. Global Development Finance 2004: Harnessing Cyclical Gains for Develop- ment. Washington, DC. ------. 2004c. India Burden of Disease Statistics. Washington, DC. ------. 2004d. Reaching Out to the Child. An Integrated Approach to Child Development. New Delhi: World Bank ------. 2004e. World Development Indicators 2004. Washington, DC. Index Note: b, f, t, and n indicate boxes, definition of malnutrition and figures, tables, and notes. undernutrition, 2b, 99n1 economic growth, effects of, 28­30, 30f age effects of, 4­10 ICDS targeting of children by, interstate and within-state variations 48­49, 48f, 81­82 in (See interstate and within- weight distributions by, 1­2, 2f, 92f state variations) (See also underweight) interventions projected to reduce, 30, anganwadi centers, 38 31f better nutritional status and MDG and, 28­30, 100n12 attendance at, 42 micronutrient deficiencies (See food supplies at, 82 micronutrient deficiencies) interstate and within-state variations prevalence of, 1­4, 3f, 4f in attendance at, 51f underweight (See underweight) monitoring and evaluation of, 59­63, childcare and feeding behaviors, 38, 79­80 60b, 102­7 cognitive development and malnutrition, personnel training, workload, and 4­5, 8­9, 10t status, 57­58, 82 communicable diseases safety and hygiene standards at, as cause of malnutrition, 35­38 55­57, 56t control and prevention measures, 80 service delivery characteristics and as effect of malnutrition, 5­6 quality, 52­59, 78 community-based interventions, 65­66, 85­87 Care India's Integrated Nutrition and crowding, 37 Health Project II (INHP II), 64­65, 96­97t data collection and information systems caste and tribe of ICDS, use of, 61, 63, 84­85 ICDS targeting of beneficiaries by, decentralization of ICDS program, 85­87 49­50, 49f demographic and socioeconomic prevalence of underweight, and characteristics. See also specific malnutrition by, 14f, 20 characteristics, e.g. gender causes of child malnutrition, 35­40, 35f ICDS targeting of beneficiaries by, child malnutrition in India, 31­32 47­51 causes of, 35­40, 35f malnutrition levels, variations in, 17­21 117 118 · Index design of ICDS program rationalizing, 87­89 mismatch between implementation income levels and, 33­34, 72­73, 78­82 ICDS targeting of beneficiaries by, rationalizing, 87­89 50­51, 50f underlying causes of child increasing ICDS coverage in poorer malnutrition and, 38­40, 39f states and districts, 82 developmental effects of undernutrition iron deficiency anemia and, 22­23 and malnutrition, 4­5, 8­9, 10t prevalence of underweight, and diarrhea, 5, 37 malnutrition by, 20­21, 20f, 21t dietary intake as cause of malnutrition, infection 35, 36 as cause of malnutrition, 35­38 disease burden control and prevention measures, 80 as cause of malnutrition, 35­38 as effect of malnutrition, 5­6 control and prevention measures, 80 information systems of ICDS, use of, 61, as effect of malnutrition, 5­6 63, 84­85 Dular Program, 65­66 INHP II (Integrated Nutrition and Health Project II), 64­65, 96­97t economic growth Integrated Child Development Services affected by malnutrition levels, Program (ICDS) 28­30, 29t, 30f anganwadi centers (See anganwadi effects of malnutrition on, 8­9, 99n4 centers) underweight, incidence of, 91 conceptual framework of underlying evaluation and monitoring of ICDS causes of child malnutrition, program and anganwadi 35­40, 35f centers, 59­63, 60b, 102n7 design of program mismatch between feeding and childcare behaviors, 38, 79­80 implementation and, 33­34, folate deficiency, 5t 72­73, 78­82 food security, 36­37 rationalizing, 87­89 underlying causes of child gender malnutrition and, 38­40, 39f ICDS targeting of beneficiaries by, 49 impact of prevalence of underweight, and empirical findings on, 40­42, 43t malnutrition by, 14f, 19t enhancement of, need for, 71­72 geographic variations. See interstate and menu of options for increasing, within-state variations 73­87, 74­77t growth-monitoring equipment and implementation of program activities, 52­53, 52f, 83 mismatch between design and, 33­34, 72­73, 78­82 high-risk groups targeted by TINP, 68­69 rationalizing, 87­89 lessons learned from project studies, ICDS. See Integrated Child 64­69 Development Services mismatch between intentions and Program implementation, 33­34, 72­73, IFPRI (International Food Policy 78­82 Research Institute), 102n32 monitoring and evaluation of, 59­63, implementation of ICDS program 60b, 102n7 mismatch between design and, 33­34, personnel 72­73, 78­82 monitoring and evaluation, 62­63 Index · 119 training, workload, and status of prevalence of, 3 anganwadi workers, 57­58, 82 protein-energy malnutrition, 6 political commitment and leadership, underlying causes and ICDS program obtaining, 78 design, 38­40, 39f Reproductive and Child Health Program, collaboration with, malaria, 5 58­59, 64­65 malnutrition. See child malnutrition in research on India; pregnant and lactating data collection and information women, malnutrition of systems, 61, 63, 84=85 MDG (Millennium Development impact of program, empirical Goals), 28­30, 100n12 findings on, 40­42, 43t measles, 5 lessons learned from project mental development affected by studies, 64­69 undernutrition, 4­5, 8­9, 10t service delivery characteristics and micronutrient deficiencies, 3, 5t, 6­8, quality, 52­59, 78 21­28 targeting folate deficiency, 5t geographic, 42­47, 44­47f, 94f, iodine deficiency, 3, 5t, 6, 8, 27­28, 27f 94t, 95t iron deficiency anemia, 3, 5t, 7, high-risk groups targeted by 21­24, 23f, 93t TINP, 68­69 Vitamin A deficiency, 3, 5t, 7, 24­26, individual targeting of 25f, 26f, 99n2 beneficiaries, 47­51 micronutrient supplementation, 80 World Bank and, 40 Millennium Development Goals Integrated Nutrition and Health Project (MDG), 28­30, 100n12 II (INHP II), 64­65, 96­97t monitoring and evaluation of ICDS International Food Policy Research program and anganwadi Institute (IFPRI), 102n32 centers, 59­63, 60b, 102n7 interstate and within-state variations morbidity and mortality, effects of in attendance at anganwadi centers, 51f undernutrition and decentralization of ICDS program to malnutrition on, 4­5, 6 address, 85­87 mothers committees, 66­68 ICDS program, geographic targeting motor development and undernutrition, of, 42­47, 44­47f, 94f, 94t, 95t 4­5, 8­9, 10t increasing ICDS coverage in poorer states and districts, 82 night blindness caused by Vitamin A iodine deficiency, 27­28 deficiency, 24­26, 99n2, 100n9 iron deficiency anemia, 24 underweight, 14f, 16f, 17­21, 17t, 18f overcrowding, 37 urban-rural locations, 18­19, 18f, 51 Vitamin A deficiency, 25 personnel interventions projected to reduce child monitoring and evaluation, 62­63 malnutrition in India, 30, 31f training, workload, and status of iodine deficiency, 3, 5t, 6, 8, 27­28, 27f anganwadi workers, 57­58, 82 iron deficiency anemia, 3, 5t, 7, 21­24, physical development and 23f, 93t undernutrition, 4­5, 8­9, 10t pneumonia, 5, 6 lactating and pregnant women, political commitment and leadership for malnutrition of ICDS program, obtaining, 78 120 · Index poverty. See income levels targeting by ICDS program pregnant and lactating women, geographic targeting, 42­47, 44­47f, malnutrition of 94f, 94t, 95t prevalence of, 3 high-risk groups targeted by TINP, protein-energy malnutrition, 6 68­69 underlying causes and ICDS program individual targeting of beneficiaries, design, 38­40, 39f 47­51 principal component analysis, 100n8 TINP (Tamil Nadu Integrated Nutrition productivity of adults, effects of Program), 40, 68­69, 103n35 undernutrition and tribe and caste malnutrition on, 8­9, 10t ICDS targeting of beneficiaries by, protein-energy malnutrition, 6, 99n3 49­50, 49f prevalence of underweight, and Reproductive and Child Health malnutrition by, 14f, 20 Program, ICDS collaboration tuberculosis, 37 with, 58­59, 64­65 research on ICDS underlying causes of child malnutrition, data collection and information 35­40, 35f systems, 61, 63, 84­85 undernutrition. See child malnutrition in impact of program, empirical India; pregnant and lactating findings on, 40­42, 43t women, malnutrition of lessons learned from project studies, underweight, 10­21 64­69 economic growth and incidence of, 91 rotavirus group A, 37 international perspective on, 10­12, rural-urban locations 10t, 12t ICDS targeting of beneficiaries by, 51 interstate and within-state variations, malnutrition disparities, 18­19, 18f 14f, 16f, 17­21, 17t, 18f patterns and trends in, 12­17, 13t, sanitation and sewage, 37 14­15f, 15t schooling, effects of undernutrition and weight-for-age distributions, 1­2, malnutrition on, 8­9 2f, 92f service delivery characteristics and United Nations Children's Fund quality of ICDS program, (UNICEF), 65 52­59, 78 urban-rural locations socioeconomic and demographic ICDS targeting of beneficiaries by, characteristics. See also specific 51 characteristics, e.g. gender malnutrition disparities, 18­19, 18f ICDS targeting of beneficiaries by, 47­51 Vitamin A deficiency, 3, 5t, 7, 24­26, malnutrition levels, variations in, 17­21 25f, 26f, 99n2 "South Asian enigma," 11b state, variations by. See interstate and water supply, 37 within-state variations wealth. See income levels supplementary nutrition and feeding, weight. See underweight 53­55, 54f, 80­81 women, malnutrition of prevalence of, 3 Tamil Nadu Integrated Nutrition protein-energy malnutrition, 6 Program (TINP), 40, 68­69, underlying causes and ICDS program 103n35 design, 38­40, 39f World Bank and ICDS, 40 ECO-AUDIT ENVIRONMENTAL BENEFITS STATEMENT The World Bank is committed to preserving endan- gered forests and natural resources. We have chosen to print India's Undernourished Children: A Call for Reform and Action on recycled paper with 30 percent post-consumer fiber. The World Bank has formally agreed to follow the recommended standards for paper usage set by the Green Press Initiative, a non- profit program supporting publishers in using fiber that is not sourced from endangered forests. For more information, visit www.greenpressinitiative.org. The printing of these books on recycled paper saved the following: · 8 trees (40' in height, 6­8 inches in diameter) · 371 pounds of solid waste · 2,888 gallons of water · 696 pounds of net greenhouse gases · 6 million BTUs of total energy T he prevalence of child undernutrition in India is among the highest in the world, nearly double that of Sub-Saharan Africa, with dire consequences for morbidity, mortality, productivity, and economic growth. Drawing on qualitative studies and quantitative evidence from large household surveys, India's Undernourished Children: A Call for Reform and Action explores the dimensions of child undernutrition in India and examines the effectiveness of the Integrated Child Development Services (ICDS) program, India's main early child development intervention, in addressing it. Although levels of undernutrition in India declined modestly during the 1990s, the reductions lagged behind those achieved by other countries with similar economic growth. Nutritional inequalities across different states and socioeconomic and demographic groups remain large. Although the ICDS program appears to be well designed and well placed to address the multi- dimensional causes of undernutrition in India, several problems exist that prevent it from reaching its potential. The book concludes with a discussion of a number of concrete actions that can be taken to bridge the gap between the policy intentions of ICDS and its actual implementation. ISBN 0-8213-6587-8