India - Tamil Nadu Nutrition Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 9259 PROJECT COMPLETION REPORT INDIA TAMIL NADU INTEGRATED NUTRITION PROJECT (CREDIT 1003-IN) DECEMBER 28, 1990 Population, Human Resources, Urban & Water Operations Division Country Department IV Asia Regional Office This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  ABBREVIATIONS CHC - Community Health Center CNC - Community Nutrition Center CNI - Community Nutrition Instructress CNS - Community Nutrition Supervisor CNW - Community Nutrition Worker DEAR - Department of Evaluation and Applied Research FST - Ferrous sulphate tablets FY - Fiscal year HSC - Health Sub-Center ICDS - Integrated Child Development Sevices IAS - Indian Administrative Service MCH - Maternal and Child Health NMP - Noon Meals Program OED - Operations Evaluation Department PCR - Project Completion Report PHC - Primary Health Center PHR - Population, Human Resources PPAR - Project Performance Audit Report SAR - Staff Appraisal Report SDR - Special Drawing Rights TINP - Tamil Nadu Integrated Nutrition Project USAID - US Agency for International Development FISCAL YEAR January 1st to December 31st  FOR OFFICIAL USE ONLY THE WORLD BANK Washington. D.C 20433 USA. Office of O,fectocGeneral OpVWatmns EvaluMtrn December 28, 1990 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report: India - Tamil Nadu Integrated Nutrition Proiect (Credit 1003-IN) Attached, for information, is a copy of a report entitled "Project Completion Report: India - Tamil Nadu Integrated Nutrition Project (Credit 1003-IN)" prepared by the Asia Regional Office. No audit of this project has been made by the Operations Evaluation Department by this time. Attachment This document has a restricted distribution and may be used by recipients only in the perfotmance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  FOR OMCIAL USE ONLY PROJECT COMPLETION REPORT INDIA TAMIL NADU INTEGRATED NUTRITION PROJECT (Credit 1003-IN) Table of Contents Paae No. Preface . . . . . . . . . . . . . . . . . . . . . . . . i Evaluation Summary . . .......... iii I. PROJECT REVIEW FROM BANK'S PERSPECTIVE A. Introduction.. . . . .1 B. Prolect DescriRtion . . . . . . . . . . . . . . . . . 1 Project Identity . . . . . . . . . . . . . . . . . . . 1 Background .. 2 Project Objectives and Strategy . . . . . . . . . . . 3 Project Components . . . . . . . . . . . . . . . . . . 4 C. Proiect Results .... . . . . . . . . . . . . . . . 5 Physical progress .... . . . . .. 5 Nutrition .... . . . . . . ..5... . . . . . . . S Health ..... . . . . . . .. 6 Cost-effectiveness .... . . . . . . 7 D. Proiect Strengths and Weaknesses: Lessons Learned . . 8 The Project Design and Implementation Process . . . . 8 The project design process . . . . . . . . . . . 8 The implementation process . . . . . . . . . . . 9 The Nutrition Component . . . . . . . . . . . . . . . 10 The worker: recruitment, tasks & routines . . . . 10 Growth monitoring ..10 Supplementation . . . . . . . . . . . . . . . . . 11 Food processing and distribution . . . . . . . . 11 Problems experienced in nutrition component . . . 12 Possible responses . . . . . . . . . . . . . . . 12 The Health Component . . . . . . . . . . . . . . . . . 14 Training and Supervision . . . . . . . . . . . . . . . 17 The Communications and Community Participation Component ..18 Monitoring and Evaluation . . . .19 Conclusions: Summary of Lessons Learned . .22 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Pate No. II. PROJECT REVIEW FROM THE BORROWER'S PERSPECTIVE . . . . . . . . . . . . . . . . . . . . . . . 23 III. STATISTICAL INFORMATION. ... 24 1. Related Bank Loans and/or Credits . . . . . . . . . . 24 2. Project Timetable . . . . . . . . . . . . . . . . . . 24 3. Credit Disbursements ..25 4. Project Results ..26 5. Project Costs and Financing ..27 6. Status of Covenants ..28 7. Use of Bank Resources ..29 ANNEXES 1. TINP'S Community Nutrition Center Summary Performance Chart . . . . . . . . . . . . . . . . 31 2. Thresholds Indicating Poor Performance . . . . . . . . . . 32 References ........................ 33  PROJECT COMPLETION REPORT INDIA TAMIL NADU INTEGRATED NUTRITION PROJECT (CREDIT 1003-IN) PREFACE This is the Project Completion Report (PCR) for the Tamil Nadu Integrated Nutrition Project ia India, for which Loan 1003-IN in the amount of US$ 32.0 million was approved on 15 April, 1980. The loan was closed on 31 March, 1989, two years behind schedule. It was almost fully disbursed, and the last disbursement was on October 17, 1989. The PCR was prepared by the Population and Human Resources, Urban and Water Operations Division of the Asia Regional Office (Preface, Evaluation Summary, Parts I and III). The Borrower was requested by the Bank to prepare Part II, but a reply has not been received. Preparation of this PCR was started after a national level workshop to evaluate the project, which was organized by the Government of Tamil Nadu in August, 1989. It is based, inter alia, on the Staff Appraisal Report; the Loan, Guarantee and Project Agreements; supervision reports; correspondence between the Bank and the Borrower; and independent evaluation studies of the project (see References).  - iii - PROJECT COMPLETION REPORT INDIA TAMIL NADU INTEGRATED NUTRITION PROJECT (CREDIT 1003-IN) E.VALUATION SUMMARY Proiect Obiectives 1. The main project objectives (para 2.07) were: (a) to halve malnutrition among children under four years of age; (b) to reduce infant mortality by 25%; (c) to reduce vi.--in A deficiency in the under fives from about 27% to about 5%; and (d) to raduce anaemia in pregnant and nursing women from about 55% to about 20%. Implementation Exy,;irience 2. Implementation experience was good. Village nutrition centers staffed by community workers were set up in about 9,000 villages of Tamil Nadu. Workers' skills and motivation were evaluated as very high. About 2,000 new Health Sub- Centers were constructed. Implementatic.. delays were not significant. The two year extension of the project was required to expend savings generated by Rupee devaluation, rather than to complete originally planned project activities (para 2.11). 3. Factors responsible for smooth implementation included the state's strong political commitment to the nutrition sector; the high quality of the managers assigned by the state to the project; the attention given by both Borrower and IDA staff to the design of project software systems during preparation; and the use of the first project year for field testing of these systems in a small area, before rapid subsequent expansion (paras 3.02-3.06). Results 4. With regard to the project's nutrition goals, it is estimated that TINP con'ributed to a reduction of a third or more in severe malnutrition among 6-24 month old, and contributed, together with another nutrition program simultaneously implemented by the state, to a reduction of about half among 6-60 month old. These achievements surpass those of most other nutrition programs. On the other hand, the project did not achieve its ambitious goal of reducing malnutrition as a whole by 50%. Even in the parts of Tamil Nadu benefiting longest from the project, severe plus moderate malnutrition together fell by 28% (from 53% to 38%) for under six children, and by only 10% for 6-36 month children; results were poorer elsewhere (paras 2.13-2.18). The main problem was failure to substantially reduce moderate malnutrition.  - iv - 5. The project's health goals were not satisfactorily met. Infant mortality went down by 12%-26Z in different areas of the project; but it is not clear how far these declines were caused by the project. Reductions in vitamin A deficiency and anaemia were not directly measured, but the targets were clearly not achieved. Process evaluation indicated that less than a third of under six children and pregnant and lactating women had been covered by the related project interventions. The main problem was lack of effective collaboration between health and nutrition workers in the field (paras 2.19-2.21 and 3.26-3.27). Sustainability 6. TINP is sustainable both institutionally and financially. By the close of the project, the state had created from scratch a nutrition infrastructure with excellent outreach through a skilled and motivated field staff and a group of managers with the commitment and capacity to objectively assess the strengths and weaknesses of their own program, and take corrective action. On the financial side, Tamil Nadu decided to continue to fund the project from its own resources after project closing, and at the time of writing has requested a second IDA credit to expand it state-wide (para 2.06). TINP's financial sustainability is indicated by the fact that its recurrent cost of about US$4.7m represented less than five percent of the state's spending on nutrition in 1989/90. Findings and Lessons Learned 7. During implementation, TINP attracted a great deal of interest from the international nutrition community. Despite the fact that it did not achieve all of its goals, the project has an unusual number of lessons for the design and implementation of other programs, some of which are already being applied. These lessons are partly technical ones related to nutrition and health. But perhaps the most important ones stem from the design of the project's software, which has a number of features which could be adapted to improve the coverage and quality of services in any of the 'people-oriented' sectors. 8. On the nutrition side, TINP showed that part time village workers, if properly trained and suppo.ted, can substantially improve their community's nutritional status. Second, the project demonstrated unambiguously that growth monitoring is an effective and feasible intervention in large scale nutrition programs. Third, the project showed that targeted supplementation based on growth faltering plus feeding of all severely malnourished children can significantly reduce severe malnutrition; but that a move to feeding all moderately malnourished children plus greater attention to maternal nutrition seems necessary to make a substantial impact on moderate malnutrition (paras 3.10-3.24). 9. On the health side, the main finding was that large scale investment in health infrastructure and supplies is not sufficient to improve performance. Complementary measures are needed on the software side which will maximize health and nutrition workers' incentive to cooperate, and make them more accountable for doing so (para 3.36). TIN's experience sug;ested six or seven such measures  -v- (paras 3.28-3.35), for experimentation in the proposed follow-on project. 10. TINP's most significant lessons lie in the design of the project's software on the nutrition side. Key features for replication include carefully defined recruitment criteria for local workers (para 3.07); limitation of field worker tasks to what is manageable and high priority (para 3.08); specification of daily and monthly work routines (para 3.09); the innovative decentralized training systet (para 3.37); supervisory ratios which facilitate on the job training (para 3.39); the use of local women's groups to support project activities (paras 3.06 and 3.42-4); the display of performance information to clients and workers at the village nutrition center (para 3.46); and a management info-mation system which could rapidly detect performers falling below established norms (para 3.48). 11. Overall, TINP can be viewed as an innovative and successful project, showing that large scale, community level interventions in the social sectors can work. The lessons from the project are significant ones, and can help to strengthen other outreach programs in India and elsewhere.  PROJECT COMPLETION REPORT INDIA TAMIL NADU INTEGRATED NUTRITION PROJECT (CREDIT 1003-IN) I. PROJECT REVIEW FROM BANK'S PERSPECTIVE A. Introduction 1.01 The Tamil Nadu Integrated Nutrition Project (TINP) is of special interest to the Bank and to the Indian and international nutrition community for several reasons. It is one of the first free-standing nutrition projects supported by IDA to be evaluated. Second, it is one of the largest nutrition education/targeted supplementary feeding programs in the world. Third, although the project did not achieve all of its planned goals, independent evaluation indicates that it has been one of the more successful efforts to date to reduce severe malnutrition. Fourth, the innovative design of TINP's software components offers many lessons for other projects and programs. 1.02 For the above reasons, this Project Completion Report is longer than the norm, and focuses particularly on the software and process aspects of the project. Part B of the report describes the sectoral environment in which the project was formulated, and the object' 's, content and results of the project. Part C focuses selectively on key design and implementation features which appear to have contributed to the project's achievements, and which may contain lessons for the design of other programs. It also focuses on aspects of the project which were less successful than planned, and the lessons from this experience. B. Project Description Proiect Identity Project Name Tamil Nadu Integrated Nutrition Project Credit No. 1003-IN Date of Credit Effectiveness August 5, 1980 Completion Date March 31, 1989 Closing Date March 31, 1989 Total Project Cost US$81.0 million equivalent Original Project Cost US$66.4 million equivalent Foreign Exchange Component (Appraisal Estimate) US$7.29 million equivalent IDA Credit US$32.0 million IDA Credit, Total Disbursed US$31.977 million IDA Credit, Total Canceled US$0.023 million RVP Unit Asia Country India Sector PHR Subsector Nutrition  -2- Background 2.01 India's concern about malnutrition dates from independence in 1947, and intensified as a result of the famine in the state of Bihar in 1966/7. Beginning witi the Fourth Plan (1968-73), the Government began to expand existing feeding programs as a way to help bridge the nutrition gap, while longer-term food production and income-generating meanures took hold. In 1975, it launched the Integrated Child Development Services Scheme on an experimental basis. This has now become the main child nutrition program, operating in about 40% of the country; however, in the late 19709, when TINP was developed, it covered only about six million people. By that time, public feeding programs of various types had expanded to reach an estimated 20 million beneficiaries. 2.02 While concern about malnutrition was reflected to greater or lesser degree in the policies and programs of all Indian states, Tamil Nadu's commitment to the nutrition sector has been particularly strong. The state commissioned studies in the mid and late 1970s, which showed that only one out of 16 major states was worse than Tamil Nadu in terms of average daily food intake; that 85% of under six children in the state were less than 90% of the recommended weight for their age; and that malnutrition was a leading or associated cause of more than 75% of the dea'hs of 0-3 year old children. By the early 1980s, Tamil Nadu had decided to substantially increase spending on nutrition, and expected this investment to make a significant contribution to the achievement of its Sixth Plan goals, which were to reduce economic and social inequalities; to raise incomes and employment; and to reduce the poverty and improve the quality of life of the weaker sections of the population. 2.03 By the late 1970s, about 25 nutrition programs were operating in Tamil Nadu, mostly under government auspices. The state was spending about US $8.8m. per annum on the sector, three quarters of which was accounted for by a school meals program reaching around two million children. However, all of the states's nutrition programs combined reached less than 10% of pre-school children, the group that studies had shown to be most affected by malnutrition. And various evaluation studies had shown that the existing nutrition programs were of limited success for the following reasons: (a) they did not reach intended beneficiaries because they failed to set or adhere to nutritional criteria with respect to target groups; (b) on-site feeding tended to substitute for food which beneficiaries would otherwise have consumed at home; (c) food taken home was shared with other family members, and thus diluted the impact on the beneficiary; (d) the supplement given was too coarse and bulky for very young children to eat; (e) feeding was mainly confined to children who came to centers on a  -3- drop-in basis, rather than through active efforts to identify those most at nutritional risk; (f) insufficient emphasis was given to nutrition education for mothers; and (g) food and education alone could not adequately improve nutritional status in the absence of complementary health care interventions. 2.04 A strong commitment to nutrition was thus combined with doubts about the appropriateness of the design of existing nutrition programs. In particular, there was increasing awareness in the State of the need to improve the targeting and lower the unit cost of its interventionn; to promote better nutrition and health practices within families; and to strengthen maternal and child health (MCH) services. This situation led to the request for IDA assistance to develop a nutrition project focusing on children in the first three years of life, the nutritionally most vulnerable group. This was only the fourth request for a free-standing nutrition project that the World Bank had at that time received. 2.05 IDA's program in India in the late 1970s focussed on the power and agriculture sectors. Of a total of US$7,206m. of IDA resources committed to India at the time TINP was appraised, only US$33m. had gone to the social sectors, for two credits for education and population which went to the Board in 1972. The latter credit, which closed in 1980 and was evaluated in a PPAR dated January 1982 (No. 3748), was implemented in selected districts of Uttar Pradesh and Karnataka. It included a small nutrition component, which was evaluated as not very successful. This was mainly because food was given on a take-home basis, and it was not clear how much was eaten by the malnourished child as opposed to shared with other family members. Nutrition components have not been incorporated in subsequent Bank-assisted projects in India. 2.K'5 In 1982, the Tamil Nadu Government initiated a second major nutrition initiative, the Chief Minister's Noon Meals Program, which supplies a daily rice- based meal to all children from three years and up. This had achieved state-wide coverage by the time the TINP credit closed ir. 1989, while TINP was operating in about half the state. At this point the state government decided to continue the operation of TINP using its own funds, and to approach IDA for a second credit which would expand TINP services to pre-school children in the rest of the state, and add several new features to the program to strengthen its impact in existing areas. Discussions on the design of the proposed follow-on project were taking place at the time of writing this report. Project Obiectives and StrategV 2.07 TINP's overall goal was to imDrove the nutrition and health conditions of pre-school children, with emphasis on those aged 6-36 months, and of pregnant and nursing women. The main project target was a 50% reduction in the incidence of protein-energy malnutrition from a level at appraisal of about 60% among children under three years of age. Additionally, the project aimed to  - 4 - contribute to the Government's goal of achieving the following objectives in the project area: a) a 252 reduction in the infant mortality rate, then about 125 per 1,000; b) reduction in the incidence of Vitamin A deficiency in children under five, from about 27% to about 5%; c) reduction in the incidence of nutritional anaemia in pregnant and nursing women, from about 55% to about 20%. 2.08 The main project strategies were to provide nutrition education and primary health care to pregnant and lactating women and children 6-36 months; to monitor the growth of children in this age group through monthly weighing and growth charting; and to provide supplementary feeding and health checks to those children whose growth was found to be faltering, as well as intensive counselling to their mothers. This was to be managed primarily through a new cadre of part time, paraprofessional Community Nutrition Workers. These were to be assisted by local women's groups created under the project, together with the existing health outreach and referral serviceo, which would be strengthened under the project. Monitoring and evaluation were to have an important role, since the Government wished to see whether, through the project, it would be possible to increase the effectiveness and reduce the unit costs of its nutrition programs. 2.09 The project was planned to take place in the rural areas of six districts of the state (later to become ten districts as a result of administrative bifurcation), chosen for having the worst nutritional status. Geographically, these districts covered about half the state, and a rural population of about nine million people. The project was to be implemented over five years, with a phased expansion on a block by block basis (the block is an Indian administrative unit covering about 100,000 people). Proiect ComRonents 2.10 The total project cost at appraisal of US $66.4m. was allocated among five components as follows: Nutrition Service Delivery (47%), Health Service Delivery (43%), Communications (6%), Monitoring and Evaluation (2%), and Project Coordination (2%). Project costs by category were distributed as follows: Salaries and Allowances (24Z), Civil Works (19%), Vehicles, Equipment and Furniture (14%), Food Supplement (13%), Drugs and Supplies (12%), Training (6Z), Vehicle and Other Operating Costs (6%), Contract Services (5%) and Project Management Fund (1%). The following main items and activities were financed under the different components: (a) Nutrition. Staff, rent, furniture and equipment for 9,000 newly created Community Nutrition Centers (CNCs); supervisory staff, accommodation and transport; running costs of supervisors' offices and CNCs, including therapeutic food for beneficiaries; (b) Health. Construction, furniture, equipment, salary and running costs for 1,600 Health Sub-Centers (HSCs), for ten institutions for classroom training of Multi-Purpose Health Workers and their supervisors, and for field training wings at 39 Primary Health Centers; bicycles, motorcycles and vehicles for health staff and supervisors; drugs, vaccines and supplies;  - 5 - (c) Communications. Costs of designing, producing and disseminating films, filmstrips, radio broadcasts, folk plays and written materials for nutrition and health education campaigns and for training neighborhood leaders. (d) Monitorinz and Evaluation. Monitoring and evaluation offices in each project district, and costs of baseline and special studies, data processing and routine monitoring activities. (e) Project Coordination. Rent, furniture, equipment, staffing and other operating costs of the project coordination office; funds for innovative activities. 2.11 During project implementation, the Dollar and SDR appreciated substantially against the Rupee, with the result that the IDA credit was able to generate significantly more Rupee funds than had been expected at appraisal. The Government therefore proposed the following additional investments to utilize these 'savings': Construction of two food supplement production plants; construction, equipping and staffing of 410 additional Health Sub-Centers; development of a Communications Center in Madras to serve as a state level resource for the planning, monitoring and evaluation of educational work in the nutrition sector. Later economic analysis led to questions about the viability of the food production plants, and these were not constructed; the remaining additional investments were implemented. C. Project Results 2.12 Physical progress. The project's closing date was extended from March 1987 to March 1989 in order to fully utilize the additional project funds generated by currency fluctuations. This extension--which was also the cause of the project's Rupee cost overrun--did not reflect delays in implementation, since the activities originally included in the project were completed more or less according to the original schedule. 2.13 Nutrition. An independent evaluation by the Tamil Nadu Government's Department of Evaluation and Applied Research (Chidambaram, 1989) compared results in four areas coming on stream in different phases of the project's development against the situation at the start of the project. This evaluation, referred to henceforth in this report as the DEAR evaluation, showed reductions in severe malnutrition (grades III and IV, i.e. less than 60Z of normal weight for age) varying from 402 to 552 among children 6-60 months old (Vol II, table 4). However, it is not clear how much of this reduction was caused by the project, how much by secular trends, and how much by benefits from the daily meal supplied to 25-60 month children by the IMP (para 2.06). Neither of the latter important influences was addressed by the DEAR evaluation. 2.14 It is very difficult to disentangle the independent effects of TINP with the data available. With regard to the effects of TINP as opposed to the NMP, DEAR data indicate (Vol II, table 2) that among 6-36 month children severe  - 6 - malnutrition fell by between 26Z and 42% in different project areas. While these data suggest a strong independent impact of TINP, they cannot be conclusive, because they still contain the 25-36 month age group also fed by the NMP. An independent assessment by Shekar (1989) made a direct comparison, and concluded that reduction in severe malnutrition was greater among 6-24 month than 25-36 month children; but this study was limited to one block. With regard to the impact of TINP as opprsed to developments in general in the state, the DEAR comparison of pilot and control blocks suggests a strong project impact over and above secular trends. But weaknesses in the evaluation design (para 3.52) make it impossible to quantify this with any confidence. 2.15 Further uncertainty is introduced by some evidence (Shekar, 1989) that the DEAR evaluation may underestimate the reduction in malnutrition achieved in the project area (para 3.53). Serious account should also therefore be taken of the project's own service statistics, which evaluators agree to be of unusually high quality. They may be a more relevant measure of project impact since they focus on nutrition status in villages covered by project services, rather than in larger population taken as the denominator by the DEAR evaluation. The monitoring data indicate a larger reduction of severe malnutrition in the 6- 36 month group, of between 36% and 66%. Again, of course, these figures include the 25-36 month age group fed by the NMP, and as with the DEAR data also reflect an unknown influence from secular trends. 2.16 Overall, it is probably fair to conclude that TINP, independently of the NMP but not independently of secular trends, achieved a reduction of a third to a half in severe malnutrition among 6-24 month children; and that the project contributed, together with the NMP and secular trends, to a reduction of about half among 6-60 month children. 2.17 There was less success in reducing moderate malnutrition (defined as between 70% and 60% of normal weight for age). In the second and last phases of the project, the DEAR evaluation indicates that second degree malnutrition was higher at project end than at the beginning. But this is inevitable, since a temporary rise in second degree malnutrition naturally follows improvement in third and fourth grade children. More significantly, however, by the end of the project grade two malnutrition in children under five had only come down by 14% in the first block to be developed, and still affected about 30% of children (DEAR, Vol. II, table 1). 2.18 TINP's overall nutrition objective was to reduce malnutrition in 6-36 month children by 50% during the project period; the appraisal report did not specify which grades, but it may be assumed that this meant moderate and severe malnutrition. According to the DEAR evaluation (Vol. II, tables 1 and 2), in the initial block to be developed moderate and severe malnutrition together fell only 10% in the 6-36 month age group, and by 28% in the 6-60 month group. The reduction was much less in later phases of the project, where the grade two population had--presumably temporarily--increased. 2.19 Health. As against the target of 25%, the DEAR evaluation indicated that infant mortality had declined by 26 to 29% in the blocks first developed,  and by 12 to 13% in the later phases of the projectl a rate for the project area as a whole was not calculated. The project therefore can be said to have contributed to meeting the state's mortality reduction target (this was how the project target was expressed), but it is not clear how far the project was actually responsible for achieving the above declines, since declines of the same average magnitude occurred in the non-project districts of the state; on the other hand, the TINP project districts were chosen because they were relatively disadvantaged, so that average mortality reduction figures may reflect more than average achievement. 2.20 Success in achieving the project's two other formal health objectives--reduction from 27% to 5% in Vitamin A deficiency in under five children, and reduction in anaemia from 55Z to 20% of pregnant and nursing women- -was not evaluated directly. Nevertheless, inferences can be drawn from coverage levels achieved in related interventions, which, according to the DEAR evaluation, had risen very significantly, but were nevertheless well short of what would have been needed to achieve the proje,zt targets. 55% of children received half or more of the due doses. 28% of children had not received a single dose. While at least half of all eligible pregnant women had received Ferrous Sulphate Tables (FST) against anaemia, only 18.5% had received the right quantity of FST. There is no data to indicate what percentage of women getting FST actually took the tables. 2.21 Unsatisfactory performance with these two interventions can partially be explained by periodic national shortages of Vitamin A, and quality problems with the FST, leading to their withdrawal at one point by health component managers. But similar performance was found at evaluation with other health interventions emphasized under the project (although not given formal targets). For example, 56% of antenatal cases were being registered by the end of the project; 45Z of eligible women received full anti-tetanus immunization; only 37% of children received regular health check-ups; and only 60% of children had received three DPT immunizations. These coverage rates are good by average Indian standards, and considerable better than in the non-project areas. But they were well below the rather ambitious, informal project targets of 80-90% for each of these services. 2.22 Cost-effectiveness. DEAR's evaluation deals with project impact, but not explicitly with the issue of cost-effectiveness, although this was a primary concern of the state government (para 2.04) at appraisal. One study (Dapice, 1988), commissioned during project implementation, attempted to compare the cost- effectiveness of TINP with the major national child development program, the Integrated Child Development Services (ICDS). The study concludedthat TINP had achieved roughly twice the reduction in severe malnutrition of ICDS for a smaller cost (about US$9.1 per direct beneficiary). However, this study needs to be treated with some caution, for three reasons. First, it compared only one TINP and one ICDS block. Second, it looked at effectiveness only in terms of reducing severe malnutrition, when the goals of both programs were to reduce malnutrition in general. Third, it is not clear that a comp&rison with ICDS is a fair one, because TINP was a nutrition/health program only, while ICDS is a child development program with broader goals, including pre-school education.  - 8 - 2.23 TINP nevertheless demonstrated that highly targeted feeding programs, complementing effective nutrition education, can have a substantial impact on severe malnutrition for far less money than universal feedings in an established TINP block, monitoring data show that no more than 251 of 6-36 month children were in feeding at any one time, with a corresponding saving in food supplement costs. The project does not, however, demonstrate that the TINP targeting model is a cost-effective formula for dealing with moderate as well as severe malnutrition, in view of the project's limited effects on the former. Experiments are currently going on in Tamil Nadu to determine the effects on moderate malnutrition of broadening the target group for supplementation (see also paras 3.22-3.24). The results should have important implications for the design of nutrition programs in the proposed follow-on project and in other countries. D. Proiect Strengths and Weaknesses: Lessons Learned 3.01 The third part of this report begins by reviewing the general design and implementation approaches that were important to TINP's achievements, and then analyses lessons learned component by component. Because of its importance, the training and supervision system is discussed as a separate component, even though it was not treated as such in the original project appraisal report. The Proiect Design and Implementation Process 3.02 The groiect design process. Project design was facilitated by an unusually strong nutrition information base, in the form of the USAID-financed Tamil Nadu Nutrition Study, completed in 1973 by a multi-disciplinary team of nutritionists, economists, biologists, food technologists, behavioral scientists, engineers and systems specialists. The study was apparently at the time the most systematic effort undertaken in a developing country to analyze the nutrition aspects of food production, distribution and consumption, with a particular focus on identifying interventions to improve child survival. The project also drew from experience with IDA-financed nutrition project then under implementation in Indonesia. Features incorporated from this project into TINP included the focus on the 0-3 age group, the use of part time community nutrition workers as the main change agents, and the calorie-dense food supplement. Finally, the design # oroject software was also influenced by the Bank-developed Training and Visit tem of agricultural extension, most notably with regard to highly defined work routines for field staff; frequent, decentralized in-service training; and reasonable spans of control for supervisory staff. 3.03 The final project design was the result of sixteen months of intensive dialogue between the Government and IDA teams, which resulted in substantial ref ineme- and simplification of the original project proposal. This had originally interaed the project to cover four states, which was felt by IDA to be too ambitious for % first operation in the sector. It had also incorporated village water supply and food production, processing and storage interventions, which vere later excluded from the project because it was felt that these would complicate project management and could better be tackled through sector-specific programs. Another result of the dialogue during  - 9 - preparation was the innovative food supplementation system, which focused on feeding very young children for relatively brief periods for growth recovery purposes, in contrast to existing programs which tended to provide prolonged feeding for older children. 3.04 The implementation process. Implementation was guided on a day to day basis by a newly created project coordination unit, headed by an Indian Administrative Service (IAS) officer who was also designated Additional Director Social Welfare. The IAS status and the rank of this officer in the Directorate were important factors in facilitating implementation; at the same time, the creation of a separate unit for managing the project meant that attention to it was not diffused in the press of managing existing departmental programs. The Project Coordinator reported to an Empowered Committee chaired by the State's Chief Secretary and consisting of the relevant Development Secretaries, which met periodically to review proL:ess and take major decisions. The high level and inter-departmental nature of this committee were important factors in achieving coordination and getting things done. 3.05 Project expansion was carefully planned. The project was confined to one administrative block in its first year, during which the implementation strategy was tested and refined. IDA's role was particularly important during this period, as it was during preparation, with three or four visits by interdisciplinary teams in a year, as against the six monthly reviews that became the routine in subsequent years. Project expansion after the first year was rapid. An important contribution to flexibility and innovation in project management both in the pilot block and the expansion phase was made by the Project Management Fund, from which small amounts could be committed for studies and operations research at the discretion of the project coordinator without bureaucratic clearance. This fund financed no less than 63 small studies which helped identify problems during implementation, and also financed an innovative program for food supplement production by village women's groups (see para. 3.14). 3.06 Project expansion was also carefully planned at the micro level, with special attention to preparing the local community whenever the project moved into a new village. After recruitment and training of the Community Nutrition Worker (CNW), but prior to the offering of any project services, a women's group was formed consisting of about 20 people with a particular interest in health and nutrition. This group was briefed on the aims of the project, and played an important part in explaining them in turn to the community. Two key elements of the program's philosophy were communicated at this stage, before services were even delivered. First, that the feeding program was to be purely rehabilitative in nature, with only needy children getting supplemented; the primary responsibility for child care and nutrition was to remain with the family. Second, that the project was the community's, and project staff accountable to local people. The CNW's supervisor made it clear at the initial briefing of the women's group that the community should approach her and complain if any of the promised services were not in fact delivered.  - 10 - The Nutrition Component 3.07 The worker: recruitment. tasks and routines. The criteria for recruitment of the CNWs were important to the project's success. They had to be womcn from the village (one was chosen for every 1,500 people), with an interest in health and nutrition; and where possible they were women who were poorer than average, but who had well nourished childt en. In addition to having good access to mothers and high motivation, such women tended to be credible sources of advice because they could demonstrate from their own experience that poor families could provide better nutrition for their children even with existing resources. These qualities were given higher weight in selecting workers than their educational levels. 3.08 The project design consciously limited CNWs' tasks to a small number which were manageable and likely to have a high impact on nutrition. These were growth monitoring; nutrition education; supplementary feeding; education in oral rehydration for diarrhea cases; twice yearly Vitamin A supplementation; and quarterly deworming. The growth monitoring, supplementation and deworming programs were limited to the under four age group both because these were the nutritionally most vulnerable, and because this group was small enough for workers to pay them and their mothers individual attention. 3.09 Workers' daily and monthly routines were clearly defined. In the mornings, they ran the supplementary feeding program. In the afternoons, they carried out home visits to families with problem children. Three predetermined days a month were set aside for child weighing, which on the first two days took place at specified locations in the village which were convenient to mothers. The third day was set aside for following up on mothers who had failed to bring their children for weighing. This outreach approach resulted in a high proportion of children being weighed every month--normally 85-90% of children in main villages, although substantial numbers of children in outlying hamlets were not reached (see para. 3.15). 3.10 Growth monitoring. A good deal of controversy persists about whether growth monitoring is a necessary and feasible intervention for large scale nutrition programs. TINP managers are convinced that regular growth monitoring was the core intervention in the program, for three reasons. First, the growth chart could detect faltering growth and incipient malnutrition long before it was visible to either mother or worker, allowing action to be taken before the problem became serious and hence more difficult to treat. Second, weighing and growth charting with the participation of mothers provided the central focus for nutrition education. Third, the child's charted growth path provided the basis for deciding whether supplementary feeding was necessary. On the question of feasibility, studies during project implementation consistently reported a high degree of accuracy in both taking weights and recording them. TINP has demonstrated unambiguously that growth monitoring is a feasible intervention for paraprofessional workers in large scale nutrition programs, providing that training and supervision are carefully planned (see paras 3.37 to 3.39).  - 11 - 3.11 Supplementation. After much debate during project preparation, a combination of nutrition grade and weight gain was used as the criterion for entry into the supplementary feeding program. All children with severe (grade 3 or 4) malnutrition were fed, as also all children who failed to gain weight in two successive monthly weighing, even if such children were in the normal category by grade. Children in the first category were graduated from feeding when they reached grade two and were on a normal growth path, children in the second category after three months, providing they were gaining weight normally. Attendance in the feeding program was evaluated as high, averaging about 95% of those eligible. 3.12 The weight gain method had three general advantages. First, the concept was educational in itself, demonstrating to mothers the importance of feeding more to maintair. growth momentum. Second, feeding growth faltering children early, even when they were in grade normal, could prevent the development of moderate and severe malnutrition. The fact that no less than 85% of children in the average TINP village suffered a growth setback and entered the feeding proeram at some time suggests that this preventive role may have been quite significant. Third, it was thought that this selection method would be cheaper, in terms of numbers in the feeding program, than supplementing all moderately malnourished children. With no more than 25% of children in the feeding program in the final year of the project, this was clearly the case. However, as noted above (para. 2.17), the project failed to achieve a substantial impact on moderate malnutrition, and it appears that this was in part because it was supplementing too few children. 3.13 The TINP food supplement was a ready-to-eat mix of wheat, gram, groundnut cake, sugar, maize and vitamin and mineral mix, delivering 380 calories a ration. It was served in the form of 'laddus', or small ball-shaped snacks. This type of supplement had three advantages. The use of locally consumed foods ensured that it was acceptable. Because the food was ready to eat, cooking costs and workers' time were saved. And, most importantly, because the food was served in the form of a snack, mothers tended to think of it as a supplement to the diet rather than as a meal which might substitute for existing consumption. This perception was reinforced by the timing of feeding at the Community Nutrition Center, which was in the first half of the morning, i.e. between meals. 3.14 Food processinR and distribution. Three quarters of the project's food requirement was supplied by a government-owned factory in Karnataka, another southern Indian state. The remainder was produced by 322 of the women's groups set up under the project (para. 3.06). Breaks in food supply from either source were not experienced during the project. The cost per kilo of food produced by village groups was similar to that of factory produced food (about Rs.7, or US$0.5). Village level production had the advantage of contributing also to income and empluyment opportunities for local women, as well as a significant educational spin-off; women in villages where the supplement was produced learned the recipe and began to make it for normal home consumption. On the other hand, problems were becoming apparent by the end of the project with quality control, and with accountability of project funds, and it is unlikely that this form of production will be expanded.  - 12 - 3.15 Problems experienced in nutrition component. Four main weaknesses in the nutrition component have been identified by various evaluators. First, it appears that up to 30% of eligible children in the project districts did not receive project services, especially regular weighing. This is a serious problem, because the excluded group will inevitably include cases of severe malnutrition which have not been detected. A recently conducted study on the characteristics of excluded beneficiaries concludes that the majority of exclusion is for geographical reasons--they live outside the current catchment area of the project--rather than self-exclusion by dropping out of the program. This probably explains why project monitoring data show 85-90% weighing rates (for villages covered by the project), while the DEAR evaluation (which sampled the entire rural population in the project districts) reports closer to 70%. 3.16 Second, just over four per cent of children failed to respond to project inputs, and remain in feeding. While this group of 'hard core malnourished' is small, it is very important, because it consists mainly of children who are severely malnourished and hence at high risk of dying. At the time of writing, a study was also being conducted to determine the characteristics of non-responders. The evidence available to date suggests that these are children with medical problems which have not been resolved and/or children from very poor families whose food resources are inadequate for the family even with the supplement provided by the project. 3.17 Third, as already discussed, the project did not achieve a rapid reduction in moderate malnutrition. Fourth, an unacceptably large number of children who graduated from feeding relapsed into supplemeL.tation within six months. A project financed study on relapse rates (Rajivan, undated) found that 42Z of the children that graduated relapsed once; 6% relapsed twice; less than 1% a third time. Thus, while the project succeeded in helping these children, they might have recovered sooner to a normal growth path had they received a longer, unbroken period of supplementary feeding. 3.18 Possible responses. At the time of writing, six possible responses to the above problems had been proposed for implementation in the next project. First, it was accepted that TINP's interventions were focused on the 6-36 month old child, and paid too little attention to children's nutrition in the womb and in the first six months of life. Due to poor maternal nutrition, about 302 of babies are born under weight, and hence at higher risk of mortality, morbidity and persistent malnutrition. A better focus on maternal nutrition and on child care during the first six months would produce bigger, stronger infants less prone to malnutrition in general, and in particular to the problems of non- response, relapse and failure to move out of grade two. 3.19 Under the TINP project design, the female Health Worker was given responsibility for growth monitoring and general care of the 0-6 month child, and for deciding which pregnant women should receive supplementation. Although this part of the program has not been specifically evaluated since 1982, it appears that this did not work well, largely because of the numbers and distances involved for a worker covering a population of 5,000. It is now proposed to make the CNW responsible for both these groups, and also for birth weight recording,  - 13 - which was very incomplete under the project. Particular emphasis would be given to increasing the percentage of antenatal cases registered; to giving a broader range of antenatal care services; and to bringing all malnourished pregnant and lactating women into supplementation. 3.20 Second, changes in the ration size and the type of food supplement are being discussed, so that supplementation has a greater effect on nutrition. This should have consequences for all of the weight gain problems mentioned above, but particularly for -hildren who are non-responders because their families are too poor to feed them adequately at home. Bigger rations may be difficult for children to finish, so consideration is also being given to reducing the bulk or viscosity of the normal ration, and in addition to enriching the supplement being given to seriously malnourished children. 3.21 Third, it is proposed to improve the frequency and quality of medical attention that malnourished children receive. This should be of particular benefit to non-responders with health problems which make it difficult for them to absorb adequate nutrients. This issue is discussed in the following section on health. Fourth, children failing to thrive may be admitted into the feeding program on the basis of a check weighing a fortnight rather than a month after inadequate weight gain is first detected, thus preventing some further decline in nutritional status. (Children are not admitted into feeding immediately when growth faltering is detected, because brief growth interruptions due to infections are common and usually self-correcting; responding to such cases with three months of supplementation is unnecessary, and wasteful of resources). 3.22 Fifth, and most significantly in financial terms, it is proposed to supplement all children with moderate malnutrition instead of only those with faltering growth as at present. This is a recognition that the existing selection criteria did not make a big enough and fast enough impact on this type of malnutrition. The growth faltering entry criterion would however be retained for grade one and normal children, because of its apparently important role in preventing moderate and severe malnutrition. Such a decision would initially double the number of children in feeding, but it is hoped that the numbers of moderately malnourished children would subsequently come down to half the present levels. If that were the case, the food costs of the program in its maintenance phase would not substantially increase from the present, even taking into account the increased unit cost of the ration, and increased numbers of pregnant women in feeding. 3.23 The impact of this move to grade two feeding is hard to predict. It could be less than hoped to the degree that moderately malnourished children are stunted as well as low weight for age, since such cases are harder to rehabilitate (a study of the incidence of stunting is now being conducted). But even if stunting is now common, the measures proposed to increase birth weights and sustain normal growth in the first year of life should help to prevent it in the future. Moreover, initial results of a small scale experiment with universal feeding of grade twos in two blocks of the TINP area show a fairly rapid decrease in the incidence of moderate malnutrition, even without the introduction of complementary measures to improve maternal and 0-6 month infant nutrition. This  - 14 - experiment is now being expanded to additional blocks to confirm its findings. 3.24 In addition to reducing moderate malnutrition, the move to supplementing all grade two children should significantly reduce the relapse rate. Many of those currently relapsing are children graduating from grade three into grade two, who would now stay in feeding for a much longer period; a TINP- financed study of relapse cases indicated that while 45Z of those fed for 90 days relapse, less than 30% of those fed for 180 days do. Related to the question of the relapse rate is whether the proposed follow-on project should graduate children immediately they reach grade one or give them an additional period of 'protective' feeding to reduce the chances of subsequent relapse. 3.25 Finally, three kinds of responses are planned to the problem of exclusion. First, in the existing project areas it is planned to develop an additional 20% of CNCs to cover villages left out of the original project because of insufficient funds at the time of appraisal. Second, a mapping exercise will be undertaken to see whether a different distribution of hamlets between CNC villages will improve access. Third, experiments will be c'rried out with different approaches to service delivery in the hamlets--for eximple, employing an additional helper to deliver the food supplement to the hamlet, or creating a special Women's Working Group in the hamlet to help in education and service provision. The Health Component 3.26 TINP' s nutrition component has received more attention from observers and evaluators than its health component, despite the fact that the latter absorbed almost 50% of project costs, spent primarily on establishment of new health sub-centers, and on additional drug supplies. As noted in paras 2.19- 2.21, these investments did not have a satisfactorily return in terms of achieving targets levels of service coverage. The problem was that despite more clinics and more supplies being available, cooperation between health workers and CNWs in the field never really developed, and hence full advantage was not taken of the potential for increased health service outreach offered by the newly created network of CNCs. 3.27 Yet the software side of the health system was not neglected during project preparation. Work routines were designed for health workers to regularly visit the CNCs in their areas to provide support to CNWs. However, these visits did not in general take place as planned; one indicator of this was the DEAR evaluation's finding that only 28% of pregnant women knew the name of their health worker, as against 69% for CNWs. The failure of health workers and CNWs to work as a team in large measure explains low rates of ante-natal registration and lower than anticipated coverage of interventions such as FST, vitamin A and immunization. A weak health/nutrition coordination was partly the consequence of poor relationships between staff in the state's departments of health and social welfare in the early years of the project. But it is also a more general problem, not unique to TINP or to Tamil Nadu and remaining equally unresolved in the case of the ICDS program. The remainder of this section reviews seven lessons suggested by the TINP experience for improving coordination.  - 15 - 3.28 First, the state government proposed a structural reform of its departments of public health and social welfare to improve cooperation between health and nutrition staff. Under the new structure being discussed at the time of writing, all female staff in the field (CNWe and health workers, and their respective supervisors) would be responsible to one officer at the district level, reporting via the Community Health Nurse, who would act as project officer at the block level. While separate departments of Public Health, Family Welfare and Social Welfare would be retained at the state level, instructions to field staff would be issued under the joint names of the three directors. 3.29 Second, it was recognized that the number of pregnant and lactating women per health worker (about 125 in her population of 5,000) made it difficult for her to provide adequate service to all. It was therefore suggested that tasks might be reallocated to give the CNW a greater role in the following areas: * registration of antenatal cases * iron supplementation * maternal growth monitoring and supplementation * birth spacing motivation * identification of likely cases for birth spacing ? taking birth weights and referral of LBWA * growth monitoring 0-6 month child and weaning education * vitamin A supplementation * identification of children for immunization