Evaluation Research of the Nutrition 
Interventions in the Integrated 
Nutrition and Health Program (INHP) 
II Areas of CARE India 
Final Report 
January 28, 2008
2/20/2008 ii 
ACKNOWLEDGEMENTS
This project, entitled Evaluation Research of the Nutrition Interventions in the Integrated Nutrition 
and Health Program (INHP) II Areas of CARE India is a collaboration between CARE India, King 
Georges Medical University and the Johns Hopkins Bloomberg School of Public Health (JHSPH), USA. 
The project is funded by USAID/India. 
Collaborating institutions: 
CARE India 
Johns Hopkins University (JHU), Bloomberg School of Public Health 
King Georges Medical University (KGMU) 
Investigators: 
CARE: Usha Kiran, Deepika Nayar Chaudhery, Dharmendra Panwar, Sanjay Kumar 
JHU (Baltimore, MD): Michele L. Dreyfuss, Laura E. Caulfield, Abdullah H. Baqui, Gary Darmstadt, 
Saifuddin Ahmed, Robert E. Black, Mathuram Santosham 
JHU (Lucknow, UP): Veena Singh, Vishwajeet Kumar 
KGMU: Vinod Srivastava, Ramesh Ahuja 
Data collection and processing was contracted to TNS India. 
This report was prepared by Michele Dreyfuss, Lora Iannotti, Yong Li, Deepika Chaudhery, Veena 
Singh, Laura Caulfield, and Saifuddin Ahmed. 
2/20/2008 iii 
TABLE OF CONTENTS
ACKNOWLEDGEMENTS............................................................................................................ ii 
LIST OF ACRONYMS ...............................................................................................................xiii 
EXECUTIVE SUMMARY ............................................................................................................ 1 
CHAPTER 1: BACKGROUND..................................................................................................... 9 
Malnutrition Worldwide and in India ......................................................................................... 9 
Program Background .................................................................................................................. 9 
Nutrition Evaluation Research.................................................................................................. 14 
SECTION I: CROSS-SECTIONAL, PRE-TEST, POST-TEST EVALUATION IN ANDHRA 
PRADESH AND UTTAR PRADESH STATES ......................................................................... 16 
CHAPTER 2: METHODOLOGY AND DATA COLLECTION ................................................ 17 
Rationale for the Study Design................................................................................................. 17 
Study Design............................................................................................................................. 17 
Sampling Design....................................................................................................................... 17 
Sample Selection....................................................................................................................... 19 
Study Instruments ..................................................................................................................... 20 
Data Collection ......................................................................................................................... 20 
Data Management ..................................................................................................................... 21 
Sample Coverage ...................................................................................................................... 21 
Statistical Methods.................................................................................................................... 23 
CHAPTER 3: SOCIODEMOGRAPHIC CHARACTERISTICS ............................................... 24 
CHAPTER 4: ANTENATAL CARE SERVICES ...................................................................... 33 
Antenatal care ........................................................................................................................... 33 
Service provider contacts.......................................................................................................... 33 
Iron-folic acid supplementation ................................................................................................ 39 
Nutrition Health Days............................................................................................................... 41 
Summary................................................................................................................................... 44 
CHAPTER 5: MATERNAL NUTRITION ................................................................................. 47 
Supplementary Nutrition........................................................................................................... 47 
Maternal Diet ............................................................................................................................ 52 
Nutritional status....................................................................................................................... 59 
Pregnancy Anemia.................................................................................................................... 63 
Summary................................................................................................................................... 65 
CHAPTER 6: INFANT FEEDING .............................................................................................. 68 
Breastfeeding initiation............................................................................................................. 68 
Complementary Feeding........................................................................................................... 75 
Supplementary Nutrition........................................................................................................... 87 
Caregiver feeding practices....................................................................................................... 88 
Summary................................................................................................................................... 90 
CHAPTER 7: MICRONUTRIENT SUPPLEMENTATION AMONG CHILDREN ................. 92 
Vitamin A supplementation ...................................................................................................... 92 
Pediatric iron supplementation ................................................................................................. 93 
Summary................................................................................................................................... 97 
CHAPTER 8: CHILD NUTRITIONAL STATUS AND ANEMIA............................................ 99 
Child Nutritional Status ............................................................................................................ 99 
2/20/2008 iv 
Wasting ................................................................................................................................... 108 
Anemia.................................................................................................................................... 113 
Summary................................................................................................................................. 114 
SECTION II: COHORT STUDY IN UTTAR PRADESH STATE........................................... 116 
CHAPTER 9: METHODOLOGY AND DATA COLLECTION FOR THE COHORT STUDY
..................................................................................................................................................... 117 
Rationale for the Cohort Study Design................................................................................... 117 
Cohort Study Design............................................................................................................... 117 
Cohort Sampling Design......................................................................................................... 117 
Cohort Sample Selection......................................................................................................... 118 
Cohort Data Collection ........................................................................................................... 118 
Cohort Data Management....................................................................................................... 119 
Cohort Sample Coverage ........................................................................................................ 119 
Statistical Methods.................................................................................................................. 120 
CHAPTER 10: SOCIODEMOGRAPHIC CHARACTERISTICS IN THE COHORT STUDY
..................................................................................................................................................... 122 
CHAPTER 11: INFANT FEEDING IN THE COHORT STUDY............................................. 130 
Breastfeeding .......................................................................................................................... 130 
Complementary Feeding......................................................................................................... 132 
Supplementary Nutrition......................................................................................................... 138 
Caregiver feeding practices..................................................................................................... 138 
Summary................................................................................................................................. 140 
CHAPTER 12: CHILD NUTRITIONAL STATUS AND PHYSICAL GROWTH IN THE 
COHORT STUDY...................................................................................................................... 141 
Underweight............................................................................................................................ 141 
Stunting................................................................................................................................... 143 
Wasting ................................................................................................................................... 145 
Longitudinal modeling of physical growth............................................................................. 149 
Summary................................................................................................................................. 150 
CHAPTER 13: SUMMARY FINDINGS AND RECOMMENDATIONS FOR FURTHER 
ANALYSES................................................................................................................................ 152 
Summary findings................................................................................................................... 152 
Recommendations for Further Analyses................................................................................. 155 
REFERENCES ........................................................................................................................... 158 
APPENDIX I: CARE India’s INHP-II Basic Package of Nutrition Interventions..................... 159 
APPENDIX II: ADDITIONAL TABLES.................................................................................. 160 
A. Antenatal Care Services..................................................................................................... 160 
B. Rest during Pregnancy and the Postpartum Period ............................................................ 168 
C. Delivery, Newborn, and Postpartum Care ......................................................................... 168 
2/20/2008 v 
LIST OF TABLES 
Table 2.1. Target sample size calculations for primary nutritional outcomes in cross-sectional 
surveys at baseline and 24 months................................................................................................ 18 
Table 2.2. Eligible population and sample coverage at each survey round of the Nutrition 
Evaluation Research Study ........................................................................................................... 22 
Table 3.1. Age distribution among mothers of children 0-23 months by age .............................. 24 
Table 3.2. Age distribution among children 0-23 months by age................................................. 25 
Table 3.3. Sex distribution among children 0-23 months by age ................................................. 26 
Table 3.4. Religion among mothers of children 0-23 months by age........................................... 26 
Table 3.5. Caste among mothers of children 0-23 months by age................................................ 27 
Table 3.6. Education level among mothers of children 0-23 months by age................................ 27 
Table 3.7. Number of live births (parity) among mothers of children 0-23 months by age ......... 28 
Table 3.8. Inter-pregnancy birth intervals among mothers of children 0-23 months by age........ 28 
Table 3.9. Household toilet facilities among mothers of children 0-23 months by age ............... 29 
Table 3.10. Household drinking water source among mothers of children 0-23 months by age . 30 
Table 3.11. Child care arrangements among mothers of children 0-23 months by age ............... 31 
Table 4.1. Antenatal care (ANC) visits among mothers of children 0-23 months of age ........... 33 
Table 4.2. Service provider home visits during the most recent pregnancy among mothers of 
children 0-23 months in Andhra Pradesh...................................................................................... 34 
Table 4.3. Service provider home visits during the most recent pregnancy among mothers of 
children 0-23 months in Uttar Pradesh ......................................................................................... 35 
Table 4.4. Third trimester pregnant women reporting 1+ service provider home visits during 
pregnancy in Andhra Pradesh ....................................................................................................... 36 
Table 4.5. Pregnancy advice on diet, rest, and IFA supplementation received during home visits 
to currently pregnant women who received any pregnancy advice in Andhra Pradesh ............... 37 
Table 4.6. Third trimester pregnant women reporting 1+ service provider home visits during 
pregnancy in Uttar Pradesh........................................................................................................... 38 
Table 4.7. Pregnancy advice on diet, rest, and IFA supplements given during home visits to 
currently pregnant women who received any advice.................................................................... 39 
Table 4.8. Receipt of any IFA tablets during the most recent pregnancy among mothers of 
children 0-23 months in Andhra Pradesh...................................................................................... 40 
Table 4.9. Receipt and consumption of IFA supplements during the most recent pregnancy 
among mothers of children 0-23 months who received any supplements in Andhra Pradesh ..... 40 
Table 4.10. Receipt of any IFA supplements during the most recent pregnancy among mothers of 
children 0-23 months in Uttar Pradesh ......................................................................................... 41 
Table 4.11 Receipt and consumption of IFA supplements during the most recent pregnancy 
among mothers of children 0-23 months who received any supplements in Uttar Pradesh......... 41 
Table 4.12. Attendance at Nutrition Health Days (NHD) in the past 3 months among pregnant 
women and mothers of children 0-23 months in Andhra Pradesh................................................ 42 
Table 4.13. Attendance at Nutrition Health Days (NHD) in the past 3 months among pregnant 
women and mothers of children 0-23 months in Uttar Pradesh.................................................... 43 
Table 5.1. Receipt of Supplementary Nutrition during pregnancy among currently pregnant 
women and mothers of children 0-23 months during the last pregnancy in Andhra Pradesh ...... 48 
2/20/2008 vi 
Table 5.2. Receipt of Supplementary Nutrition during the postpartum period among mothers of 
children 0-5 months in Andhra Pradesh....................................................................................... 48 
Table 5.3. Receipt of Supplementary Nutrition during pregnancy among currently pregnant 
women and mothers of children 0- 23 months during the last pregnancy in Uttar Pradesh......... 50 
Table 5.4. Receipt of Supplementary Nutrition during the postpartum period among mothers of 
children 0-5 months in Uttar Pradesh ........................................................................................... 51 
Table 5.7. Meal consumption among pregnant women and mothers of children 0-23 months in 
Andhra Pradesh............................................................................................................................. 53 
Table 5.10. Snack consumption among pregnant women and mothers of children 0-23 months in 
Uttar Pradesh................................................................................................................................. 54 
Table 5.11. Dietary intake during the past 7 days among pregnant women in Andhra Pradesh .. 56 
Table 5.12. Dietary intake during the past 7 days among mothers of children 0-5 months in 
Andhra Pradesh............................................................................................................................. 56 
Table 5.13. Dietary intake during the past 7 days among mothers of children 6-23 months in 
Andhra Pradesh............................................................................................................................. 57 
Table 5.14. Dietary intake during the past 7 days among pregnant women in Uttar Pradesh...... 58 
Table 5.15. Dietary intake during the past 7 days among mothers of children 0-5 months in Uttar 
Pradesh.......................................................................................................................................... 58 
Table 5.16. Dietary intake during the past 7 days among mothers of children 6-23 months in 
Uttar Pradesh................................................................................................................................. 59 
Table 5.17. Body Mass Index (BMI) among pregnant women stratified by trimester in Andhra 
Pradesh.......................................................................................................................................... 60 
Table 5.18. Mean Body Mass Index among non-pregnant women in Andhra Pradesh ............... 60 
Table 5.19. Malnutrition (BMI<18.5 Kg/m2
) among pregnant women stratified by trimester in 
Andhra Pradesh............................................................................................................................. 61 
Table 5.20. Malnutrition (BMI<18.5 Kg/m2
) among non-pregnant women in Andhra Pradesh.. 61 
Table 5.21. Mean Body Mass Index among pregnant women stratified by trimester in Uttar 
Pradesh.......................................................................................................................................... 62 
Table 5.22. Mean Body Mass Index among non-pregnant women in Uttar Pradesh ................... 62 
Table 5.23. Malnutrition (BMI<18.5 Kg/m2
) among pregnant women stratified by trimester in 
Uttar Pradesh................................................................................................................................. 63 
Table 5.24. Malnutrition (BMI<18.5 Kg/m2
) among non-pregnant women in Uttar Pradesh ..... 63 
Table 5.25. Anemia among pregnant women stratified by trimester in Andhra Pradesh............. 64 
Table 5.26. Moderate-severe anemia among pregnant women stratified by trimester in Andhra 
Pradesh.......................................................................................................................................... 64 
Table 5.27. Anemia among pregnant women stratified by trimester in Uttar Pradesh................. 65 
Table 5.28. Moderate-severe anemia among pregnant women stratified by trimester in Uttar 
Pradesh.......................................................................................................................................... 65 
Table 6.1. Breastfeeding initiation among mothers of infants 0-5 months of age....................... 68 
Table 6.2. Prelacteal feeds of newborn infants among mothers of infants 0-5 months............... 70 
Table 6.3. Exclusive breastfeeding among infants 0-5 months of age in Andhra Pradesh........... 72 
Table 6.4. Early introduction of complementary foods and liquids among infants 0-5 months of 
age in Andhra Pradesh .................................................................................................................. 73 
Table 6.5. Exclusive breastfeeding among infants 0-5 months of age in Uttar Pradesh .............. 74 
Table 6.6. Early introduction of complementary foods and liquids among infants 0-5 months of 
age in Uttar Pradesh...................................................................................................................... 74 
2/20/2008 vii 
Table 6.7. Reported introduction of solids at 6-8 months among mothers of children 6-23 
months of age................................................................................................................................ 76 
Table 6.8. Foods consumption in the past 24 hours among breastfeeding children 6-23 months of 
age in Andhra Pradesh .................................................................................................................. 77 
Table 6.9. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding 
children 6-23 months of age in Andhra Pradesh........................................................................... 79 
Table 6.10. Foods consumption in the past 24 hours among breastfeeding children 6-8 months of 
age in Uttar Pradesh...................................................................................................................... 80 
Table 6.11. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding 
children 6-23 months of age in Uttar Pradesh .............................................................................. 82 
Table 6.12. Consumption of at least half the recommended quantity of solids among children 6-
23 months of age........................................................................................................................... 83 
Table 6.13. Minimum recommended frequency of solids consumed per day among breastfeeding 
children 6-23 months of age ......................................................................................................... 84 
Table 6.14. Consumption of ≥3 food groups in the past 24 hours among children 6-23 months of 
age................................................................................................................................................. 85 
Table 6.15. Consumption of oil added to solids among breastfeeding children 6-23 months...... 86 
Table 6.16. Receipt and consumption of Supplementary Nutrition by children 6-23 months of 
age in Andhra Pradesh .................................................................................................................. 87 
Table 6.17. Receipt and consumption of Supplementary Nutrition by children 6-23 months of 
age in Uttar Pradesh...................................................................................................................... 88 
Table 6.18. Child feeding by mothers among children 6-23 months of age................................. 89 
Table 6.19. Feeding from a separate plate among children 6-23 months..................................... 90 
Table 7.1. Number of IFA supplements received by children 12-23 months of age at endline ... 96 
Table 7.2. Number of days IFA supplements consumed by children 12-23 months of age at 
endline........................................................................................................................................... 96 
Table 7.3. Current consumption of IFA supplements among children 12-23 months of age at 
endline........................................................................................................................................... 97 
Table 8.1. Mean weight-for-age Z-score (WAZ) among children 0-23 months by sex and age in 
Andhra Pradesh........................................................................................................................... 100 
Table 8.2. Underweight (<-2SD WAZ) among children 0-23 months by sex and age in Andhra 
Pradesh........................................................................................................................................ 100 
Table 8.3: Severe underweight (<-3SD WAZ) among children 0-23 months by sex and age in 
Andhra Pradesh........................................................................................................................... 101 
Table 8.4. Mean weight-for-age Z-score (WAZ) among children 0-23 months by sex and age in 
Uttar Pradesh............................................................................................................................... 102 
Table 8.5. Underweight (<-2SD WAZ) among children 0-23 months by sex and age in Uttar 
Pradesh........................................................................................................................................ 103 
Table 8.6. Severe underweight (<-3SD WAZ) among children 0-23 months by sex and age in 
Uttar Pradesh............................................................................................................................... 103 
Table 8.7. Mean height-for-age Z-score (HAZ) among children 0-23 months by sex and age in 
Andhra Pradesh........................................................................................................................... 104 
Table 8.8. Stunting (<-2SD HAZ) among children 0-23 months by sex and age in Andhra 
Pradesh........................................................................................................................................ 105 
Table 8.9. Severe stunting (<-3SD HAZ) among children 0-23 months by sex and age in Andhra 
Pradesh........................................................................................................................................ 105 
2/20/2008 viii 
Table 8.10. Mean height-for-age Z-score (HAZ) among children 0-23 months by sex and age in 
Uttar Pradesh............................................................................................................................... 106 
Table 8.11. Stunting (<-2SD HAZ) among children 0-23 months by sex and age in Uttar Pradesh
..................................................................................................................................................... 107 
Table 8.12. Severe stunting (<-3SD HAZ) among children 0-23 months by sex and age in Uttar 
Pradesh........................................................................................................................................ 107 
Table 8.13. Mean weight-for-height Z-score (WHZ) among children 0-23 months by sex and age 
in Andhra Pradesh....................................................................................................................... 109 
Table 8.14. Wasting (<-2SD WHZ) among children 0-23 months by sex and age in Andhra 
Pradesh........................................................................................................................................ 109 
Table 8.15. Severe wasting (<-3SD WHZ) among children 0-23 months by sex and age in 
Andhra Pradesh........................................................................................................................... 110 
Table 8.16. Mean weight-for-height Z-score (WHZ) among children 0-23 months by sex and age 
in Uttar Pradesh........................................................................................................................... 111 
Table 8.17. Wasting (<-2SD WHZ) among children 0-23 months by sex and age in Uttar Pradesh
..................................................................................................................................................... 112 
Table 8.18. Severe wasting (<-3SD WHZ) among children 0-23 months by sex and age in Uttar 
Pradesh........................................................................................................................................ 112 
Table 8.19. Anemia among children 12-23 months of age in Andhra Pradesh.......................... 114 
Table 8.20. Anemia among children 12-23 months of age in Uttar Pradesh.............................. 114 
Table 10.1. Age distribution among pregnant women in the cohort and cross sectional studies, 
Uttar Pradesh............................................................................................................................... 122 
Table 10.2. Religion among pregnant women in the cohort and cross sectional studies, Uttar 
Pradesh........................................................................................................................................ 123 
Table 10.3. Caste among pregnant women in the cohort and cross sectional studies, Uttar 
Pradesh........................................................................................................................................ 123 
Table 10.4. Education level among pregnant women in the cohort and cross sectional studies, 
Uttar Pradesh............................................................................................................................... 124 
Table 10.5. Number of live births (parity) among pregnant women in the cohort and cross 
sectional studies, Uttar Pradesh .................................................................................................. 125 
Table 10.6. Inter-pregnancy birth intervals among pregnant mothers in the cohort and cross 
sectional studies, Uttar Pradesh .................................................................................................. 125 
Table 10.7. Household toilet facilities among pregnant women in the cohort and cross sectional 
studies, ........................................................................................................................................ 126 
Table 10.8. Household drinking water source among pregnant women in the cohort and cross 
sectional studies, Uttar Pradesh .................................................................................................. 127 
Table 10.9. Child care arrangements among pregnant women in the cohort and cross sectional 
studies, ........................................................................................................................................ 128 
Table 10.10. Socioeconomic Status (SES) Index among pregnant women in the cohort and.... 129 
Table 11. 1 Current breastfeeding status among children in the cohort sample 0-20 months of 
age, by district............................................................................................................................. 131 
Table 11.2. Consumption of at least half the recommended quantity of solids among children in 
the cohort sample at follow-up from 9 to 20 months of age, by district..................................... 136 
Table 11.3. Consumption of oil added to solids among breastfeeding children in the cohort 
sample at follow-up from 9 to 20 months of age, by district...................................................... 137 
2/20/2008 ix 
Table 11.4. Receipt of Supplementary Nutrition among children in the cohort sample at follow￾up from 12 to 20 months of age, by district................................................................................ 138 
Table 11.5. Plate sharing practices among children in the cohort sample at follow-up from 9 to 
20 months, by district.................................................................................................................. 139 
Table 12.1. Mean weight-for-age Z score (WAZ) among children in the cohort sample at follow￾up from 0-18 months, by district................................................................................................. 142 
Table 12.2. Underweight (% <-2SD WAZ) among children in the cohort sample at follow-up 
from 0-18 months, by district...................................................................................................... 143 
Table 12.3. Mean height-for-age Z score (HAZ) among children in the cohort sample at follow￾up from 0-18 months, by district................................................................................................. 144 
Table 12.4. Stunting (% <-2SD HAZ) among children in the cohort sample at follow-up from 0-
18 months, by district.................................................................................................................. 145 
Table 12.5. Mean weight-for-height Z score (WHZ) among children in the cohort sample at 
follow-up from 0-18 months, by district..................................................................................... 146 
Table 12.6. Wasting (% <-2SD WHZ) among children in the cohort sample at follow-up from 0-
18 months, by district.................................................................................................................. 149 
2/20/2008 x 
LIST OF FIGURES 
Figure 4.1. Home visits by AWW during the most recent pregnancy among mothers of children 
0-23 months in Andhra Pradesh.................................................................................................... 34 
Figure 4.2, Home visits by AWW during the most recent pregnancy among mothers of children 
0-23 months in Uttar Pradesh ....................................................................................................... 35 
Figure 4.3. Third trimester pregnant women reporting 1-2, 3+ AWW home visit during 
pregnancy in Andhra Pradesh ....................................................................................................... 36 
Figure 4.4. Third trimester pregnant women reporting 1-2, 3+ AWW home visit during 
pregnancy in Uttar Pradesh........................................................................................................... 38 
Figure 4.5. Attendance at Nutrition Health Days in the past 3 months at each survey round 
among pregnant women and mothers of children 0-23 months in Andhra Pradesh intervention 
district (Karimnagar)..................................................................................................................... 43 
Figure 4.6. Attendance at Nutrition Health Days in the past 3 months at each survey round 
among pregnant women and mothers of children 0-23 months in Uttar Pradesh intervention 
district (Barabanki) ....................................................................................................................... 44 
Figure 5.1. Receipt of Supplementary Nutrition during the last pregnancy and the postpartum 
period in Andhra Pradesh intervention district (Karimnagar) by survey round ........................... 49 
Figure 5.1. Receipt of Supplementary Nutrition during the last pregnancy and the postpartum 
period in Uttar Pradesh intervention district (Barabanki) by survey round.................................. 51 
Figure 6.1. Breastfeeding initiation within 1 hour of delivery among mothers of infants 0-5 
months in Andhra Pradesh by place of delivery ........................................................................... 69 
Figure 6.2. Breastfeeding initiation within 1 hour of delivery among mothers of infants 0-5 
months across all survey rounds ................................................................................................... 69 
Figure 6.3. Prelacteal feeds of newborn infants among mothers of infants 0-5 months in Andhra 
Pradesh by place of delivery......................................................................................................... 70 
Figure 6.4. Prelacteal feeds of newborn infants among mothers of infants 0-5 months across all 
survey rounds................................................................................................................................ 71 
Figure 6.5. Exclusive breastfeeding among children 0-11 months of age in Andhra Pradesh..... 72 
Figure 6.6. Exclusive breastfeeding among children 0-11 months of age in Uttar Pradesh......... 74 
Figure 6.7. Exclusive breastfeeding and early introduction of complementary foods and liquids 
among children 0-5 months of age in Uttar Pradesh across all survey rounds............................. 75 
Figure 6.8. Cereal consumption during the past 24 hours among children 6-23 months in Andhra 
Pradesh intervention district (Karimnagar.................................................................................... 78 
Figure 6.9. Legume consumption during the past week among children 6-23 months in Andhra 
Pradesh intervention district (Karimnagar)................................................................................... 79 
Figure 6.10. Cereal consumption during the past 24 hours among children 6-23 months in Uttar 
Pradesh intervention district (Barabanki) ..................................................................................... 81 
Figure 6.11. Legume consumption during the past week among children 6-23 months in Uttar 
Pradesh intervention district (Barabanki) ..................................................................................... 82 
Figure 7.1. Vitamin A supplementation among children 12-23 months of age............................ 92 
Figure 7.2. Vitamin A supplementation among children 12-23 months of age by survey round in 
Andhra Pradesh and Uttar Pradesh intervention districts ............................................................. 93 
Figure 7.3. Pediatric iron-folic acid (IFA) supplementation among children 12-23 months of age
....................................................................................................................................................... 93 
2/20/2008 xi 
Figure 7.4. Pediatric iron-folic acid (IFA) supplementation among children 12-23 months of age 
by survey round in Andhra Pradesh and Uttar Pradesh intervention districts .............................. 94 
Figure 7.5. Iron-folic acid supplementation by type of service provider among children 12-23 
months of age who received IFA supplements in Andhra Pradesh1
............................................. 95 
Figure 7.6. Iron-folic acid supplementation by type of service provider among children 12-23 
months of age who received IFA supplements in Uttar Pradesh1
................................................. 95 
Figure 8.1. Underweight (<-2 SD WAZ) among children 0-23 months by geographical block in 
Andhra Pradesh intervention district (Karimnagar).................................................................... 102 
Figure 8.2. Underweight (<-2 SD WAZ) among children 0-23 months by geographical block in 
Uttar Pradesh intervention district (Barabanki) .......................................................................... 104 
Figure 8.3. Stunting (<-2 SD HAZ) among children 0-23 months by geographical block in 
Andhra Pradesh intervention district (Karimnagar).................................................................... 106 
Figure 8.4. Stunting (<-2 SD HAZ) among children 0-23 months by geographical block in Uttar 
Pradesh intervention district (Barabanki) ................................................................................... 108 
Figure 8.5. Wasting (<-2 SD WHZ) among children 0-23 months by geographical block in 
Andhra Pradesh intervention district (Karimnagar).................................................................... 111 
Figure 8.6. Wasting (<-2 SD WHZ) among children 0-23 months by geographical block in Uttar 
Pradesh intervention district (Barabanki) ................................................................................... 113 
Figure 9.1. Cohort Study Profile................................................................................................ 120 
Figure 11.1. Breastfeeding initiation among mothers of children in the cohort sample, by district
..................................................................................................................................................... 130 
Figure 11.2. Cereal consumption in the past 24 hours among breastfeeding children in the cohort 
sample from 9 to 20 months, by district ..................................................................................... 133 
Figure 11.3. Legume consumption in the past 24 hours among breastfeeding children in the 
cohort sample from 9 to 20 months, by district .......................................................................... 133 
Figure 11.4. Fruit and vegetable consumption in the past 24 hours among breastfeeding children 
in the cohort sample from 9 to 20 months, by district ................................................................ 134 
Figure 11.5. Consumption of eggs, meat, chicken or fish in the previous 24 hours among 
breastfeeding children in the cohort sample from 9 to 20 months, district................................. 134 
Figure 11.6. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding 
children in the cohort sample 9 to 20 months of age, by district................................................ 135 
Figure 11.7. Minimum recommended frequency of solids consumed per day among 
breastfeeding children in the cohort sample at follow-up from 9 to 20 months of age, by district
..................................................................................................................................................... 136 
Figure 11.8. Consumption of ≥3 food groups in the past 24 hours among children in the cohort 
sample at follow-up from 9 to 20 months of age, by district...................................................... 137 
Figure 11.9. Child feeding by mothers among children in the cohort sample at follow-up from 9 
to 20 months of age, by district................................................................................................... 138 
Figure 11.10. Child feeding without assistance among children in the cohort sample at follow-up 
from 9 to 20 months of age, by district....................................................................................... 139 
Figure 12.1. Mean WAZ among children in the cohort sample at follow-up from 0-18 months, 
by district .................................................................................................................................... 142 
Figure 12.2. Mean height-for-age Z score (HAZ) among children in the cohort sample at follow￾up from 0-18 months, by district................................................................................................. 144 
Figure 12.3. Mean weight-for-height Z score (WHZ) among children in the cohort sample at 
follow-up from 0-18 months, by district..................................................................................... 146 
2/20/2008 xii 
Figure 12.4. Mean weight-for-height Z score (WHZ) among boys in the cohort sample at 
follow-up from 0-18 months, by district..................................................................................... 147 
Figure 12.5. Mean weight-for-height Z score (WHZ) among girls in the cohort sample at follow￾up from 0-18 months, by district................................................................................................. 148 
2/20/2008 xiii 
LIST OF ACRONYMS 
ANC Antenatal care 
ANM Auxiliary nurse midwife 
AWC Anganwadi center 
AWW Anganwadi worker 
ICDS Integrated Child Development Service 
IFA Iron-folic acid (supplements) 
INHP Integrated Nutrition and Health Project 
GOI Government of India 
HAZ Height-for-age Z score 
JHSPH Johns Hopkins School of Public Health 
KGMU King George Medical University 
MTR Mid-term Review 
NFHS National Family Health Survey of India 
NHD Nutrition and Health Day 
RACHNA Reproductive and Child Health, Nutrition and HIV/AIDS Program 
RDW Recently delivered women 
THR Take-home ration 
TT Tetanus toxoid (immunization) 
USAID United States Agency for International Development 
WAZ Weight-for-age Z score 
2/20/2008 1 
EXECUTIVE SUMMARY
Background 
The Reproductive and Child Health, Nutrition and HIV/AIDS Program (RACHNA) is an 
umbrella program of CARE India with the goal of improving child survival and nutritional 
status, women's reproductive health and prevention of transmission of HIV/AIDS. It is a 
collaboration between CARE India and the Government of India (GOI) funded by the United 
States Agency for International Development (USAID), and consists of 2 projects: the Integrated 
Nutrition and Health Project (INHP) and the Chayan project. The INHP-II, a 5-year program 
initiated in October 2001, targets pregnant and lactating women, and children less than 2 years of 
age to reduce child mortality and improve nutritional status. The program’s primary objective is 
to strengthen the Government of India’s (GOI) Integrated Child Development Services (ICDS) 
program and increase convergence with the Ministry of Health and Family Welfare’s (MOHFW) 
Reproductive and Child Health (RCH) Program for improved provision of health and nutrition 
services. Interventions include antenatal care, supplementary feeding, pediatric vitamin A and 
iron-folic acid supplementation, immunization, community-based newborn care, and promotion 
of appropriate breastfeeding and complementary feeding practices among children 0-2 years of 
age. INHP-II is currently implemented in ICDS catchment areas across 78 districts in nine states 
namely Andhra Pradesh, Bihar, Chhattisgarh, Jharkhand, Madhya Pradesh, Orissa, Rajasthan, 
Uttar Pradesh and West Bengal. 
The operational strategy of INHP-II is to demonstrate “best practices” in 10% of anganwadi 
center (AWC) areas while simultaneously replicating and scaling up these practices through 
government systems in other areas. The 4 “best practices” promoted in this strategy are 1) 
community-based volunteers (Change Agents) to promote healthy maternal and child care 
practices, 2) fixed-day, fixed site provision of health services at monthly Nutrition and Health 
Days (NHD), 3) community-based monitoring systems, and 4) Block-Level Resource Mapping. 
After the mid-term review (MTR) in late 2004, CARE reformulated its INHP-II intervention 
strategies to better increase coverage of program components and their impact on program 
outcomes. The program focused more specifically on improving supportive supervision of 
service providers and more strongly emphasizing home visits by ANM and AWW to encourage 
behavior change during pregnancy and the first 2 years of life. Locally adaptable tools were 
developed for field use to track progress and encourage action. The 4 “best practices” were de￾emphasized with the exception of NHD. 
Nutrition Evaluation Research
The Johns Hopkins University was funded by USAID in 2003 to evaluate over a 2-year period 
the impact of the basic package of INHP-II interventions compared to the ICDS program alone 
in reducing malnutrition and anemia among children 0-23 months of age. A secondary study 
objective was to assess whether the INHP-II strategy was more suited by design to reducing 
child malnutrition only in ecological settings with relatively low economic and cultural 
constraints. To accomplish these objectives, the evaluation was planned as a pre-test, post-test 
quasi-experimental study design in an INHP-II intervention district and an ICDS comparison 
district in each of 2 states. Andhra Pradesh was chosen to represent a relatively low constraint 
ecological setting in India, and Uttar Pradesh was chosen to represent a relatively high-constraint 
setting. In Andhra Pradesh, the district of Karimnagar was chosen as the experimental district 
2/20/2008 2 
while the district of Rangareddy was selected to serve as a comparison. In Uttar Pradesh, 
Barabanki was selected as the experimental district and Unnao as the comparison. 
Baseline data were collected from samples of pregnant women, mothers, and their children 0-23 
months of age in all 4 study districts in January-February 2004. The baseline survey interviewed 
and collected measurements from pregnant women, and mothers and their children 0-23 months 
of age to determine baseline levels of process indicators and outcomes for maternal and child 
nutrition prior to the intervention period. After 2 years of program intervention, another cross￾sectional survey of the same target groups in the study area was conducted in January-March 
2006 to assess the impact of the intervention. Two adequacy surveys were conducted during the 
2-year interval, in January 2005 and July-August 2005, to provide feedback to CARE on the 
pace of implementation of various program components. Data were collected from the pregnant 
women and mothers of children 0-23 months to assess program exposure at the household level. 
A second study component – a longitudinal cohort – was also conducted in the same two Uttar 
Pradesh districts to collect more detailed information on program exposures and nutrition 
outcomes. A cohort of pregnant women was enrolled starting in May 2004 and they and their 
infants were followed from delivery through 18 months of age to collect more detailed data on 
intervention exposures and dietary intakes, and repeated follow-up of children to assess 
nutritional outcomes, including child growth, over time. However, the cohort sample size was 
not designed to provide the primary means of assessing the program’s impact on child nutritional 
status as this was provided by the cross-sectional evaluation. 
Key Findings 
• Child malnutrition and anemia did not decrease as a result of the intervention in the 2 
districts included in the Nutrition ER cross-sectional, pre-test, post-test evaluation. In 
fact, a few specific measures of nutritional status worsened significantly among all 
children or in sex- or age-specific categories in the intervention relative to the 
comparison district in both states. However, small, but significant gains in weight and 
height were observed in the cohort study conducted in UP districts only. However, the 
prevalence of underweight, stunting, wasting and anemia among children under 2 years 
remain alarmingly high in both states. These results are incongruent with CARE’s own 
before and after surveys across 8 states that demonstrated a decrease of 8 percentage 
points in underweight among young children. However, this change was documented 
over a 5-year rather than 2-year period, and data from control districts were not available 
for comparison as they were in this evaluation.
• The INHP-II intervention was effective at improving some infant and young child 
feeding practices, particularly in UP state where baseline proportions were quite low for 
all feeding indicators. Timely initiation of breastfeeding increased and inappropriate early 
introduction of complementary foods decreased in both states. Indicators of diet quantity, 
quality, and frequency increased significantly in UP, but not AP, where they were more 
prevalent initially. However, endline prevalence of these indicators remained too low, 
particularly diet quantity, quality and frequency of feeds, and appear to have been 
insufficient to improve child nutritional status.
2/20/2008 3 
• Outreach by service providers to pregnant women for antenatal care and health and 
nutrition advice improved dramatically in the INHP-II intervention districts compared to 
the ICDS districts. Coverage and frequency of Anganwadi worker (AWW) home visits 
and contacts in the community during pregnancy increased significantly in both states, 
but contacts with ANM and CA improved in UP state only. 
• Coverage and utilization of health and nutrition services for pregnant women and their 
children increased in both states, but the effects of the intervention were greater for many 
services in UP. Improvements in both states were seen in distribution and consumption of 
SN among pregnant and postpartum women and children 6-23 months, attendance at 
NHD, and coverage of pediatric vitamin A and IFA supplementation. Coverage of ANC 
visits and antenatal IFA supplementation distribution and consumption were improved 
significantly in UP, but not AP. 
• The INHP-II intervention yielded tangible improvements in the dietary intake and 
nutritional status of both pregnant and postpartum women in both states. Although 
nutrition advice in INHP-II was targeted at women during pregnancy and the first 6 
months postpartum, broad improvements in dietary intake among all women were 
observed in both states. The gains in nutritional status among pregnant women and 
mothers of young children were an unexpected result. Unfortunately, they did not 
translate into gains in child nutritional status. 
Summary Results 
Antenatal care
• Receipt of ANC visits was universal in AP. INHP-II significantly increased ANC 
coverage in UP so that 50% of women reported at least one visit and 25% reported ≥3 
visits during the most recent pregnancy. 
• INHP-II was highly effective in both states at increasing outreach by service providers to 
provide ANC services and health and nutrition advice to pregnant women through home 
visits and contacts in the community. Approximately 50% of AP women and 60% of UP 
women reported at least one home visit by a service provider during the most recent 
pregnancy. Home visits by AWW increased most dramatically in both states. However, 
ANM and CA also increased their outreach significantly in UP. Service provider contacts 
at the Anganwadi Centers and Nutrition Health Days increased significantly also. 
• Improvements in coverage and consumption of antenatal IFA supplementation were 
limited in AP. However, UP state recorded significant increases in the proportion of 
women who received (65%) and consumed (36%) ≥90 tablets during their most recent 
pregnancy. In both states, the proportion of women who consumed all the IFA tablets 
they received more than doubled by endline. 
• Recent attendance at NHD increased significantly among pregnant women and mothers 
of children 0-23 months in both states. Coverage during the past 3 months was only 
2/20/2008 4 
~15% in AP, but 34-39% in UP at endline. However, the proportion of women in both 
states who remain unaware of NHD is still too high. 
Maternal nutrition
• Receipt of SN among pregnant and postpartum women increased significantly to ~50% in 
both states. At endline, consumption of SN was more common among intervention 
district than comparison district women also. In the recall of dietary intake during the 
past 7 days, SN intake was significantly increased among all groups of women in UP and 
among pregnant women only in AP. 
• Consumption of at least 3 meals a day increased but snack consumption decreased in AP 
state where the vast majority of women already consumed 3 meals a day. In UP state, 
only 25% of women ate 3 meals a day at baseline, but both meal and snack consumption 
increased at endline. In both states, these changes in meal and snack consumption 
occurred among pregnant and recently delivered women as well as mothers of children 6-
23 months of age. 
• There were multiple improvements in women’s diets attributable to INHP-II in both 
states. Recent consumption of legumes, dark green leafy vegetables, and yellow-orange 
fruits was increased among all groups of women in AP. In UP, only yellow-orange fruit 
intake increased significantly among all women, but pregnant women and mothers of 
children 6-23 months increased their weekly intake of meat, fish, chicken and eggs and 
recently delivered women increased their dairy intake. Dietary messages delivered to 
women during pregnancy and the early postpartum period appeared to have a positive 
effect on all women’s diets in many cases. 
• Malnutrition was very high among pregnant and non-pregnant women in both states, but 
it was particularly alarming in the AP intervention district where over 50% of mothers of 
children 6-23 months were malnourished. However, malnutrition prevalence decreased 
significantly among first trimester pregnant women and mothers of children 6-23 months 
of age in AP. In UP, malnutrition decreased among third trimester pregnant women only. 
The increase in BMI of ~1 unit among the first (AP) and third (UP) trimester pregnant 
women in the intervention districts over a 2-year evaluation period is an impressive 
improvement. 
• The INHP-II was not effective in reducing the prevalence of anemia among pregnant 
women in either state. Overall prevalence was comparable to the NFHS-2 prevalence 
although the severity of anemia was greater in this study population. 
Infant and child feeding
• The INHP-II program was highly effective at improving newborn feeding practices, 
including increasing early initiation of breastfeeding and reducing the provision of 
prelacteal feeds. The intervention effects were particularly dramatic in UP state. More 
modest changes in these feeding practices occurred in the UP cohort. 
2/20/2008 5 
• Exclusive breastfeeding among infants 0-5 months increased at endline in UP, but 
declined in AP. However, inappropriate early introduction of complementary foods and 
liquids before 6 months of age decreased significantly in both states. The decline in 
exclusive breastfeeding in AP was explained by an increase in breastfeeding with water, 
but the reasons for this are unknown. In the UP cohort, exclusive breastfeeding at 3-5 
months of age was more prevalent in the intervention than comparison district, but early 
introduction of complementary foods was not affected by the intervention. 
• The proportion of mothers of children 6-23 months of age who reported timely 
introduction of complementary foods at 6-8 months increased in both states, but only the 
increase in AP was attributable to the INHP-II intervention. Increases in cereal intake 
from the 24-hour recall among children 6-8 months were observed also in both states, 
but the increase in the intervention relative to the comparison district was significant in 
UP state only. By contrast, the intervention did not influence timely introduction of 
complementary foods among UP cohort mothers when defined by maternal recall or 
reported current dietary intake. 
• In UP, the INHP-II program was highly effective at increasing consumption of cereals 
and legumes among children 6-23 months. In AP, cereal intake increased equally in both 
districts and legume intake remained low. Fruit and vegetable intake was not influenced 
by INHP-II complementary feeding messages in either state, but intake increased in the 
UP districts and decreased in the AP districts. The intervention was highly effective also 
at increasing legume consumption from 9 to 18 months among children in the UP cohort. 
However, cereal intake increased equally in both districts and fruit and vegetable intake 
increased more rapidly among cohort children in the comparison than the intervention 
district. 
• INHP-II was effective at increasing complementary feeding indicators of diet quantity, 
quality and feeding frequency among children 6-23 months of age in UP, but not AP 
state. Diet quantity, quality, and frequency did increase in AP and remained more 
prevalent than in UP, but improvements were comparable in the two AP districts. 
However, the quality, frequency, and particularly quantity of dietary intake among 
children 6-23 months remained quite inadequate at the end of the evaluation. The 
intervention was less effective in the UP cohort, increasing the feeding frequency, but 
not diet quantity or quality among children from 9 to 18 months of age. 
• Receipt and consumption of SN among children 6-23 months was more prevalent in the 
intervention than comparison district in both states. However, recent declines in 
coverage and utilization of SN in late 2005 in the AP intervention district are of concern 
for the ongoing success of the program. The intervention was effective at distributing SN 
to families of weaning age children in the UP cohort as well. 
• Mothers in the intervention district of both states were more likely than those in the 
comparison district to report that they usually fed their child. In UP only, the 
intervention was also effective at encouraging mothers to feed their young children from 
a separate, rather than a shared, plate. In the UP cohort, some variation in the 
2/20/2008 6 
intervention effects was observed. While intervention district mothers were more likely 
to usually feed their child than comparison district mothers, they were less likely to feed 
them from a separate plate. The explanation for the inconsistency in the intervention 
effect in UP is unknown. 
Micronutrient supplementation
• Vitamin A supplementation coverage increased significantly in both states. Receipt of at 
least one dose of vitamin A syrup among children 12-23 months increased by ~50% to 
79% coverage in AP and four-fold to 69% coverage in UP. 
• Pediatric IFA supplementation coverage increased modestly to 30% in AP and 
dramatically to 69% in UP. Gains in coverage occurred during the second year of the 
intervention. However, few children reported consumption of more than a handful of 
tablets. 
Child nutritional status and anemia
• The INHP-II intervention did not have an effect on child nutritional status in the study 
populations in AP and UP states. 
• In AP state, underweight and stunting among all children did not change significantly. 
However, the mean HAZ score declined and severe wasting increased significantly in the 
intervention district. Stunting was increased significantly also among boys, but not girls. 
• In UP state, underweight, stunting, and wasting did not change significantly. However, 
the underweight prevalence increased and mean HAZ score declined significantly among 
girls, but not boys in the intervention district. 
• In the UP cohort, all 3 indicators of nutritional status followed the same worsening trend 
over the first 18 months of life that were observed in the cross-sectional data. However, 
there was limited evidence of a positive effect of the intervention on nutritional status and 
growth in the cohort sample. Mean WHZ was increased and wasting prevalence reduced 
among intervention children, particularly among girls from 3 to 12 months of age. The 
same benefits of the intervention were observed for mean WAZ and underweight 
prevalence at 6 and 9 months only. No effects on HAZ and stunting were observed. 
• Longitudinal modeling of weight and height gains among cohort children identified small 
positive effects of the intervention on physical growth from 0 to 18 months of age. The 
borderline significant weight gain in intervention district children was larger and 
statistically significant for girls only. The height gain among intervention district children 
was statistically significant also, but was larger for girls also. When calculated over a 12-
month period, children in the intervention district were predicted to have an additional 
200 g weight gain and 0.6 cm gain in height relative to those in the comparison district. 
The comparable gains for girls only were an additional 300 g weight gain and 0.8 cm 
gain in height. 
2/20/2008 7 
• The INHP-II program was not effective at reducing anemia among children 12-23 months 
of age in both states, and anemia prevalence remained >85% at endline. 
Recommendations for Further Analyses 
The INHP-II program was quite successful at improving maternal dietary practices and nutrition, 
infant and child feeding practices, and micronutrient supplementation in the intervention districts 
of both states, but particularly in UP. The pre-test, post-test, quasi-experimental design of this 
study allowed for evaluation of program impact on various indicators of these outcomes that are 
considered key to improved child nutritional status and health. However, study data from both 
the pre-test, post-test evaluation and the cohort also provide a valuable and unique resource to 
explore in detail the sociodemographic, environmental, health care utilization, and program 
factors that are associated with or predict key study outcomes. Listed below are some examples 
of additional analyses of the Nutrition Evaluation Research data that would be of high priority to 
better understand which factors and program inputs were most strongly associated with positive 
maternal and child nutrition outcomes. The following list is a starting point only. 
Infant and child feeding
• Sociodemographic, environmental, pregnancy, antenatal care utilization and program 
factors/predictors associated with early initiation of breastfeeding 
o Survey data: cross-sectional association of the factors with early breastfeeding 
initiation in the endline data in AP and UP. 
o Cohort data: Factors assessed at pregnancy enrollment and immediately 
postpartum that predicted early breastfeeding initiation in the 2 UP districts 
• Sociodemographic, environmental, pregnancy, health care utilization, delivery, infant 
feeding, and program factors/predictors associated with exclusive breastfeeding and early 
introduction of complementary foods and liquids. 
o Survey data: cross-sectional association of factors with exclusive breastfeeding 
and early introduction of complementary foods and liquids among children 0-5 
months of age in AP and UP 
o Cohort data: Factors assessed at pregnancy enrollment, and 1, 3, and 6 months 
postpartum that predicted a) exclusive breastfeeding until 6 months of age or b) 
early introduction of complementary foods and liquids in 2 UP districts. 
• Sociodemographic, environmental, health care utilization, infant, and program 
factors/predictors associated with introduction of solids to breastfeeding infants 6-11 
months of age 
• Sociodemographic, environmental, health care utilization, infant, and program 
factors/predictors associated with adequate quantity, frequency, and diversity of dietary 
intake among children 6-23 months of age 
• The association between diversity in the diets of mothers and their children (cross￾sectionally from survey data and longitudinally from cohort data). 
2/20/2008 8 
• Evaluation of variation in intervention impact on infant and child feeding practices by 
child sex and household socioeconomic status (survey data). 
Child nutritional status and anemia
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
factors/predictors of a) malnutrition, and b) good nutritional status of children at 6, 12, 
18, and 24 months of age 
o Survey data: cross-sectional association of factors with a) malnutrition, and b) 
good child nutritional status 
o Cohort data: Factors assessed at pregnancy enrollment, and at 1, 3, 6, 9, 12, 15, 
and 18 months of age that predicted a) malnutrition and b) good nutritional status 
as 6, 12, and 18 months of age. 
• Intervention effect on nutritional status and growth from 0-18 months of age in the 2 UP 
districts (cohort – in process) 
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
predictors of good physical growth between birth and 18 months of age in the 2 UP 
districts (cohort). 
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
factors associated with anemia among children 12-23 months of age 
• The association between receipt and consumption of pediatric IFA supplements and child 
anemia. 
Micronutrient supplementation
• Sociodemographic, environmental, health care utilization, maternal, infant and program 
factors/predictors of receipt and consumption of pediatric IFA supplements (survey and 
cohort data) 
Maternal Nutrition
• Sociodemographic, environmental, dietary, and program factors/predictors (advice given, 
receipt of SN, etc.) of women’s BMI and malnutrition 
o Survey data: cross-sectional association of factors with BMI and malnutrition 
o Cohort data: Factors assessed at pregnancy enrollment associated with third 
trimester BMI 
• Association of third trimester BMI and neonatal weight (cohort data) 
• Sociodemographic, environmental, dietary, health utilization, pregnancy, delivery, and 
program (contacts with service providers, attendance at NHD, advice given, etc.) factors 
associated with the receipt and consumption of SN 
2/20/2008 9 
CHAPTER 1: BACKGROUND
Malnutrition Worldwide and in India 
Malnutrition is responsible, directly or indirectly, for 60% of the 10.9 million deaths annually 
among children under five. Within this age range, the sharpest increase in malnutrition occurs 
between 6 to 24 months of age, the time when children grow most rapidly. This is due primarily 
to poor feeding practices that include not exclusively breastfeeding during the first six months of 
life, early cessation of breastfeeding, early introduction of complementary foods, and the feeding 
of inadequate quantity, and quality of complementary foods. These poor feeding practices 
compound the dangers for infants who are at highest risk of mortality because of their exposure 
to disease and limited access to health services. 
Moderately and severely malnourished children are five to eight times more likely to die than are 
adequately nourished children. Even mildly malnourished children have a greater risk of death 
than children who are normally nourished. The total number of child deaths attributable to mild 
and moderate malnutrition is far greater than those attributable to severe malnutrition. Among 
children who survive, malnutrition contributes to long-term negative effects on growth, health, 
cognitive development, and educational achievement. As adults, malnourished women are more 
likely to deliver intrauterine growth retarded infants, thereby creating an intergenerational cycle 
of malnutrition. 
India is the world’s largest contributor to malnutrition. According to the National Family Health 
Survey (NFHS-2) of 1998-99, 47% of children under the age of 3 years are underweight and 
46% are stunted; 74% of children between 6 months and 3 years of age suffer from anemia. 
Malnutrition is strikingly high in Madhya Pradesh (55%), Bihar (45 %), Orissa (45 %) and Uttar 
Pradesh (52 %) states. Malnutrition in the girl child continues into adulthood, with over a third of 
women 15-49 years chronically malnourished. Malnutrition is prevalent in both women and men, 
but it is more serious in women because of the intergenerational effect. The prevalence of anemia 
among women is ~50%, and approximately one-third of these women have moderate to severe 
anemia. 
Program Background 
The Reproductive and Child Health, Nutrition and HIV/AIDS Program (RACHNA) is an 
umbrella program of CARE India, supported by USAID and aimed at impacting on child and 
women’s health and nutrition, women's reproductive health and prevention of transmission of 
HIV/AIDS. It consists of the Integrated Nutrition and Health Project (INHP) and the Chayan
project. In October 1996, CARE, GOI and USAID initiated the Integrated Nutrition and Health 
Project (INHP), which transformed the program from primarily provision of supplementary food 
support to the ICDS into a maternal health and child survival initiative. INHP is a ten year 
project implemented in partnership with the Ministry of Human Resources Development's 
Department of Women and Child Development and Ministry of Health and Family Welfare's 
(MoH&FW) Department of Family Welfare, local NGOs and Community Based Organizations. 
While the first phase was successful in several key areas, some of the recommendations from 
final evaluation of the first phase of INHP included: the strengthening of INHP’s focus on 
2/20/2008 10 
nutritional impact, clearer definition of operational approaches, strengthening of capacity 
building and communication approaches to deliver the intervention package and a deliberate 
strategy for replication. The first five-year phase of the project was concluded in September 
2001 and the second five-year phase, referred to as INHP II, began in October 2001. INHP II is 
currently implemented in ICDS catchment areas across 78 districts in nine states namely Andhra 
Pradesh, Bihar1
, Chhattisgarh, Jharkhand, Madhya Pradesh, Orissa, Rajasthan, Uttar Pradesh and 
West Bengal. 
In July 2002, USAID funded a six-year reproductive health and HIV/AIDS prevention project, 
called “Chayan” (meaning Choices or Selection), to build upon the INHP program platform in 
29 districts from four out of nine INHP states2
. HIV prevention interventions among high-risk 
behavior groups and youth are implemented in 22 cities3
 in these four states. In the same year, a 
shift from project-based approach to program approach and organizational restructuring within 
CARE led to the creation of the RACHNA program. Thus, INHP and Chayan along with other 
health and nutrition projects were operationally and managerially integrated under the RACHNA 
program umbrella at all levels. The Chayan project, originally proposed to last until 2008, was 
synchronized with INHP II and both are scheduled now to end in late 2006. 
Overview of RACHNA
RACHNA’s goal is for vulnerable families to achieve sustainable improvement in the nutrition 
and health status of women and children in approximately 70 districts across 8 states in India by 
2006. 
INHP II intends to fulfill the following two sub-objectives:
• Service providers (especially auxiliary nurse midwives [ANM] and anganwadi 
workers[AWW]) improve the quality and coverage of maternal and child health services and 
key systems, including training, supply chain management and information management 
• Communities sustain activities for improved maternal and child survival. 
Impact and outcome commitments
INHP II is committed to specified improvements in a set of impact and outcome indicators by 
2006, at program scale: 
Impact: 
• A 10 percentage point increase in the proportion of normally nourished children of age 12-23 
months 
• A reduction in IMR (not to be directly measured) 
Outcome: 
• A set of indicators related to maternal and newborn care, nutrition and immunization, 
including reported practice of key behaviors and coverage of specified services, each having 
 
1
 Following bifurcation of Bihar, INHP was discontinued in the state in year 2000 since majority of program areas fell iinn the 
newly formed state of Jharkhand. Subsequently, in year 2003, the program was restarted in three districts of Bihar based a request 
from Government of India and state Government. 2
 The reproductive health interventions of Chayan project is implemented in rural areas of 29 districts from Uttar Pradesh, 
Jharkhand, Chhattisgarh, Rajasthan. These districts form a subset of 78 districts covered by INHP in nine states 3
 HIV prevention interventions are in 22 cities from the same four states, in addition to selected slums of National Capital 
Territory (NCT) of Delhi 
2/20/2008 11 
a specified increment to be achieved by the end-of-program in 2006. 
In addition, there are specified management indicators that are tracked annually by the program 
donor. 
Interventions package
The interventions chosen are those that have been proven to be associated with the program’s 
impact targets based on global evidence and the experiences from the first phase of INHP. The 
package includes: 
• Antenatal care: emphasis on check-ups, Tetanus Toxoid (TT) vaccination, iron-folic acid 
(IFA) supplementation, Supplementary Nutrition (SN), increased dietary intake, reduced 
workload, rest and preparations for safe delivery 
• Community-based newborn care: emphasis on clean childbirth, adequate warmth, early and 
exclusive breastfeeding, clean handling including cord care, recognition and extra care for 
the weak newborn (low-birth weight or premature) 
• Infant feeding practices: exclusive breastfeeding, appropriate complementary feeding 
(appropriate timing, quantity, quality and responsive feeding) with emphasis on feeding 
during and after illness, vitamin A supplementation, pediatric IFA supplementation and SN 
• Child immunization: complete and timely vaccination of all children with BCG, DPT, OPV 
and Measles by the completion of one year. 
Implementation approaches
The project was implemented by a hierarchy of district, state and national level teams. According 
to its operational strategy implemented since early 2002, innovation and demonstration in 
selected areas provided lessons on the ground that were simultaneously replicated elsewhere. 
Initially, the main lessons came from the previous phase of the project in the form of four “best 
practices”, and these were the mainstay of the program until the mid term review (MTR) of the 
program. The MTR provided an opportunity to thoroughly review lessons learnt and to plan for 
the rest of the life of the program. A brief overview of the operational strategy is provided here: 
Pre-MTR strategies / approaches: 
• Building on the food platform: Having supplied food commodity (corn-soya-blend and 
refined vegetable oil) to ICDS for decades using USAID Title II supplies, ICDS was the 
natural and nodal point for all operations. The AWC was the unit of implementation. The 
RCH program was a necessary partner for fulfilling several health-related objectives. All 
these were carry-overs from the previous phase. 
• Food supplementation: While the intention was to continue food supplementation throughout 
the program period, commodity supply was disrupted due to Genetically Modified 
Organisms (GMO) considerations by the GoI in mid-2002. The corn-soya blend supply 
ceased by about early 2003, while oil supplies continue to be programmed. 
• The Best Practices: Based on experience in INHP I, a number of approaches were assessed 
and the following four were chosen for wide application during INHP II: 
ƒ Change Agents (CA): community-based volunteers, each representing / catering to a 
cluster of 15-25 households, working with support of the AWW and community 
2/20/2008 12 
organizations, primarily to promote healthy maternal-child care practices. 
ƒ Nutrition-and-Health Day (NHD): The use of food distribution by ICDS as incentive 
to enhance effectiveness of fixed-day, fixed-site service provision by the RCH 
program, with value addition through involvement of community organizations 
ƒ Community-based monitoring systems (CBMS): The use of individual (Self￾Monitoring Tool) and collective (Social Map) tools to track caring practices and 
service utilization, leading to identification and addressing of problems by 
community groups. 
ƒ Block Level Resource Mapping (BLRM): A process of program review and action by 
concerned system functionaries at the block level, with participation of NGOs and 
people’s representatives 
• Technical interventions: using specified interventions that were known from global evidence 
to be effective in reducing mortality and malnutrition in the broad areas of antenatal care, 
essential newborn care, maternal infant and child feeding, and supplementary nutrition, and 
primary immunization (see Appendix I for details). 
• Behavior Change Communication (BCC): Using multiple channels to reach communities 
with relevant information support 
• System strengthening: to improve supply chains, information use and training institutions 
• Capacity building (CB) and advocacy: as the primary tools for achieving all of the above. All 
INHP staff, and virtually all systems and NGO functionaries in the program universe, 
including resource persons from government training bodies participated in structured CB 
efforts. 
• Gender and equity: increasing attention was paid through the program period to learning to 
assess the influence of gender and poverty / inequity on maternal and child health, and find 
ways to address them. 
• Demonstration Sites (DS), NGO partnerships and Replication: Partnering with local NGOs 
for the development of about 10 per cent of AWCs as DS, where program implementation 
could be demonstrated, that the ICDS and RCH programs could adapt elsewhere in a phased 
manner (Replication Sites, or RS), while ongoing learning in the DS would contribute to 
program refinements. 
• Early Learning Sites (ELS): Where fast-paced implementation over the first few months in a 
handful of sites would give field teams hands-on learning opportunities. 
• Rapid assessments in a panel of districts (RAPs): Using independently conducted annual 
household surveys to assess progression of outcomes and processes in one district in each 
state, and using this information for planning program refinements with government and 
other partners. 
• HMIS: A largely process-based that ensured process integrity across the program 
Using these approaches, interventions had reached about 40% of the program universe (up to 
10% DS and the rest RS) by December 2004. There was clear evidence of improvements in 
several outcome indicators as revealed by the RAPs, but this was not uniform, and the changes 
were not of sufficient magnitude to bring about impact level changes at scale. What was clear 
was that the processes as in use were not adequate to achieve this. Detailed independent and 
internal reviews, including the MTR, helped formulate approaches that had the best chance of 
success. 
2/20/2008 13 
Post-MTR strategies / approaches 
The primary change post-MTR was from an implicit process-orientation to an explicit results￾orientation of the program. The main elements of the changed approach are: 
• Closely engaged existing systems at multiple levels to get these interventions to scale: 
ƒ Primarily, used the supervisory level of ICDS functionaries to deliver interventions to 
full scale. 
ƒ Prioritized the mechanism of targeted home visits by AWW and ANM as primary 
vehicle for taking interventions to families. 
ƒ Utilized NGOs beyond conventional DS-associated roles to assist ICDS supervisors 
and CDPOs to implement / monitor interventions, and to try out any innovations at 
small scale before wider adaptation. 
ƒ Reorganized internally, assigning geographical regions within districts to individual 
district team (DT) members for ensuring implementation 
ƒ Used locally adaptable tools for AWW, ICDS supervisors, NGO staff and DTs to 
systematically track progress and take appropriate action. These tools emphasized use 
of information where generated, minimized upwards reporting and encouraged 
supportive supervision at all levels. 
ƒ Closely engaged district / block administration to initiate these mass-level processes 
and to monitor progress internally 
ƒ Used locally generated process / output data to inform program planning in each 
district, wherever possible (“mini-RAPs” were used in 53 districts, besides RAPs in 
the panel of 8 districts as previously) 
• BCC: was continued / universalized / optimized use of BCC materials 
• The “Best Practices”: 
ƒ Continued to look for ways to enhance effective community involvement in program 
monitoring, but not necessarily using the previous CBMS tools. 
ƒ Continued to use change agents where available and effective, but desisted from 
engaging in time-consuming structured recruitment and training of new CAs, mainly 
from the feasibility perspective. 
ƒ Continued promoting NHD, but with primary emphasis on fixed-day, fixed-site 
service provision 
ƒ The function of BLRM evolved into more frequent block level program reviews using 
existing forums within ICDS/health 
• Gender and equity: Used the best available experience to make the program sensitive to 
gender and equity issues while going to full scale 
• CB: Transformed structured CB to ongoing, on-the-job CB, using primarily ICDS sector 
meetings for strengthening program focus and content 
• Strengthening supply chains: Continued emphasis on attention to supply chains, as 
appropriate in each district 
• Demonstration and Replication: 
ƒ Specific focus and intensity on maintaining 10 per cent DS with NGO partners was 
considerably reduced 
ƒ With focus shifting to reaching every mother and child in every village, “replication” 
was no longer limited to selected sites. 
ƒ With the systems taking over the implementation in entirety, effectively, the 
distinctions between “DS”, “RS” and the rest became blurred by the end of program 
2/20/2008 14 
• HMIS: pruned down to include only donor-required data coming up to Delhi; the rest 
decentralized and given strong results-orientation, and largely integrated into government 
systems 
• Overall, there was considerably greater implementation flexibility and freedom to adapt these 
approaches to local contexts in individual states and districts 
In addition, common themes at state and national levels through the program life included 
strategies to build alliances for organizational learning and advocacy, with national and 
international agencies, and attempts to create synergy with other CARE and non-CARE projects. 
INHP II underwent through many challenges. One of them was the Genetic Engineering 
Approval Committee (GEAC) of Government of India's decision not to approve the import of 
food supplement i.e. Corn Soy Blend (one of the commodities used in INHP to support the ICDS 
Program), and its implications on the program. Non-import of CSB and availability of reduced 
resources levels from the donor implied significant change in the resource levels along with other 
implications of sudden stoppage of CSB. 
Nutrition Evaluation Research 
Objectives
This report presents findings from the program evaluation of CARE India’s INHP-II intervention 
impact on maternal and child nutritional outcomes. Findings presented in this report come from 
two study components: a pre-test, post-test cross-sectional evaluation in AP and UP states and a 
longitudinal cohort in UP state only. For the pre-test, post-test, cross-sectional evaluation, the 
baseline survey was conducted in January – February 2004, two adequacy surveys were carried 
out in January and July 2005 during program implementation, and the endline survey was 
conducted in January – March 2006. The objectives of this analysis are as follows: 
1. Compare the intervention districts (AP – Karimnagar, UP – Barabanki) to the comparison 
districts (AP – Rangareddy, UP – Unnao) in AP and UP states with regard to socio￾demographic characteristics. 
2. Describe the change in key process indicators (service provider contacts and advice, 
antenatal care, maternal and child micronutrient supplementation, supplemental feeding) 
from baseline to endline in the intervention and comparison districts in AP and UP states. 
3. Describe the changes in key maternal and child nutrition outcomes (underweight, 
stunting, anemia, infant and child feeding practices) from baseline to endline in the 
intervention compared to the comparison district in AP and UP states, respectively. 
4. Compare and contrast the impact of the intervention on key maternal and child nutrition 
indicators in AP and UP states. 
 
The longitudinal cohort was carried out in the same intervention and comparison districts in UP 
state that were selected for the cross-sectional evaluation surveys. A cohort of pregnant women 
was enrolled starting in May 2004 and they and their infants were followed from delivery 
through 18 months of age until June 2006. The objective of this study component was to collect 
more detailed data on intervention exposures and nutrition outcomes to further explore the 
intervention impact on key child nutrition indicators over time. 
2/20/2008 15 
The purpose of this program evaluation was to determine the impact of the basic package of 
INHP-II interventions in reducing malnutrition among children 0-23 months of age compared to 
the ICDS package alone. A secondary study objective was to assess whether the INHP-II 
strategy was more suited by design to reducing child malnutrition only in ecological settings with 
relatively low economic and cultural constraints. Child malnutrition and anemia are key impact 
indicators for INHP-II, but it is not possible to collect these outcomes data with a quasi￾experimental study design that includes comparison areas in all 8 states where INHP-II is being 
implemented. Therefore, the program impact using this rigorous study design was evaluated in 2 
districts each in AP and UP states only. To address the secondary objective, the evaluation 
research examined also the differential impact of INHP-II in a relatively low-constraint 
ecological setting in India, AP state, and a relatively high-constraint setting, UP state. 
Study sites 
A key focus of the study was to determine whether programs seeking to enhance child nutrition 
were likely to be more effective in areas where constraints to adequate child nutrition and growth 
are less formidable. Given this background, it was important to select areas with distinct 
ecological settings, including child nutritional status and resource endowments. The need to 
compare project impact in different settings was balanced with the importance of having the 
necessary service delivery infrastructure to maintain project fidelity. This balance was met by the 
selection of the two states of Andhra Pradesh and Uttar Pradesh. While the state of UP has high 
rates of child malnutrition and infant mortality, the state of AP has some of the better nutritional 
status figures in the whole of India. Andhra Pradesh also has a receptive state government, high 
potential uptake of evaluation results, and capacity to operationalize and monitor implementation 
of the evaluation activities. Uttar Pradesh also had basic implementation resources including its 
ongoing participation in the INHP Newborn Care Project that ensured an additional layer of site 
supervision. 
To select districts within each of the two states, data from the Reproductive and Child Health 
(RCH) rapid survey were used to compare outcomes targeted by INHP II. Only districts 
implementing the government’s ICDS program could be selected for comparison. In Andhra 
Pradesh, the district of Karimnagar was chosen as the experimental district while the district of 
Rangareddy was selected to serve as a comparison. In Uttar Pradesh, Barabanki was selected as 
the experimental district and Unnao as the comparison. . 
Collaborating partners
The INHP-II program is implemented by CARE India. Johns Hopkins Bloomberg School of 
Public Health (JHSPH) is conducting the evaluation of the nutrition impact of CARE India’s 
INHP-II program. JHSPH, in partnership with King Georges Medical University (KGMU), has 
designed the evaluation and is carrying out its implementation. Evaluation data were collected by 
TNS, a private survey research firm in India, under the supervision of JHSPH and KGMU. 
2/20/2008 16 
SECTION I: CROSS-SECTIONAL, PRE-TEST, POST￾TEST EVALUATION IN ANDHRA PRADESH AND 
UTTAR PRADESH STATES 
2/20/2008 17 
CHAPTER 2: METHODOLOGY AND DATA COLLECTION 
Rationale for the Study Design 
The purpose of this study was to evaluate the effectiveness of the INHP-II program to reduce the 
prevalence of malnutrition and anemia among children less than 2 years of age in 2 rural districts 
in India over a 2-year period. Health workers (ANM), community-based health and development 
workers (AWW), and change agents (community volunteers) were trained to provide counseling 
to women during pregnancy, after delivery, and during the first 2 years of their children’s lives to 
utilize antenatal care, delivery, and postpartum care services and to practice good maternal and 
child feeding behaviors to improve maternal and child nutritional status and health. To assess 
rigorously the impact of the program’s services, a pre-test, post-test, quasi-experimental 
evaluation design was employed with an intervention and comparison area in each state. The 
adequacy of the intervention implementation was assessed twice during the 2-year evaluation 
period; the first one after one year of program implementation, and the second 6 months later. 
Study Design
In each state, an INHP district was selected as an intervention area while a similar non-INHP 
(ICDS) district was chosen as comparison area for the nutrition evaluation research. In Uttar 
Pradesh, Barabanki was selected as the INHP district and Unnao was selected as the comparison 
district where the GOI’s ICDS program was in place. In Andhra Pradesh state, Karimnagar was 
selected as CARE’s INHP district and Rangareddy was selected as the ICDS comparison district. 
The baseline survey interviewed and collected measurements from pregnant women, and 
mothers and their children 0-23 months of age to determine baseline levels of process and 
outcome indicators for maternal and child nutrition prior to the intervention period. After 2 years 
of program intervention, another cross-sectional survey of the same target groups in the study 
area was conducted to assess the impact of the intervention. Two adequacy surveys were 
conducted during the 2-year interval to provide feedback to CARE on the pace of 
implementation of various program components. Data were collected from the pregnant women 
and mothers of children 0-23 months to assess program exposure at the household level. 
Sampling Design 
Sample size for baseline and endline surveys
Sample size was calculated to detect differences in various nutrition outcomes between each 
intervention district and its comparison district. Assumptions in making these calculations 
included a statistical power of 80% to detect a difference and a significance level of 95% (alpha 
= 0.05). The sample size was not increased to account for a design effect, which in nutrition 
studies has been small or trivial. The sample size estimations for children were based on three 
primary outcomes: shift in average z-scores in length and weight, stunting and anemia. 
Calculations for the primary outcomes for children and pregnant women are found in below. 
2/20/2008 18 
Table 2.1. Target sample size calculations for primary nutritional outcomes in cross￾sectional surveys at baseline and 24 months 
Outcome Measure SD or 
P0
Detectable difference in 
means by intervention arm 
Sample size 
per intervention arm* 
Nutritional status
 Children 
 Shift in average z-score 
 12-17 months 
 18-23 months 
 Stunting (%) 
 12-23 months 
1.1 
1.1 
 60% 
.18 z-score 
.18 z-score 
 Absolute reduction 6% 
587 ~600 
587 ~600 
1101 ~1200 
 Anemia
 Children 12-23 months 
 
 Pregnant women 
80% 
50% 
 Absolute reduction 6% 
 Absolute reduction 8% 
772 ~800 
633 ~800 
*Estimated sample size rounded to nearest ‘00. 
We needed approximately 1200 children aged 12-23 months for nutritional status outcomes and 
at least 800 children for anemia. The NFHS-2 data show that the prevalence of stunting (height￾for-age <-2 SD of median of reference population) among children 12-23 months was 57.5%. 
We estimated a minimum sample size requirement of 1101 children for the detection of 6% 
absolute reduction in stunting from a 60% prevalence of stunting at baseline. We rounded the 
target sample to 1200 children aged 12-23 months. The targeted 1200 sample size was also 
sufficient to detect a change of at least 0.18 z-score units for anthropometric z-scores. 
NFHS-2 reported that the prevalence of anemia among children aged 12-23 months nationally 
was 77.7% (children aged <3 years: 74.3% nationally, 73.9% in Uttar Pradesh, and 72.3% in 
Andhra Pradesh). We estimated that a sample of 772 children could detect a 6% reduction in 
anemia from the assumed 80% prevalence at baseline among children aged 12-23 months. 
However, recruitment for hemoglobin assessment included all children 12-23 months of age who 
were enrolled in the surveys for anthropometry and an interview, yielding a larger sample. For 
women of reproductive age, the prevalence of anemia was 51.8% nationally (48.7% in Uttar 
Pradesh and 49.8% in Andhra Pradesh). We estimated a minimum sample size requirement of 
633 (~800) pregnant women to detect a 8% reduction in anemia prevalence. 
Sample size for adequacy surveys
For the 2 adequacy surveys, we targeted 100 currently pregnant women and 500 mothers of 
children 0-23 months of age for interviews. The purpose of the adequacy surveys was to monitor 
the progress and implementation of the INHP-II intervention for various process indicators (e.g., 
home visits received, antenatal iron-folic acid supplements received), and statistical power was 
not a primary consideration in sample size estimation. However, we estimated the required 
sample size of at least 97 currently pregnant women to estimate the process indicators with 10% 
margin-of-error, and rounded the sample size to 100 pregnant women for the adequacy surveys. 
2/20/2008 19 
With a targeted sample of 500 mothers of children 0-23 months, we expected to detect at least a 
9% difference in program coverage between the arms. 
Based on population size, we needed 24 and 13 AWC areas in AP and UP districts, respectively, 
to meet the sample size requirements. 
Sample Selection 
In each of the 4 districts in the evaluation, rural blocks were selected for the evaluation surveys. 
In UP state, the intervention district of Barabanki has 16 blocks, and CARE worked in 14 of the 
16 blocks in the district, all of which were rural. The evaluation study was planned for the 10 
blocks where CARE was able to start their replication strategy immediately as the other 4 blocks 
had too many sector supervisor vacancies. The comparison district of Unnao has 16 blocks also, 
15 of which are rural. Nine blocks were randomly selected to represent this district in the 
evaluation. In AP state, each district selected for the evaluation has 15 blocks. In the intervention 
district of Karimnagar, CARE worked in 9 of the 15 blocks, and the evaluation was conducted in 
these 9 CARE blocks. In the comparison district of Rangareddy, 8 blocks were selected 
randomly for the purpose of this evaluation. 
Each block contains 3-6 sectors, an area with 15-25 AWC and an estimated population of 
20,000-25,000 persons. In the intervention districts in each state, the sampling universe for 
survey sample selection included only those sectors and AWC areas that were participating in 
CARE’s INHP-II program as replication sites. All sectors in the selected blocks had AWC areas 
participating in the intervention. In the comparison districts, the sampling universe included all 
functioning AWC from all sectors of the selected blocks in the district. 
Baseline survey
We used multi-stage survey sampling design to achieve the targeted sample of children and 
women. 
Uttar Pradesh. In the intervention district (Barabanki), 2 sectors were randomly sampled from 
each block (n=20 sectors). In the comparison district (Unnao), 2-4 sectors were randomly 
sampled per block in proportion to the number of AWC per block (n=20 sectors). In both 
districts, 3 AWC areas were randomly sampled from each selected sector (n=60 AWC). Women 
were randomly selected from the sampled AWC areas. 
Andhra Pradesh. In both districts, 2-3 sectors per block were sampled in proportion to the 
number of AWC per block (n=20 sectors), and 4-6 AWC areas were randomly sampled per 
sector (n=110 AWC). Women were randomly selected from the sampled AWC areas. 
Endline survey
The same sample selection process used at baseline was employed again at endline. However, 
the sampling universe in the intervention district in Uttar Pradesh was modified slightly by the 
removal of some AWC that did not exist at the time of the endline survey. 
2/20/2008 20 
Adequacy surveys
In all districts, one sector was randomly sampled from each block for selection of the Adequacy 
Survey sample. In AP districts, 2-3 AWC per sector were sampled in proportion to the number of 
AWC per block for a total of 24 selected AWC areas. In UP districts, 1-2 AWC per sector were 
sampled in the same manner for a total of 13 selected AWC areas. The same sample of AWC 
areas were used for both adequacy survey rounds to allow for tracking of program exposures in 
the same AWCs over time. 
Study Instruments 
The study included three separate instruments: 
i. Sociodemographic Questionnaire: This questionnaire was administered to pregnant 
women and mothers of children 0-23 months. It collected data on the respondents 
sociodemographic characteristics including pregnancy history, age, education, 
religion, caste, parity, birth intervals, occupation, work history, land ownership, 
source of income, and basic household amenities. 
ii. Currently Pregnant Women’s (CPW) questionnaire: This questionnaire was 
administered only to women who were pregnant at the time of survey. It collected 
information on a wide variety of issues including antenatal care utilization, health 
behaviors during pregnancy, micronutrient supplementation, dietary intake and 
practices, rest and workload, and nutritional status. 
iii. Mothers’ (MOM) questionnaire: This questionnaire was administered to women with 
children 0-23 months of age. It collected information on similar topics as those 
covered in the pregnant women’s questionnaire. In addition, it included questions on 
delivery, postpartum contacts with service providers, infant and child feeding 
practices and dietary intake, immunization, micronutrient supplementation, and 
maternal and child nutritional status. 
Data Collection
The evaluation involved 4 rounds of cross-sectional survey data collection: the baseline survey in 
January-February 2004, the Adequacy I survey in January 2005, the Adequacy II survey in July￾August 2005, and the endline survey in January-March 2006. At each survey round, we 
conducted interviews with currently pregnant women and mothers of children 0-23 months of 
age. Anthropometric measures of weight and height/length were obtained at the baseline, 
Adequacy II and endline surveys. Hemoglobin measurement for assessment of anemia was 
carried out among currently pregnant women and children 12-23 months of age at the baseline 
and endline surveys only. The surveys were implemented by TNS with guidance from Johns 
Hopkins and CARE India investigators. 
House listing
TNS conducted a household listing operation in all the sampled AWC areas selected for the 
samples at each survey round. The purpose of the household listing was to identify households 
with eligible women and children only. If more than one pregnant woman or mother of a child 0-
23 months of age was identified in a household, interviewers selected one eligible woman 
randomly from that household. The sample of AWC areas was selected with the expectation of 
2/20/2008 21 
identifying at least 2,400 mothers of children 0-23 months of age and 800 currently pregnant 
women in each district. For the adequacy surveys, the sample targets were 500 mothers of 
children 0-23 months of age and 100 currently pregnant women. However, all eligible women 
from sampled AWC areas were recruited for survey enrollment. This resulted in actual survey 
samples that may have fallen short of or exceeded the sample targets. 
Training and field work 
Training of interviewers and supervisors for each survey round was carried out by TNS with on￾site technical guidance from JHU and CARE personnel. The interview teams consisted of 4 
interviewers, 1 supervisor and 1 field editor each. The supervisors were trained to take the 
weight and height/length measurements from women and children, and to collect a fingerprick 
blood sample for hemoglobin assessment. The interviewers and field editors were trained to 
assist in both these tasks. The JHU field office in Lucknow organized, trained, and managed 
independent quality control teams to supervise the TNS interview teams and assure data quality. 
The surveys in Uttar Pradesh and Andhra Pradesh were conducted simultaneously at all survey 
rounds. 
Data Management
All questionnaires for baseline and endline surveys were returned to the TNS office from the 
field immediately after interview completion and data were entered by TNS. Quality of data was 
checked and validated both at TNS and Johns Hopkins University (JHU) for completion, range 
check, skip-pattern matching, heaping and and consistency. In case of discrepancies, the data 
were cross-checked with the questionnaire for immediate resolution. The final statistical 
analyses were conducted at JHU. 
Sample Coverage 
The eligible population identified for each survey round and coverage rates are reported below 
(Table 2.2). A total of 4878 mothers of children 0-23 months and 1205 pregnant women eligible 
for recruitment were found in the sampled households of the 2 AP districts at baseline, and 
coverage for the two groups was 83.7% and 78.2%, respectively. In UP districts, a total of 4895 
mothers of children 0-23 months and 1437 pregnant women eligible for recruitment were found 
in the sampled households at baseline, and coverage was 80.3% and 86.5% in the 2 groups, 
respectively. At endline, coverage in the AP districts was 80.5% among the 6892 mothers and 
69.4% among the 1902 pregnant women identified as eligible for the survey. In the UP districts, 
coverage was 82.8% among 5935 mothers and 83.5% among 1755 pregnant women. The main 
reason for non-response was failure to find the women at home despite 3 re-visits for all eligible 
survey respondents. 
2/20/2008 22 
Table 2.2. Eligible population and sample coverage at each survey round of the Nutrition Evaluation Research Study 
Andhra Pradesh Uttar Pradesh 
Intervention Comparison Intervention Comparison 
Population Coverage Population Coverage Population Coverage Population Coverage 
Baseline
 MOM 
 CPW 
2964 
763 
2375 
(80.1%) 
574 
(75.2%) 
2866 
777 
2503 
(87.3%) 
631 
(81.2%) 
3126 
853 
2443 
(78.2%) 
734 
(86.0%) 
2972 
809 
2452 
(82.5%) 
703 
(86.9%) 
Adequacy I
 MOM 
 CPW 
652 
165 
529 
(81.1%) 
120 
(72.7%) 
735 
183 
556 
(75.6%) 
140 
(76.5%) 
686 
171 
548 
(79.9%) 
133 
(77.8%) 
633 
186 
520 
(82.1%) 
136 
(73.1%) 
Adequacy II
 MOM 
 CPW 
606 
129 
498 
(82.2%) 
87 
(67.4%) 
877 
229 
576 
(65.7%) 
161 
(70.3%) 
672 
238 
542 
(80.7%) 
191 
(80.3%) 
650 
214 
531 
(81.7%) 
164 
(76.6%) 
Endline
 MOM 
 CPW 
3323 
809 
2668 
(80.3%) 
550 
(68.0%) 
3569 
1093 
2878 
(80.6%) 
770 
(70.4%) 
3186 
867 
2512 
(78.8%) 
761 
(87.8%) 
2749 
888 
2403 
(87.4%) 
704 
(79.3%) 
2/20/2008 23 
Statistical Methods
After careful evaluation of the quality of data, we examined the frequency distribution of the 
sample characteristics with descriptive statistics. We compared the similarity in the 
characteristics of sample distribution between intervention and comparison arms with Student’s 
t-test and Pearson’s chi-square for metric (continuous) and non-metric (categorical) variables, 
respectively. 
The main analysis of this study is based on “difference-in-difference” analysis (Wooldridge, 
2002) that compares the differences in health outcomes between baseline and endline survey 
periods by intervention and comparison arms. 
Statistical comparisons of intervention and comparison arms
The tables presented in the result section compare results from the baseline and endline surveys 
of this evaluation. For each indicator given in the tables, the change from baseline to endline is 
reported within each district (Δ = endline minus baseline value). A positive change indicates that 
the prevalence of an indicator has increased from baseline to endline and a negative change 
indicates it has decreased over time. The difference of differences (difference of Δ) is the test of 
the impact of the intervention. It is calculated as the change in the intervention district relative to 
the change in the comparison district (intervention Δ – comparison Δ). For an indicator where an 
increased proportion represents improvement (e.g. proportion of mothers consuming IFA 
supplements), a large positive difference of differences demonstrates program effectiveness. For 
an indicator where a decreased proportion represents improvement (e.g. proportion of children 
with weight-for-age Z-score <-2 SD), a large negative difference of differences demonstrates 
program effectiveness. We used probability linear model for statistical significance testing in 
“difference-in-difference” analysis (baseline to endline differences between intervention and 
comparison arms). One advantage of using regression model is to adjust the effect of 
intervention net of confounding covariates. The addition of variables for various 
sociodemographic characteristics to the regression models to test for confounding did not change 
significantly the evaluation results. Therefore, unadjusted results are presented in the subsequent 
chapters to facilitate interpretation of the “difference of differences” results. 
All analyses were conducted with STATA 8.0 (Stata Corp, TX). 
2/20/2008 24 
CHAPTER 3: SOCIODEMOGRAPHIC CHARACTERISTICS
The endline survey included two groups of respondents: mothers of children 0-23 months of age 
and currently pregnant women. Mothers in the former group who were also pregnant at the time 
of the interview contributed data to both questionnaires. Since most of the data presented in the 
following chapters come from the mothers of children 0-23 months, we present 
sociodemographic characteristics on this respondent group only in this chapter. The distributions 
of characteristics from the currently pregnant women closely matched the distributions presented 
here for mothers of children 0-23 months. Because of the large sample sizes, even very small 
differences in the distributions of two districts may be statistically significant. However, such 
statistical significance may not reflect a meaningful difference in the distribution of a 
characteristic that could affect the evaluation results in subsequent chapters. 
Maternal Age
In the AP districts, approximately three-quarters of the women were <25 years of age and only 
~5% were ≥30 years of age (Table 3.1). The only significant difference in the age distributions in 
the 2 districts was the smaller proportion of adolescent girls (<20 years) in Karimnagar district. 
The age distribution of women in UP state was skewed older than in AP. Only 9% of the samples 
were adolescent girls and approximately two-thirds were 20-29 years of age in both UP districts. 
In contrast to AP, more than 25% of the UP sample was ≥30 years of age. Although the age 
distributions of the 2 UP districts were statistically different, the differences in proportions for 
each age category were quite small. 
Table 3.1. Age distribution among mothers of children 0-23 months by age 
Age of mothers of children 0-23 months (%) 
Age categories Andhra Pradesh Uttar Pradesh 
(years) Karimnagar* Rangareddy* Barabanki * Unnao * 
< 20 16.2 22.1 8.7 9.5 
20-24 57.6 55.5 31.7 34.3 
25-29 20.8 18.0 31.6 30.4 
30-34 4.4 3.4 17.1 17.3 
35-49 1.1 0.9 10.9 8.6 
Total no. of 
mothers 
2575 2656 2413 2306 
* Statistical test comparing distributions between districts within a state, significant at p<0.05
2/20/2008 25 
Child Age
Overall, the samples were split approximately evenly between the first and second years of life in 
all districts, but the age distribution of children in AP skewed older than in UP. In the AP 
districts, there were 5% fewer children <12 months of age and 5% more children ≥18 months of 
age in Karimnagar than Rangareddy district. In the UP districts, ~30% of children were <6 
months of age compared to 22-3% in the AP districts. The oldest children 18-23 months of age 
comprised ~20% of the sample in both UP districts. 
Table 3.2. Age distribution among children 0-23 months by age 
Child Age (%) 
Age categories Andhra Pradesh Uttar Pradesh 
(months) Karimnagar* Rangareddy* Barabanki * Unnao * 
0-5 21.9 23.4 31.6 29.6 
6-11 25.9 28.4 19.4 22.6 
12-17 27.1 27.6 29.8 26.8 
18-23 25.2 20.5 19.2 21.0 
Total no. of 
children 0-23 
months 
2575 2656 2413 2306 
* Statistical test comparing distributions between districts within a state, significant at p<0.05 
Child Sex
The proportion of boys in the UP districts was slightly higher than in the AP districts (Table 3.3). 
In AP, there was no difference in sex distribution with almost equal percentage of male and 
female children in both districts. In UP, the distribution was skewed more to boys by a couple of 
percentage points in Unnao compared to Barabanki district, but this difference was not 
statistically significant. 
2/20/2008 26 
Table 3.3. Sex distribution among children 0-23 months by age 
Child Sex (%) 
Andhra Pradesh Uttar Pradesh 
Sex Karimnagar Rangareddy Barabanki Unnao 
Girl 49.2 50.8 48.7 46.5 
Boy 50.8 49.2 51.3 53.5 
Total no. of 
children 0-23 
months 
2575 2656 2413 2306 
Religion 
The majority of respondents in both states were Hindu. In AP districts, >90% in both districts 
were Hindi with Muslims at 4-6% and other religions – primarily Christian, Sikh, Buddist, and 
Jain – at 3%. However, the proportion of Hindus and Muslims in the 2 UP districts differed 
significantly. Barabanki District had a much larger Muslim population (19%) whereas it was 
<5% in Unnao District. 
Table 3.4. Religion among mothers of children 0-23 months by age 
Andhra Pradesh (%) Uttar Pradesh (%) 
Religion Karimnagar * Rangareddy * Barabanki * Unnao * 
Hindu 
93.2 90.7 80.9 95.5 
Muslim 4.1 6.0 19.1 4.4 
Others 2.7 3.3 0.1 0.1 
Total no. of 
mothers 
2575 2656 2413 2306 
* Statistical test between districts within a state, significant at p<0.05
Caste 
Other backward caste was the most common caste group in all four districts (Table 3.5). 
However, there was significant variation in the population distribution among caste groups 
between the intervention and comparison district in each state. In AP, Karimnagar had more 
other backward caste women than Rangareddy (60% vs. 43%) whereas Karimnagar had more 
scheduled tribe women. In UP, Barabanki had more other backward caste women than Unnao 
(52% vs. 41%), but scheduled caste women were more common in Unnao. Scheduled tribe 
women were virtually non-existent in both UP districts. 
2/20/2008 27 
Table 3.5. Caste among mothers of children 0-23 months by age 
Andhra Pradesh Uttar Pradesh 
Caste groups Karimnagar * Rangareddy * Barabanki * Unnao * 
Scheduled caste 
23.7 23.6 33.9 41.8 
Scheduled tribe 4.6 19.2 0.9 1.7 
Other backward 
caste 
61.3 42.9 51.6 41.4 
Not applicable 0.1 0.1 0.3 0.1 
Other 10.3 14.2 13.3 15.0 
Total no. of 
mothers 
2575 2656 2413 2306 
 * Statistical test between districts within a state, significant at p<0.05
Education 
Education level was reported as the number of school years completed by mothers using the 
categories from the NFHS II survey (Table 3.6). Illiteracy was more prevalent and high school 
education less prevalent in the UP than AP districts. In AP, approximately three-quarters of 
women had ≤7 years of education and ~15% had 10 or more years. However, there was some 
variation among the less educated women as illiteracy was more prevalent in Rangareddy than 
Karimnagar. In the UP districts, approximately two-thirds of women were illiterate and only 8% 
had 10 or more years of education. The distribution of women by education level differed 
statistically between UP districts, but the distributions were quite similar and unlike to bias any 
evaluation results. 
Table 3.6. Education level among mothers of children 0-23 months by age 
Education level Andhra Pradesh Uttar Pradesh 
(years) Karimnagar * Rangareddy * Barabanki * Unnao * 
None 
45.5 57.8 66.9 65.3 
1-7 
30.7 21.2 15.5 13.7 
8-9 
7.3 6.2 10.0 12.9 
10+ 
16.4 14.8 7.8 8.1 
Total no. of mothers 
2575 2656 2413 2306 
* Statistical test between districts within a state, significant at p<0.05
2/20/2008 28 
Parity 
Parity was much lower among women in AP than UP (Table 3.7). Over one-third of women in 
AP were primiparous compared to ~20% of women in UP. In the AP districts, the number of live 
births was skewed slightly higher in Rangareddy where 11% of women had ≥4 live births 
compared to 4% in Karimnagar. The distribution of live births were quite similar in the 2 UP 
districts, but were skewed slightly higher in Barabanki where 46% of women had ≥4 live births 
compared to 40% in Unnao. 
 Table 3.7. Number of live births (parity) among mothers of children 0-23 months by age 
 
Number of live Andhra Pradesh Uttar Pradesh 
births Karimnagar * Rangareddy * Barabanki* Unnao* 
1 38.5 34.2 18.1 20.6 
2-3 57.7 54.3 35.7 39.3 
4-5 3.5 10.3 26.6 23.2 
6+ 0.3 1.2 19.4 16.7 
Total no. of 
mothers 
2575 2656 2413 2306 
 * Statistical test between districts within a state, significant at p<0.05
Birth Interval 
Intervals between 2 births that are too short or too long not only pose difficulty for the mother, 
but also influence health outcomes for the newborn. In AP state, birth intervals were significantly 
longer in Karimnagar compared to Rangareddy district (Table 3.8). In Karimnagar, 45% of 
mothers had a birth interval of <48 months compared to 55% in Rangareddy. In UP, the birth 
interval distribution did not differ between districts. Just under 20% of women in both districts 
had birth intervals of <24 and ≥48 months each. Primiparous women were almost twice as 
common in AP as in UP districts, and did not contribute to the birth interval distribution. 
Table 3.8. Inter-pregnancy birth intervals among mothers of children 0-23 months by age 
Birth interval Andhra Pradesh Uttar Pradesh 
(months) Karimnagar * Rangareddy * Barabanki Unnao 
<24 months 
15.5 22.2 19.8 18.7 
24-47 months 
29.1 32.9 41.6 41.5 
48+ months 
15.6 10.2 19.8 18.0 
N/A 
39.8 34.7 18.9 21.8 
Total no. of 
women 
2575 2656 2413 2306 
* Statistical test between districts within a state, significant at p<0.05
2/20/2008 29 
Household Toliet Facilities 
Women in UP districts were much more likely than those in AP to have no toilet in their 
households and much less likely to have a flush toilet (Table 3.9). In AP, the type of toilet 
available in the household differed significantly between districts. Households in Rangareddy 
were twice as likely as those in Karimnagar to have a flush toilet. Three-quarters of Karimnagar 
households and a little more than half of Rangareddy households did not have a toilet in the 
home. In the UP districts, approximately 90% of households did not have a toilet. 
Table 3.9. Household toilet facilities among mothers of children 0-23 months by age 
Andhra Pradesh Uttar Pradesh 
Type of toilet facility Karimnagar * Rangareddy * Barabanki * Unnao * 
Flush toilet 
22.6 40.9 9.1 6.8 
Open/shared/public 
Toilet 
0.4 0.7 2.0 1.2 
No facility 
76.9 58.4 88.6 91.9 
Other 0.2 0.1 0.3 0.1 
Total no. of mothers 2575 2656 2413 2306 
* Statistical test between districts within a state, significant at p<0.05 
Drinking Water Source 
Approximately 80-90% of the mothers in all four districts reported piped water as their 
household’s drinking water source (Table 3.10). However, the piped water was more likely to 
come from a public tap than from within the household in AP compared to UP. In both states, the 
distribution of drinking water source differed significantly between districts. In AP, piped water 
within the home was more common in Rangareddy, but a little more than half of mothers in both 
districts obtained piped water from a public tap. In Karimnagar, well water was available within 
the home, but this was not a drinking water source in Rangareddy. 
In UP, piped water was available in over half of Barabanki households, but only 40% of Unnao 
households. More than a third of households in both districts had piped water from a public tap. 
A public well was the drinking water source for 5% of Barabanki households and 13% of Unnao 
households. 
2/20/2008 30 
Table 3.10. Household drinking water source among mothers of children 0-23 months by age 
 
Drinking water Andhra Pradesh Uttar Pradesh 
source type Karimnagar * Rangareddy * Barabanki * Unnao * 
Piped water in 
residence 
28.9 34.9 56.0 40.2 
Piped water in 
public tap 
51.5 56.7 36.4 43.4 
Well water in 
residence 
12.6 0.6 1.9 1.8 
Public Well 
5.0 1.5 5.0 13.1 
Natural water/other 
2.1 6.3 0.7 1.6 
Total no. of 
mothers 
2575 2656 2413 2306 
* Statistical test between districts within a state, significant at p<0.05
Child Care for Mothers who Work Outside the Home 
For mothers who worked outside the home, their child care arrangements were somewhat 
different in AP than UP. In AP, mother-in-laws were the most commonly used source of child 
care, but this option was more common in Karimnagar than Rangareddy. By contrast, women in 
Rangareddy were 3 times more likely than those in Karimnagar to take their child with them to 
work. Other relatives or neighbors were also a significant source of child care (8-10%). 
In UP, mothers in Barabanki were less likely than those in Unnao to work outside the home. For 
women who required child care, older children were the most common source in both districts at 
16%. Mothers-in-law or taking the child to work were additional options that were frequently 
used, but they were more common in Unnao than Barabanki. 
2/20/2008 31 
Table 3.11. Child care arrangements among mothers of children 0-23 months by age 
Child care Andhra Pradesh Uttar Pradesh 
arrangments Karimnagar * Rangareddy * Barabanki * Unnao * 
Take child to 
work 
3.6 11.6 7.3 12.2 
Husband 
1.8 0.5 0.7 0.3 
Mother in law 
25.4 18.0 8.9 13.0 
Older children 
1.6 4.5 15.7 16.1 
Other relatives/ 
neighbor 
10.4 7.7 3.7 3.6 
Do not work 
outside the 
home 
57.2 57.7 63.7 54.8 
Total no. of 
mothers 
2575 2656 2413 2306 
 * Statistical test between districts within a state, significant at p<0.05 
Socioeconomic Status Index 
In developing countries, income data are difficult to gather and are of poor quality. Several 
studies have suggested that a “wealth index” measured from household assets as a score through 
principal component analysis (PCA) is quite robust and serves as a good proxy for 
socioeconomic status (SES) (Filmer, and Pritchett, 2001; Bollen, Glanville, and Stecklov, 2001). 
Wagstaff and Watanabe (2002) showed that inequity in malnutrition, whether measured by a 
wealth index or consumption – a direct measurement of economic condition – provides similar 
degrees of inequity magnitudes. Currently, the wealth index is widely used for inequity studies in 
developing countries. We also used the wealth index to stratify respondents in SES terciles based 
on a principal component analysis of domestic assets and household possessions. The lowest 
tercile indicates the poorest group and the highest tercile indicates the richest group. The SES 
index distribution was determined separately for each state, and then the distribution in terciles 
was compared between the intervention and comparison districts within a state. 
The distribution of households across the SES index differed significantly between districts in 
both states. Approximately one-third of women were classified in the medium tercile of the 
index in all 4 districts. However, a larger proportion of households were classified in the low 
tercile and a smaller proportion in the high tercile in Karimnagar (intervention) district in AP 
state and in Unnao (comparison) district in UP state. 
2/20/2008 32 
Table 3.12. Socioeconomic (SES) Index among mothers of children 0-23 months of age 
Andhra Pradesh Uttar Pradesh 
SES Index level Karimnagar * Rangareddy * Barabanki * Unnao * 
Low 
37.6 29.5 27.7 39.3 
Medium 
35.2 31.4 33.5 33.2 
High 
27.3 39.1 38.8 27.5 
Total no. of 
mothers 
2575 2656 2413 2306 
 * Statistical test between districts within a state, significant at p<0.05 
2/20/2008 33 
CHAPTER 4: ANTENATAL CARE SERVICES 
Antenatal care
Antenatal care (ANC) is critical for the health and wellbeing of both the mother and the 
developing fetus. Frequent and regular use of quality antenatal care throughout pregnancy 
protects against adverse birth outcomes and maternal health complications. The GOI 
recommends that women have a minimum of 3 antenatal visits during pregnancy. 
Andhra Pradesh. Virtually all women in both AP districts reported receipt of some antenatal care 
during their most recent pregnancy (Table 4.1a). Even though ~85% of women made at least 3 
ANC visits at baseline, the intervention was successful at increasing significantly this proportion 
to 95% at endline. 
Table 4.1. Antenatal care (ANC) visits among mothers of children 0-23 months of age 
a. Andhra Pradesh 
Frequency of ANC visits among mothers of 0-23 month olds (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
ANC visits 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1+ visits 97.9 99.5 1.6 94.7 98.2 3.5 -1.9 
3+ visits 86.4 95.1 8.7 84.7 91.6 6.9 1.8 
Total # of 
mothers 
2369 2574 2476 2656 
b. Uttar Pradesh 
Frequency of ANC visits among mothers of 0-23 month olds (%) 
Intervention (Barabanki) Comparison (Unnao) 
ANC visits 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1+ visits 35.1 53.4 18.3 29.7 28.0 -1.7 20.0* 
3+ visits 11.5 25.3 13.8 8.9 8.8 -0.1 13.9* 
Total # of 
mothers 
2445 2413 2388 2306 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. Utilization of ANC was dramatically lower in UP than AP districts. Nonetheless, 
the INHP-II intervention was effective at improving ANC coverage and intensity of exposure. 
The proportion of women in the intervention district who received 1+ or 3+ ANC visits increased 
significantly to 53% and 25%, respectively (Table 4.1b). The proportions in the comparison 
district were unchanged. 
Service provider contacts
Mothers of children 0-23 months during their most recent pregnancy
Andhra Pradesh. Service provider home visits during the most recent pregnancy were reported 
by only one-third of women at baseline in the intervention district, but this increased to 
approximately half by endline (Table 4.2). This improvement in coverage of home visits during 
pregnancy was attributable to a significant increase in the visits by AWW in the intervention 
relative to the comparison district. Data in Figure 4.1 illustrate significant increases in the 
2/20/2008 34 
intensity of contact with AWW as well. The proportion of women reporting visits by ANM did 
not change over time, and there was little evidence of CA participation in this intervention 
activity. 
Table 4.2. Service provider home visits during pregnancy among mothers of children 0-23 months in Andhra 
Pradesh 
1+ service provider home visit during the most recent pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 36.1 49.4 13.3 22.1 28.5 6.4 6.9* 
ANM 31.5 35.6 4.1 20.0 22.1 2.1 2.0 
AWW 15.7 41.3 25.6 8.4 15.2 6.8 18.8* 
CA 0.3 2.4 2.1 0.1 0.1 0 2.1* 
Total # of 
mothers 
2369 2574 2476 2656 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Figure 4.1. Anganwadi worker (AWW) home visits during pregnancy among mothers of children 0-23 
months in Andhra Pradesh
6 15
4 7
10
26
4 8
0
20
40
60
80
100
Baseline Endline Baseline Endline
Service provider home visits (%)
3+ home visits
1-2 home visits
AP Intervention AP comparison
Women were asked to report on pregnancy advice received outside the home other than at ANC 
visits. Women in the intervention district were significantly more likely to report receiving 
advice at the Anganwadi Center (AWC) (49 vs. 37%) or at a Nutrition Health Day (NHD) (12 
vs. 7%) than women in the comparison district during their most recent pregnancy. 
Uttar Pradesh. The proportion of women who reported any service provider home visits during 
the most recent pregnancy was ≤20% at baseline in both UP districts (Table 4.3). However, this 
proportion increased significantly to 60% at endline in the intervention district relative to little 
change in the comparison district. This dramatic increase was a result of improved outreach to 
pregnant women by ANM and AWW. Change Agents played a role as well. By endline, 40% of 
women reported at least one home visit by ANM and 44% by AWW in the intervention district. 
2/20/2008 35 
Significantly increased intensity of contacts was reported for ANM (data not shown) and AWW 
(Figure 4.2) in the intervention relative to the comparison district. 
Table 4.3. Service provider home visits during pregnancy among mothers of children 0-23 months in Uttar 
Pradesh 
1+ service provider home visit during the most recent pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 20.2 59.7 39.5 17.8 22.4 4.6 34.9* 
ANM 15.0 39.8 24.8 12.0 17.0 5 19.8* 
AWW 6.8 44.0 37.2 3.9 7.3 3.4 33.8* 
CA 0.8 19.4 18.6 0.1 0.1 0 18.6* 
Total # of 
mothers 
2445 2413 2386 2306 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Figure 4.2. Anganwadi worker (AWW) home visits during pregnancy among mothers of children 0-23 
months in Uttar Pradesh 
4
17
5 3
27
2 3
0
20
40
60
80
100
Baseline Endline Baseline Endline
Service provider home visits (%)
3+ home visits
1-2 home visits
UP intervention UP comparison
 
Intervention district women in UP were more than twice as likely as comparison district women 
(47 vs. 21%) to report receipt of pregnancy advice outside the home other than at ANC visits. 
Pregnancy advice was significantly more common for intervention than comparison district 
women at the Anganwadi Center (AWC) (31 vs. 7%) and at a Nutrition Health Day (NHD) (26 
vs. 2%). 
Currently pregnant women
Andhra Pradesh. Third trimester pregnant women in the intervention district were twice as likely 
to be visited at home by a service provider at endline (70%) compared to baseline (36.8%; Table 
4.4). In contrast to the findings among mothers of children 0-23 months, this improved coverage 
was due to increased contacts with both ANM and AWW. Home visits by CA increased 
significantly, but remained rare at endline. Nevertheless, the greatest improvements were 
observed with AWW home visits that increased more than 3-fold in the intervention district 
2/20/2008 36 
relative to no change in the comparison district. Increased intensity of visits was observed for 
AWW only (Figure 4.3). 
Table 4.4. Service provider home visits during pregnancy among third trimester pregnant women in Andhra 
Pradesh 
Third trimester pregnant women reporting 1+ service provider home visits during 
pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 36.8 70.0 33.2 22.9 24.6 1.7 31.5* 
ANM 29.0 43.0 14.0 16.9 15.6 -1.3 15.3* 
AWW 21.3 69.0 47.7 13.3 13.6 0.3 47.4* 
CA 2.6 8.0 5.4 1.2 0 -1.2 6.6* 
Total # of 3rd
trimester 
pregnant 
women 
155 100 166 199 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Figure 4.3. Frequency of AWW home visits during pregnancy among third trimester pregnant women in 
Andhra Pradesh 
7
36
6 11
14
33
7 3
0
20
40
60
80
100
Baseline Endline Baseline Endline
Service provider home visits (%)
3+ home visits
1-2 home visits
AP intervention AP comparison
Approximately two-thirds of third trimester pregnant women in both AP districts reported receipt 
of pregnancy advice from places outside the home (intervention: 67%, comparison: 60%). 
However, women in the intervention district were significantly more likely than those in the 
comparison district to have received pregnancy advice at an AWC (61 vs. 42%) or NHD (16 vs. 
8%). 
At baseline, advice on maternal diet was received by >50% of women who received any 
pregnancy advice from a service provider during home visits (Table 4.5). Specific advice on 
consumption of IFA supplements was also received by ~40% of women whereas advice on extra 
rest during pregnancy was less common. The intervention was effective only at increasing the 
2/20/2008 37 
proportion of women who received advice on extra rest. The occurrence of advice on maternal 
diet, IFA supplementation, and breastfeeding did not change significantly. 
Table 4.5. Pregnancy advice on diet, rest, and IFA supplementation received during home visits to currently 
pregnant women who received any pregnancy advice in Andhra Pradesh
Advice received (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Pregnancy advice 
by service 
provider type Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Auxilary Nurse 
Midwife (ANM) 
 
Diet 50.4 52.1 1.7 45 49.3 4.3 -2.6 
Rest 23.5 52.1 28.6 28.3 30.7 2.4 26.2* 
IFA supplements 44.5 58.2 13.7 43.3 40 -3.3 17.0 
Breastfeeding 2.5 2.1 -0.4 3.3 1.3 -2.0 1.6 
Total # pregnant 
women who 
received advice 
from ANM 
119 146 60 75 
Anganwadi 
worker (AWW) 
 
Diet 61.3 56.5 -4.8 40.9 42.2 1.3 -6.1 
Rest 30.1 55.8 25.7 27.3 31.3 4.0 21.7 
IFA supplements 46.2 41.2 -5.0 34.1 23.4 -10.7 5.7 
Breastfeeding 4.3 3.5 -0.8 3.5 1.6 -1.9 0.8 
Total # pregnant 
women who 
received advice 
from AWW 
93 260 44 64 
Change Agent 
 
Diet 55.6 60.0 4.4 0 n/a 
Rest 44.4 65.0 20.6 0 n/a 
IFA supplements 33.3 50.0 16.7 40 n/a 
Breastfeeding 0 5.0 5.0 0 n/a 
Total # of 
pregnant women 
who received 
advice from CA 
9 20 5 n/a 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. Service provider home visits for pregnancy advice reported by third trimester 
pregnant women were more common than those reported by mothers of children 0-23 months, 
but followed the same increasing trend in the intervention district. Over 80% of third trimester 
pregnant women at endline reported a service provider home visit during pregnancy compared to 
only 21% at baseline - a highly significant increase relative to the modest improvement observed 
in the comparison district over time (Table 4.6). This dramatic intervention effect was a result of 
dramatic increases in home visits by ANM, AWW, and CA so that almost two-thirds of women 
reported a home visit by ANM and by AWW at endline. Change Agent visits were most 
2/20/2008 38 
commonly reported by this target group compared to other program indicators. Increased 
coverage was accompanied by significantly increased intensity of exposure to this program 
component for all three service provider types. Figure 4.4 illustrates the significantly increased 
proportion of intervention district pregnant women receiving 1-2 and 3+ AWW home visits. The 
same trend was observed for ANM and to a lesser extent, CA. 
Table 4.6. Third trimester pregnant women reporting 1+ service provider home visits during pregnancy in 
Uttar Pradesh 
Third trimester pregnant women reporting 1+ service provider home visits during 
pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 20.7 82.9 62.2 17.8 31.0 13.2 49.0* 
ANM 15.2 58.6 43.4 13.6 23.4 9.8 33.6* 
AWW 8.0 67.0 59 1.9 9.8 7.9 51.1* 
CA 0.8 32.2 31.4 0 1.1 1.1 30.3* 
Total # of 3rd
trimester CPW 
237 239 214 184 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Figure 4.4. Third trimester pregnant women reporting 1-2, 3+ AWW home visit during pregnancy in Uttar 
Pradesh 
5
30
1 5 3
37
1
4
0
20
40
60
80
100
Baseline Endline Baseline Endline
Service provider home visits (%)
3+ home visits
1-2 home visits
UP Intervention UP comparison
Pregnant women in the intervention district were also much more likely than those in the 
comparison district to receive pregnancy advice from places outside the home (63 vs. 13%). This 
difference was due to increased contact with service providers at the AWC (46 vs. 11%) and 
NHD (38 vs. 3%). 
Information on IFA supplementation was the most common type of pregnancy advice given by 
ANM at baseline (Table 4.7). Advice on maternal diet and rest were much less frequently 
mentioned. However, the proportion of pregnant women who received advice from ANM on diet 
and rest was significantly increased at endline in the intervention relative to the comparison 
district. Advice from AWW on diet and rest was also more likely at endline, but the increase did 
2/20/2008 39 
not differ significantly between the intervention and comparison districts. Advice on IFA 
supplementation by both ANM and AWW also increased equally in the two districts. 
Table 4.7. Pregnancy advice on diet, rest, and IFA supplements given during home visits to currently 
pregnant women who received any advice 
Advice received (%) 
Intervention (Barabanki) Comparison (Unnao) 
Pregnancy advice 
by service 
provider type Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Auxilary Nurse 
Midwife (ANM) 
 
Diet 10.7 54.5 43.8 17.3 16.2 -1.1 44.9* 
Rest 10.7 44.2 33.5 5.8 17.2 11.4 22.1* 
IFA supplements 34.5 55.3 20.8 34.6 56.6 22.0 -1.2 
Breastfeeding 0 1.9 1.9 1.9 1.0 -0.9 2.8 
Total # of 
pregnant women 
who received 
advice from ANM 
84 371 52 99 
Anganwadi 
worker (AWW) 
 
Diet 33.3 59.7 26.4 22.2 30.4 8.2 18.2 
Rest 30.6 49.2 18.6 11.1 17.4 6.3 12.3 
IFA supplements 36.1 52.8 16.7 33.3 56.5 23.2 -6.5 
Breastfeeding 2.8 1.8 -1.0 0 2.2 -2.2 1.2 
Total # of 
pregnant women 
who received 
advice from 
AWW 
36 449 9 46 
Change Agent 
 
Diet 62.5 63.8 1.3 n/a n/a 
Rest 37.5 53.7 16.2 n/a n/a 
IFA supplements 62.5 40.8 -21.7 n/a n/a 
Breastfeeding 0 0.9 0.9 
Total # of 
pregnant women 
who received 
advice from CA 
8 218 n/a n/a 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Iron-folic acid supplementation
The GOI policy is that all women should receive and consume at least 3 months’ supply of IFA 
supplements during pregnancy. ANM work together with AWW to distribute these supplements 
and encourage their consumption by pregnant women as part of antenatal care visits and during 
other service provider contacts at health centers, AWC, NHD, and at womens’ homes. 
2/20/2008 40 
Andhra Pradesh. At baseline, receipt of any IFA supplements during the most recent pregnancy 
was significantly lower in the intervention than the comparison district (Table 4.8). However, the 
~30% increase in this proportion in the intervention district at endline was significantly greater 
than the small increase in the comparison district. Nonetheless, the 54% coverage at endline in 
the intervention district still falls short of desired coverage levels. 
Table 4.8. Receipt of any IFA tablets during the most recent pregnancy among mothers of children 0-23 
months in Andhra Pradesh 
Receipt of any IFA supplements (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Receipt of any 
IFA 
supplements 
40.9 54.5 13.6 74.4 75.8 1.4 12.2* 
Total # of 
mothers 
2369 2575 2476 2656 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Among women who received any IFA supplements during their most recent pregnancy, only 
one-quarter at baseline received the recommended quantity of 90+ supplements and this 
proportion remained unchanged at endline in both districts (Table 4.9). Consumption of at least 
90 supplements during pregnancy was reported by <10% of women at baseline and increased by 
~10 percentage points by endline. However, the increase was comparable in the intervention and 
comparison districts, providing no evidence that the intervention contributed to this 
improvement. In contrast to the lack of intervention effect on the program targets for receipt and 
consumption of IFA supplements, there was a significant positive effect on compliance with IFA 
supplement intake. The proportion of women who consumed all of the supplements they 
received approximately doubled in both districts, but the increase was significantly larger in the 
intervention than comparison district. 
Table 4.9. Receipt and consumption of IFA supplements during the most recent pregnancy among mothers of 
children 0-23 months who received any supplements in Andhra Pradesh 
% 
Intervention (Karimnagar) Comparison (Rangareddy) 
Indicators of 
IFA 
supplementation Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Received 90+ 
supplements 
26.9 26.5 -0.4 28.9 31.5 2.6 -3 
Consumed 90+ 
supplements 
7.5 17.8 10.3 6.3 18.4 12.1 -1.8 
Consumed all 
supplements 
received 
24.7 57.2 32.5 22.5 44.4 21.9 10.6* 
Total # of 
mothers who 
received any 
IFA 
968 1403 1841 2014 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 41 
Uttar Pradesh. Receipt of any IFA supplements during the most recent pregnancy was >50% in 
both districts at baseline (Table 4.10). This proportion increased significantly by endline in the 
intervention district, but was unchanged in the comparison district. By endline, IFA 
supplementation coverage had reach 76% in the intervention district. 
Table 4.10. Receipt of any IFA supplements during the most recent pregnancy among mothers of children 0-
23 months in Uttar Pradesh 
Receipt of any IFAsupplements (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Receipt of any 
IFA 
supplements 
62.7 75.6 12.9 56.5 55.4 -1.1 14.0* 
Total # of 
mothers 
2445 2410 2388 2305 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Receipt and consumption of at least 90 IFA supplements during pregnancy increased 
significantly at endline in the intervention relative to the comparison district (Table 4.11). 
Among women receiving any IFA supplements, receipt of at least 90 supplements during 
pregnancy increased from 50 to 65% in the intervention district, but remained unchanged in the 
comparison district. Consumption of at least 90 supplements doubled at endline in the 
intervention district to achieve approximately 30% coverage relative to <20% coverage in the 
comparison district. The project intervention was also effective at increasing compliance with 
IFA supplementation, almost doubling at endline the proportion of women who reported 
consumption of all IFA supplements received. 
Table 4.11 Receipt and consumption of IFA supplements during the most recent pregnancy among mothers 
of children 0-23 months who received any supplements in Uttar Pradesh 
% 
Intervention (Barabanki) Comparison (Unnao) 
IFA 
supplementation 
indicator Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Received 90+ 
supplements 
49.8 65.3 15.5 52.7 53.1 0.4 15.1* 
Consumed 90+ 
supplements 
14.6 36.4 21.8 13.4 18.6 5.2 16.6* 
Consumed all 
supplements 
received 
25.3 45.4 20.1 21.5 31.7 10.2 9.9* 
Total # of 
mothers 
1534 1821 1348 1278 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Nutrition Health Days
Nutrition Health Days (NHD), fixed-day, fixed-site delivery of health and nutrition services for 
mothers and their children, were among the four “best practices” employed by the INHP-II 
2/20/2008 42 
intervention to improve maternal and child health outcomes. CARE integrated distribution of 
supplementary food from the ICDS into community-based provision of health and nutrition 
services from the RCH program to increase NHD coverage. The services offered for pregnant 
women included provision of SN, ANC check-ups, tetanus toxoid (TT) immunizations, antenatal 
IFA supplements, monitoring of pregnancy weight gain, provision of safe birth kits, and health 
education and counseling. Services for lactating women, infants and young children included 
provision of SN, immunization, growth monitoring, vitamin A syrup, and health education and 
counseling. 
Andhra Pradesh. Recent attendance at NHD was low (<20%) at baseline in both districts, but it 
was twice as common among mothers of children 0-23 months compared to currently pregnant 
women (Table 4.12). Participation doubled approximately among pregnant women in the 
intervention district relative to a slight decrease in the comparison district. However, 
participation by mothers of children 0-23 months was unchanged in the intervention district, but 
decreased by more than 30% in the comparison district, yielding a significant relative difference. 
Among pregnant women, the 4 most commonly provided NHD services were provision of SN 
(as take-home rations), monitoring of pregnancy weight gain, TT immunization, and antenatal 
IFA supplements. Mothers and their children 0-23 months of age were most likely to receive SN, 
child immunizations, growth monitoring, and health education and counseling. 
Table 4.12. Attendance at Nutrition Health Days (NHD) in the past 3 months among pregnant women and 
mothers of children 0-23 months in Andhra Pradesh 
Attendance at NHD in the past 3 months, % (n) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
6.5 
(582) 
14.1 
(573) 
7.6 4.6 
(632) 
4.3 
(705) 
-0.3 7.9* 
Mothers of 
children 0-23 
months 
13.4 
(2270) 
15.1 
(2446) 
1.7 8.5 
(2291) 
4.9 
(2430) 
-3.6 5.3* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Coverage of NHD varied across the evaluation period in the AP districts. Among pregnant 
women, coverage almost tripled to 20% by the Adequacy I survey in early 2005 (Figure 4.5). 
Increases were more modest among mothers of children 0-23 months. However, coverage 
decreased to <10% in both groups by the Adequacy II survey. By the endline survey in early 
2006, coverage had increased again, approximately doubling in the last 6 months of 2005. The 
reasons for the variability in coverage across the evaluation period are unclear. Lack of 
awareness of NHD among mothers was much higher in the comparison (82-6%) than the 
intervention (59%) district at endline. Nonetheless, the proportion of mothers unaware of NHD 
in the intervention district was quite high and represented an increase from baseline (42-4%). 
2/20/2008 43 
Figure 4.5. Attendance at Nutrition Health Days in the past 3 months at each survey round among pregnant 
women and mothers of children 0-23 months in Andhra Pradesh intervention district (Karimnagar) 
7
20
9 13 14 16
8
15
0
10
20
30
40
50
60
70
80
90
100
Baseline Adequacy I Adequacy II Endline
Attended NHD in past 3 mo (%)
Pregnant women
Mothers of children < 2y
Uttar Pradesh. Attendance at NHD was virtually non-existent in both UP districts at baseline, 
suggesting that NHD may not have been held in the months immediately proceeding the baseline 
survey (Table 4.13). The reasons for this are not known. However, dramatic increases in NHD 
participation were observed at endline among pregnant women and mothers of children 0-23 
months in the intervention district. Coverage in the intervention district was approximately one￾third in each target group at endline compared to <3% in the comparison district. The list of most 
commonly provided services was similar to the AP intervention district: provision of SN, TT 
immunization, antenatal IFA supplements, and monitoring of pregnancy weight gain among 
pregnant women, and provision of SN, child immunizations, growth monitoring, and provision 
of vitamin A syrup among mothers and their children 0-23 months of age. 
Table 4.13. Attendance at Nutrition Health Days (NHD) in the past 3 months among pregnant women and 
mothers of children 0-23 months in Uttar Pradesh 
Attendance at NHD in the past 3 months, % (n) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
1.5 
(745) 
38.9 
(723) 
37.4 0 
(706) 
2.2 
(673) 
2.2 35.2* 
Mothers of 
children 0-23 
months 
1.2 
(2288) 
33.7 
(2273) 
32.5 0.04 
(2230) 
2.8 
(2150) 
2.76 29.7* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Participation in NHD among pregnant women and mothers of children 0-23 months in the UP 
intervention district did not improve during the first year of the evaluation period (in 2004), but 
thereafter increased markedly in 2005 (Figure 4.6). Coverage in both target groups increased 
eight-fold between the Adequacy I and endline surveys. Unlike in AP, lack of awareness of NHD 
dropped dramatically in the intervention district from 57-59% at baseline to 15-20% at endline. 
Three-quarters of women surveyed in the comparison district were unaware of NHD at endline. 
2/20/2008 44 
Figure 4.6. Attendance at Nutrition Health Days in the past 3 months at each survey round among pregnant 
women and mothers of children 0-23 months in Uttar Pradesh intervention district (Barabanki)
2 5
24
39
1 4
19
34
0
10
20
30
40
50
60
70
80
90
100
Baseline Adequacy I Adequacy II Endline
Attended NHD in past 3 mo (%)
Pregnant women
Mothers of children < 2y
Summary
The task facing CARE with regard to improvement of coverage and quality of antenatal care 
services varied by state and by specific program component. With regard to receipt of ANC 
visits by pregnant women, the underlying conditions in AP and UP states differed tremendously. 
The facility-based health care infrastructure for provision of ANC services in AP state was 
already strong prior to the evaluation, so coverage of ANC visits among pregnant women and 
achievement of the program target of at least 3 visits was nearly universal in the intervention and 
comparison districts at baseline and endline. Extensive use of health services in the private sector 
was common also. By contrast, baseline receipt of ANC visits was quite low in UP. However, 
INHP-II was effective at increasing receipt of ANC visits in the intervention district so that 50% 
of women reported any ANC visits during their previous pregnancy and 25% reported receipt of 
3 or more visits. 
The INHP-II program was highly effective in both states at increasing service provider contacts 
with pregnant women at home and in communities. In AP state, the proportion of women who 
reported at least one home visit by a service provider during pregnancy increased from 
approximately one-third at baseline to 50% among mothers of children 0-23 months and to 70% 
among third trimester pregnant women at endline. In UP, the coverage was even lower at 
baseline (~20%), and increased to 60% and over 80% among mothers and third-trimester 
pregnant women, respectively. In both states, the dramatic improvements were driven primarily 
by large increases in the coverage and frequency of home visits by AWW. In UP state, home 
contacts during pregnancy with ANM and CA increased significantly also. However, increased 
contacts with ANM were more modest in AP state, and were significant among third trimester 
pregnant women only. Change Agents did not make home visits to provide advice in AP, but did 
have some success in UP. Despite the significant increases in home visits by service providers 
during pregnancy, women were not more likely to report receipt of maternal nutrition advice 
2/20/2008 45 
during these visits. Only pregnant women in the UP intervention district were more likely to 
receive advice on maternal diet and rest from ANM during these contacts. 
Coverage of home contacts by service providers was consistently higher among third trimester 
pregnant women than mothers reporting on their most recent pregnancy in both states. This 
discrepancy could be explained by differences in the recall period as mothers could be reporting 
on a pregnancy that took place as far back as 2 and a half years ago. Alternatively, reported 
contacts by currently pregnant women may reflect more accurately the current state of the 
program because exposure was more recent. The original INHP-II strategy for increasing home 
contacts and community outreach by service providers was to encourage these activities through 
Block Level Advisory Committees (BLAC). Medical Officers attending these meetings with the 
CDPO had macro level discussions of operational issues, but this did not translate into corrective 
actions at the field level very often. After the MTR, the strategy shifted to better align efforts 
between ICDS and the health sector. ICDS supervisors attended PHC monthly reviews to discuss 
service provider activities in communities with the Medical Officers, thus increasing 
convergence between the ANM and AWW in their outreach efforts. 
Recruitment and training of CA to support AWW in their community outreach efforts was an 
initial strategy of INHP-II that was de-emphasized after the MTR. Nonetheless, CA activities did 
continue where the program was successful at putting them in place. At endline, the data 
suggested some success with recruitment and training of CA to provide health and nutrition 
advice to pregnant women and their children in UP, but not AP. One reason for this discrepancy 
may relate to the CA candidates who were recruited in the different states. In the UP intervention 
district, CA were recruited from among Dai, mothers-in-law, and other men and women of 
varied ages. After their training, retention was fairly good. By contrast, ~50% of the CA 
recruited in the AP intervention district were adolescent girls. The dropout rate in this group of 
CA was very high because of migration due to marriage. This necessitated additional recruitment 
and training that was not entirely successful because newly recruited CA and existing CA were 
trained together. The reasons for this difference in the types of CA candidates are unknown. 
 The effectiveness of INHP-II at increasing coverage and consumption of antenatal IFA 
supplementation varied by state. In AP state, coverage was much lower in the intervention than 
comparison district initially, and the program had only limited success at improving it. Coverage 
had increased to over 50% at endline, but only one quarter of women reported receipt of at least 
90 tablets during pregnancy and less than 20% consumed that amount. More sizeable increases 
in coverage were achieved in the UP intervention district where 65% of women received at least 
90 tablets during pregnancy and consumption of at least 90 tablets more than doubled to 36% at 
endline. In both states, women in the intervention districts were at least twice as likely at endline 
to consume all the IFA supplements they received. Increased contacts with service providers 
during pregnancy may have influenced compliance positively. 
The state-specific trends with NHD followed along the same lines as antenatal IFA 
supplementation. Attendance at NHD in the past 3 months by pregnant women and mothers of 
children 0-23 months increased significantly in the intervention relative to the comparison 
district in both states. However, the changes in AP were quite modest, increasing coverage to 
~15% only in both groups, whereas the improvements in UP were dramatic, with attendance 
2/20/2008 46 
increasing from almost nothing to ~34-39% in the two groups. The pace of implementation also 
varied by state. The increases in AP occurred within the first year of the evaluation period, but 
were not maintained consistently or improved upon further in year 2. In UP, the gains in 
coverage occurred almost entirely in the second year after the MTR. 
2/20/2008 47 
CHAPTER 5: MATERNAL NUTRITION 
Supplementary Nutrition
The INHP-II Program was designed to support the ICDS supplementary feeding intervention for 
women during pregnancy and the first 6 months postpartum. USAID’s Title II Program provided 
two commodities, corn soya blend (CSB) and refined vegetable oil (RVO), for distribution in this 
program. However, the GOI disallowed import of CSB in mid 2002 because of concerns about 
genetically modified foods. Subsequently, CARE’s INHP-II mandate was amended in 2003 to 
facilitate the transition from imported CSB to locally available foods in the ICDS Program. 
Vegetable oil continued to be distributed as part of the feeding intervention. 
In AP state, the period during which Supplementary Nutrition (SN) was unavailable was 
relatively brief as the State government initiated provision of locally available food through 
ICDS as of January 2003. Therefore, recall of receipt of SN during the most recent pregnancy 
and postpartum at the time of the baseline survey in early 2004 did include the ~6 month period 
in 2002 when SN was unavailable. By the time of the endline survey, the recall period occurred 
after the transition to the state government’s program with locally available food. 
The shift from CSB to locally available foods for SN was more prolonged in UP state. There 
were some efforts to supply locally available food during mid-2003, but this initial effort was 
unsuccessful. The UP State government succeeded in re-initiating the ICDS supplementary 
feeding program successfully in February 2004. Therefore, recall of SN receipt at the baseline 
survey included up to a one and a half year period during which provision of SN was sporadic or 
non-existent. Endline recall covered a period when the SN program had already transitioned to 
locally available food with the exception of mothers with the oldest children who were pregnant 
in 2003. 
Andhra Pradesh. Receipt of SN during the most recent pregnancy among mothers of children 0-
23 months was fairly low at baseline, but was higher in the intervention (38%) than in the 
comparison (22%) district (Table 5.1). By endline, this proportion had increased by over 20 
percentage points in both districts, so that almost two-thirds of intervention district women and 
almost half of comparison district women reported receipt of SN sometime during their most 
recent pregnancy. However, the increase in the intervention district cannot be attributed to the 
INHP-II program because of the comparable increase in the comparison district during the same 
period. It is important to keep in mind that these proportions were not completely reflective of 
the current status of the ICDS supplementary feeding program because these data referred to a 
period as far back as two and a half years prior to each survey. The large increases in coverage in 
both districts most likely reflect the fact that the endline data cover the period when the AP state 
program had completed its transition to provision of locally available food. It should be kept in 
mind that CARE played a key role in facilitating this transition by the ICDS program at the state 
level during that period, so the INHP-II program impact may not be as easily discernable at the 
district level for this SN indicator. 
2/20/2008 48 
Table 5.1. Receipt of Supplementary Nutrition during pregnancy among currently pregnant women and 
mothers of children 0-23 months during the last pregnancy in Andhra Pradesh 
Supplementary nutrition use during pregnancy, % (n) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-23 
months (during 
last pregnancy) 
38.1 
(2369) 
62.6 
(2575) 
24.5 22.4 
(2476) 
45.5 
(2656) 
23.1 1.4 
Pregnant 
women 
11.0 
(582) 
48.9 
(573) 
37.9 8.2 
(632) 
25.4 
(705) 
17.2 20.7* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Coverage of SN was lower among currently pregnant women in both districts at baseline (Table 
5.1). These numbers are expected to be slightly lower than among women recalling receipt of SN 
after the end of pregnancy because women in the pregnancy survey samples were at various 
stages of pregnancy. However, this bias in the data is consistent between districts, and therefore, 
should not affect the evaluation of program impact. Coverage among currently pregnant women 
increased more than twice as much in the intervention (38 percentage points) as in the 
comparison (17 percentage points) district by endline, resulting in a statistically significant 
intervention effect on SN receipt. 
Receipt of SN during the postpartum period among mothers of children 0-5 months followed the 
same trends as were observed among currently pregnant women (Table 5.2). Coverage was low 
overall at baseline, but twice as high in the intervention as compared to the comparison district. 
At endline, the increase in the proportion of mothers who received SN postpartum was 
significantly greater in the intervention relative to the comparison district. These results among 
currently pregnant and postpartum women represent the current state of program activities, 
suggesting that the INHP-II program was more effective at improving SN coverage among 
pregnant and postpartum women than the ICDS services alone. 
Table 5.2. Receipt of Supplementary Nutrition during the postpartum period among mothers of children 0-5 
months in Andhra Pradesh 
Supplementary Nutrition use during the postpartum period (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-5 
months 
19.2 42.1 22.9 11.0 21.2 10.2 12.7* 
Total # of 
mothers 
625 563 648 622 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
The data across all survey rounds suggest that the intervention was active early in the evaluation 
period with respect to the provision of SN through ICDS services (Figure 5.1). Coverage among 
both pregnant and postpartum women increased dramatically in 2004 between the baseline and 
2/20/2008 49 
Adequacy I surveys. However, there is some evidence that the program was encountering some 
difficulties in late 2005 as the endline numbers among postpartum women indicate a large drop 
in coverage. 
Figure 5.1. Receipt of Supplementary Nutrition during the last pregnancy and the postpartum period in 
Andhra Pradesh intervention district (Karimnagar) by survey round 
38
56
70
63
19
66 63
42
0
10
20
30
40
50
60
70
80
90
100
Baseline Adequacy I Adequacy II Endline
Receipt of Supplementary Nutrition (%)
Pregnancy
Postpartum period
Information on consumption of SN was collected at the endline survey in early 2006. Therefore, 
the proportion of women who consumed SN at endline can be compared between the 
intervention and comparison districts as a proxy for assessing whether the INHP-II program was 
effective at increasing SN intake in the intervention districts. Both during pregnancy and the 
postpartum period, women in the intervention district were significantly more likely than those in 
the comparison district to have consumed SN (Figure 5.2). When these proportions are compared 
to the endline coverage of SN receipt reported in Tables 5.1 and 5.2, it is apparent that SN was 
consumed by virtually all women who received any SN during pregnancy or the postpartum 
period. 
2/20/2008 50 
Figure 5.2. Consumption of Supplementary Nutrition during pregnancy and the postpartum period at 
endline in Andhra Pradesh 
62
48
36
44
25
18
0
10
20
30
40
50
60
70
80
90
100
Pregnant (moms of children 0-
23 mo)
Currently pregnant women Postpartum women
Consumption of SN (%)
AP intervention
AP comparison
* Statistical test result comparing intervention and comparison districts, significant at p<0.05 for all 3 target groups. 
Uttar Pradesh. Baseline coverage of SN in UP districts was lower than in AP districts, but the 
effectiveness of the intervention at improving coverage was greater. Among mothers of children 
0-23 months at baseline, more than one-third in the intervention district and one-quarter in the 
comparison district received SN during their most recent pregnancy (Table 5.3). This proportion 
increased in the intervention district, but decreased in the comparison district, yielding a 
statistically significant relative difference for this indicator. The positive effect of the INHP-II 
program on receipt of SN was even stronger among currently pregnant women, increasing from 
5% at baseline to 48% at endline in the intervention district. A more modest increase to only 
18% at endline was observed in the comparison district. 
Table 5.3. Receipt of Supplementary Nutrition during pregnancy among currently pregnant women and 
mothers of children 0- 23 months during the last pregnancy in Uttar Pradesh 
Supplementary nutrition use during pregnancy, % (n) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-23 
months (during 
last pregnancy) 
38.9 
(2445) 
50.2 
(2413) 
11.3 24.9 
(2388) 
19.5 
(2306) 
-5.4 16.7* 
Pregnant 
women 
4.7 
(745) 
48.1 
(724) 
43.4 3.7 
(699) 
17.6 
(676) 
13.9 29.5* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The trend for receipt of SN among women in the postpartum period followed that of currently 
pregnant women. Baseline coverage was <10% in both districts, most likely reflecting the 
current status of the supplementary feeding program in UP state in late 2003 when provision of 
locally available commodities by the ICDS was not yet up and running fully (Table 5.4). 
However, the proportion of mothers who received SN during the postpartum period increased 
2/20/2008 51 
dramatically (by 37 percentage points) at endline in the intervention district relative to a modest 
increase in the comparison district. 
Table 5.4. Receipt of Supplementary Nutrition during the postpartum period among mothers of children 0-5 
months in Uttar Pradesh 
Supplementary Nutrition use during the postpartum period (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-5 
months 
7.3 44.0 36.7 5.5 14.2 8.7 28.0* 
Total # of 
mothers 
804 763 731 682 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The pace of implementation for distribution of SN was similar for all target groups in the UP 
intervention district. Improvements in coverage occurred early in the evaluation period, with 
significant increases observed at the Adequacy I survey in early 2005. Coverage remained at the 
same level during 2005 until the endline survey in early 2006. These data suggest that the 
intervention was fairly effective at facilitating supplementary food distribution through the ICDS 
services once the UP state government had transitioned its program to the use of locally available 
food in early 2004. 
Figure 5.3. Receipt of Supplementary Nutrition during the last pregnancy and the postpartum period in 
Uttar Pradesh intervention district (Barabanki) by survey round 
39
48
53 50
7
41
46 44
0
10
20
30
40
50
60
70
80
90
100
Baseline Adequacy I Adequacy II Endline
Receipt of Supplementary Nutrition (%)
Pregnancy
Postpartum period
As with AP districts, UP data were available at endline only for assessment of program 
effectiveness on SN intake. Both pregnant and postpartum women in the UP intervention district 
were significantly more likely than those in the comparison district to have consumed SN (Figure 
5.4). When these proportions are compared to the endline coverage of SN receipt reported in 
Tables 5.3 and 5.4, some gap between receipt and consumption of SN is noticeable, particularly 
among women during the postpartum period (receipt=44%, consumption=27%). Supplementary 
food from ICDS contributed to the food supply in many intervention district households, but was 
2/20/2008 52 
not necessarily consumed by women themselves for reasons that are unknown. Pregnant and 
postpartum women may have disliked the locally available SN commodity or may have 
prioritized its use for other household members. 
Figure 5.4. Consumption of Supplementary Nutrition during pregnancy and the postpartum period at 
endline in Uttar Pradesh 
38 38
27
15 13
5
0
10
20
30
40
50
60
70
80
90
100
Pregnant (moms of children 0-
23 mo)
Currently pregnant women Postpartum women
Consumption of SN (%)
UP intervention
UP comparison
* Statistical test result comparing intervention and comparison districts, significant at p<0.05 for all 3 target groups 
Maternal Diet
INHP-II advice on maternal diet focuses on two messages for women during pregnancy and the 
first 6 months postpartum: 1) consume an additional meal every day, and 2) consume all 
available foods in the home. In addition to assessing whether women were more likely to receive 
these messages during pregnancy from service providers (see chapter 4), data were collected on 
current dietary practices and intake during pregnancy and postpartum to evaluate the impact of 
the INHP-II intervention on maternal diet. 
Meals and snacks
Andhra Pradesh. More than two-thirds of women in the AP districts usually consumed at least 3 
meals a day at baseline (Table 5.7). Pregnant women were the least likely and mothers of 
children 6-23 months of age were most likely to consume ≥3 meals a day in both districts. At 
endline, the proportion of women consuming at least 3 meals per day increased significantly in 
the intervention district relative to little or no change in the comparison district. It is interesting 
to note that the message about increased meal intake, targeted at women during pregnancy and 
the first 6 months postpartum, had a positive effect on the number of meals consumed among all 
groups of women surveyed. 
2/20/2008 53 
Table 5.7. Meal consumption among pregnant women and mothers of children 0-23 months in Andhra 
Pradesh 
≥3 meals consumed per day, % (n) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Target group 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
68.7 
(582) 
83.9 
(572) 
15.2 73.9 
(632) 
77.5 
(703) 
3.6 11.6* 
Mothers of 
children 0-5 
months 
78.0 
(624) 
94.0 
(550) 
16.0 83.1 
(637) 
83.4 
(596) 
0.3 15.7* 
Mothers of 
children 6-23 
months 
81.5 
(1646) 
91.6 
(1886) 
10.1 84.4 
(1652) 
82.8 
(1815) 
-1.6 11.7* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Trends in usual snack consumption differed from meal consumption in the AP districts. Snacks 
were most common among pregnant and least common among newly delivered mothers of 
children 0-5 months of age in both districts (Table 5.8). Snack intake decreased among all groups 
of women in the intervention district relative to increases in the comparison district. One possible 
explanation for these opposing trends in the intervention district may relate to how women acted 
upon the nutrition message they received to add one meal a day to their diets. A period of dietary 
intake at some point in the day may have been increased from a snack to a full meal by the added 
consumption of a staple food such as rice or other grains. It is not known whether the net 
difference between an added meal and an omitted snack resulted in additional caloric intake, but 
it is a reasonable assumption given the Indian definition of a meal as including a staple food. 
Alternatively, micronutrient intake may have been unchanged or even reduced by this 
substitution. 
Table 5.8. Snack consumption among pregnant women and mothers of children 0-23 months in Andhra 
Pradesh 
≥1 snack consumed per day, % (n) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Target group 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
45.4 
(582) 
38.3 
(572) 
-7.1 37.5 
(632) 
40.4 
(703) 
2.9 -10.0* 
Mothers of 
children 0-5 
months 
30.3 
(624) 
21.5 
(550) 
-8.8 19.6 
(637) 
30.4 
(596) 
10.8 -19.6* 
Mothers of 
children 6-23 
months 
34.6 
(1646) 
25.5 
(1886) 
-9.1 24.8 
(1652) 
35.7 
(1815) 
10.9 -20.0* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. Consumption of at least 3 meals a day was much less common among women in 
UP than AP districts (Table 5.9). Only 14% of pregnant women and 25% of postpartum women 
in the intervention district consumed at least 3 meals a day at baseline, and the proportions 
comparable in the comparison district. By endline, consumption of at least 3 meals a day had 
2/20/2008 54 
increased 4 times as much among pregnant women and twice as much among postpartum women 
in the intervention relative to the comparison district. 
Table 5.9. Meal consumption among pregnant women and mothers of children 0-23 months in Uttar Pradesh 
≥3 meals consumed per day, % (n) 
Intervention (Barabanki) Comparison (Unnao) 
Target group 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
14.2 
(745) 
37.8 
(723) 
23.6 20.0 
(699) 
26.0 
(676) 
6.0 17.6* 
Mothers of 
children 0-5 
months 
25.4 
(802) 
42.6 
(758) 
17.2 26.2 
(724) 
34.7 
(677) 
8.5 8.7* 
Mothers of 
children 6-23 
months 
25.0 
(1485) 
37.6 
(1519) 
12.6 29.5 
(1504) 
36.4 
(1473) 
6.9 5.7* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Snack consumption was much more common in the UP intervention district than in the AP 
districts. This may be compensation for the lower usual meal intake by UP women. However, 
women in the UP comparison district were approximately half as likely as intervention district 
women to have reported snack consumption at baseline (Table 5.10). Snack consumption among 
pregnant women increased modestly in both districts, resulting in no impact of the intervention 
in this target group. By contrast, the program was very effective at increasing snack intake 
among postpartum women so that 90% reported consumption of at least one snack a day in the 
intervention district relative to 43-4% in the comparison district at endline. The more consistent 
positive effects of the intervention on meal and snack intake in UP state suggest that women in 
the intervention district may have increased their caloric and micronutrient intake relative to their 
comparison district peers. 
Table 5.10. Snack consumption among pregnant women and mothers of children 0-23 months in Uttar 
Pradesh 
≥1 snack consumed per day, % (n) 
Intervention (Barabanki) Comparison (Unnao) 
Target group 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
63.1 
(745) 
76.4 
(723) 
13.3 34.5 
(699) 
44.4 
(676) 
9.9 3.4 
Mothers of 
children 0-5 
months 
59.9 
(802) 
90.0 
(758) 
30.1 34.8 
(724) 
44.2 
(668) 
9.4 20.9* 
Mothers of 
children 6-23 
months 
60.8 
(1485) 
89.7 
(1519) 
28.9 34.8 
(1505) 
42.6 
(1473) 
7.8 21.1* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
It is interesting to note that the maternal nutrition messages aimed at women during pregnancy 
and the early postpartum period (0-5 months after delivery) had a significant positive effect on 
the dietary practices of mothers of older children (6-23 months) as well. Whether this 
2/20/2008 55 
phenomena is due to insufficient targeting of health messages to the appropriate audiences, 
overly general information in the messages, or the desire of women to improve their diets in 
general is not clear. In all fairness, women of reproductive age in the study population are 
cycling rapidly through these various reproductive stages so that distinctions made by the 
intervention (e.g. recently delivered mothers of children 0-5 months versus mothers of children 
6-23 months) may not seem very relevant in the real life context. 
Dietary intake
Usual dietary intake among women was characterized by data collected from a 7-day food 
frequency recall. Results are reported here as intake of specific categories of food one or more 
times in the past 7 days. Staple foods included rice, bread, and other foods made from cereal 
grains. Intake of staple foods is not reported here because intake was 100% in all groups. 
Andhra Pradesh. AP state women in all target groups had very diverse diets that included 
frequent consumption of legumes, meat and eggs, dairy products, fruits and vegetables in 
addition to cereal grains (Tables 5.11-5.13). Legumes (65-95%), meat/fish/chicken and eggs (65-
79%), and dark green leafy vegetables (DGLV; 44-72%) were the most commonly consumed 
foods among pregnant and postpartum women at baseline and endline. Usual intake of legumes 
and DGLV increased significantly in the intervention relative to the comparison district for all 3 
target groups of women. Intake of yellow-orange fruits was <11% among non-pregnant and 18-
19% among pregnant women at baseline, but these proportions increased significantly for all 3 
groups at endline in the intervention relative to the comparison district. Fewer than one-quarter 
of women reported any intake of yellow-orange vegetables in the past week at baseline (or the 
Adequacy I survey for non-pregnant women). This proportion remained unchanged or decreased 
significantly at endline among non-pregnant women. By contrast, it increased to over 80% 
among currently pregnant women in both districts. The reason for this dramatic increase in 
consumption of yellow-orange vegetables cannot be explained by seasonal variation as baseline 
and endline surveys were conducted in the same season of the year. It is possible that nutrition 
advice given to pregnant women emphasized the inclusion of yellow-orange vegetables in the 
diet as a source of micronutrients to improve pregnancy outcomes. Dairy products were 
consumed by 24-52% of women at baseline, but this proportion decreased significantly in the 
intervention relative to the comparison district among non-pregnant women only. Intake of SN 
was low at baseline and increased significantly among pregnant women only in the intervention 
relative to the comparison district. 
2/20/2008 56 
Table 5.11. Dietary intake during the past 7 days among pregnant women in Andhra Pradesh 
Intake ≥1 times during the past 7 days among pregnant women (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Legumes 85.4 94.9 9.5 92.7 91.0 -1.7 11.2* 
Dairy 52.2 49.8 -2.4 38.1 35.0 -3.1 0.7 
Meat, chicken, 
fish and eggs 
70.1 79.4 9.4 70.6 76.8 6.2 3.2 
Dark green 
leafy vegetables 
48.8 68.2 19.4 72.0 70.1 -1.9 21.3* 
Yellow-orange 
vegetables + 
14.3 82.3 78 21.0 89.6 68.6 9.4 
Yellow-orange 
fruits 
17.9 29.8 11.9 18.7 19.4 0.7 11.2* 
Other fruits 51.9 64.1 12.2 47.9 65.4 17.5 -5.3 
Butter/oil/fat 10.5 10.8 0.3 9.5 6.1 -3.4 3.7 
Supplementary 
Nutrition 
14.8 24.0 9.2 10.4 10.2 -0.2 9.4* 
Total # of 
pregnant 
women 
582 572 632 702 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
+ sweet potatoes are not included 
Table 5.12. Dietary intake during the past 7 days among mothers of children 0-5 months in Andhra Pradesh 
Intake ≥1 times during the past 7 days among mothers of children 0-5 months (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Legumes 65.5 76.7 11.6 84.0 85.5 1.5 10.1* 
Dairy 46.6 36.4 -10.2 30.4 28.4 -2.0 -8.2* 
Meat, chicken, 
fish and eggs 
65.4 75.3 9.9 67.9 75.3 7.4 2.5 
Dark green 
leafy vegetables 
56.1 72.9 16.8 76.2 75.3 -0.9 17.7* 
Yellow-orange 
vegetables #+ 
21.7 23.3 1.6 7.2 14.8 7.6 -6 
Yellow-orange 
fruits 
8.7 24.6 15.9 9.9 16.1 6.2 9.7* 
Other fruits 34.6 48.6 14.0 26.8 49.3 22.5 -8.5* 
Butter/oil/fat 13.5 8.7 -5.2 9.1 7.7 -1.4 -3.8 
Supplementary 
Nutrition 
9.8 18.8 9.0 5.3 10.8 5.5 3.5 
Total # of 
mothers of 
children 0-5 mo 
624 550 638 594 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05. 
# Baseline data from Adequacy I data 
+ Sweet potatoes not included 
2/20/2008 57 
Table 5.13. Dietary intake during the past 7 days among mothers of children 6-23 months in Andhra Pradesh 
Intake ≥1 times during the past 7 days among mothers of children 6-23 months (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Legumes 83.6 92.1 8.5 92.3 94.0 1.7 6.8* 
Dairy 41.1 33.7 -7.4 26.1 24.0 -2.1 -5.3* 
Meat, chicken, 
fish and eggs 
64.8 77.7 12.9 68.5 77.2 8.7 4.2* 
Dark green 
leafy vegetables 
44.1 62.8 18.7 72.3 69.4 -2.9 21.6* 
Yellow-orange 
vegetables #+ 
22.2 18.0 -4.2 2.8 11.2 8.4 -12.6* 
Yellow-orange 
fruits 
9.9 22.2 12.3 11.0 17.3 6.2 6.1* 
Other fruits 37.4 46.3 8.9 37.5 56.2 18.7 -9.8* 
Butter/oil/fat 9.1 9.7 0.6 6.2 6.1 -0.1 0.7 
Supplementary 
Nutrition 
13.2 10.1 -3.1 6.5 8.7 2.2 -5.3* 
Total # of 
mothers of 
children 6-23 
mo 
1646 1886 1653 1814 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
 # Baseline data from Adequacy I data 
+ Sweet potatoes not included 
Uttar Pradesh. Legume and DGLV intake in the past week was nearly universal (legumes: 
≥94%, DGLV: ≥88%) at baseline among women in the UP intervention district (Tables 5.14-16). 
Significant increases were not observed at endline because of the high baseline intake. Dairy 
consumption was ~40% at baseline in the intervention district for pregnant and non-pregnant 
women. Large increases in consumption were reported at endline in both districts except among 
mothers of children 0-5 months of age where the relative difference was statistically significant. 
Intake of meat, fish, chicken and eggs was much lower in UP than AP at baseline (16-24%), and 
modest increases in this proportion were observed at endline in all target groups. However, the 
relative difference in the increase in the intervention relative to the comparison district was 
statistically significant for pregnant women and mothers of children 6-23 months only. 
Consumption of yellow-orange vegetables at baseline varied greatly among pregnant (18%) and 
postpartum (0-5 months: 34%, 6-23 months: 65%) women. Intake increased fairly dramatically 
at endline for pregnant women and mothers of children 0-5 months, but these increases were 
comparable in the intervention and comparison districts. By contrast, there was a slight decrease 
in the proportion consuming yellow-orange vegetables among mothers of children 6-23 months 
who had the highest intake at baseline. Yellow-orange fruits were consumed by <7% of all 
women at baseline, but significant increases were reported in the intervention relative to the 
comparison district for all three target groups. Consumption of butter, oil, or fat increased to 
~40% at endline for pregnant and postpartum women, but this change was statistically significant 
for non-pregnant women only. Finally, intake of SN from the ICDS was essentially non-existent 
at baseline, but increased dramatically to between 23 and 31% by endline. 
2/20/2008 58 
Table 5.14. Dietary intake during the past 7 days among pregnant women in Uttar Pradesh 
Intake ≥1 times during the past 7 days pregnant women (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pulses 94.9 95.6 0.7 73.4 75.7 2.3 -1.6 
Dairy 42.7 62.1 19.4 33.3 54.8 21.5 -2.1 
Meat, chicken, 
fish and eggs 
16.9 28.3 11.4 23.8 19.8 -4.0 15.4* 
Dark green 
leafy vegetables 
85.4 88.1 2.7 56.9 63.1 6.2 -3.5 
Yellow-orange 
vegetables + 
18.0 34.2 16.2 22.2 46.7 24.5 -8.3 
Yellow-orange 
fruits 
5.9 16.6 10.7 6.6 10.2 3.6 7.1* 
Other fruits 29.9 48.0 18.1 32.8 45.0 12.2 5.9 
Butter/oil/fat 20.4 41.9 21.5 21.2 36.4 15.2 6.3 
Supplementary 
Nutrition 
1.3 28.4 27.1 3.0 7.4 4.4 22.7* 
Total # of 
pregnant 
women 
745 723 699 676 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
+ Sweet potatoes not included 
Table 5.15. Dietary intake during the past 7 days among mothers of children 0-5 months in Uttar Pradesh 
Intake ≥1 times during the past 7 days among mothers of children 0-5 months (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pulses 97.3 96.7 -0.6 77.5 75.6 -1.9 1.3 
Dairy 39.0 60.8 21.8 37.7 47.7 10.0 11.8* 
Meat, chicken, 
fish and eggs 
16.7 24.7 8.0 18.0 20.3 2.3 5.7 
Dark green 
leafy vegetables 
85.2 86.9 1.7 61.6 67.8 6.2 -4.5 
Yellow-orange 
vegetables # 
33.6 66.3 32.7 31.7 60.7 29.0 3.7 
Yellow-orange 
fruits 
4.5 15.8 11.3 2.6 8.5 5.9 5.4* 
Other fruits 23.6 35.6 12.0 20.4 30.1 9.7 2.3 
Butter/oil/fat 20.8 41.8 21.0 20.4 32.1 11.7 9.3* 
Supplementary 
Nutrition 
1.3 31.1 29.8 1.2 9.0 7.8 22.0* 
Total # of 
mothers of 
children 0-5 mo 
802 758 724 677 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
# baseline data from Adequacy I data 
2/20/2008 59 
Table 5.16. Dietary intake during the past 7 days among mothers of children 6-23 months in Uttar Pradesh 
Intake ≥1 times during the past 7 days among mothers of children 6-23 months (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pulses 97.9 97.6 -0.3 74.7 74.2 -0.5 0.2 
Dairy 38.6 57.9 19.3 31.6 47.5 15.9 3.4 
Meat, chicken, 
fish and eggs 
16.1 25.5 9.4 20.3 18.6 -1.7 11.1* 
Dark green 
leafy vegetables 
86.5 88.5 2 63.8 69.3 5.5 -3.5 
Yellow-orange 
vegetables # 
65.1 56.9 -8.2 30.6 30.4 -0.2 -8.0* 
Yellow-orange 
fruits 
4.6 13.2 8.6 4.1 7.8 3.7 4.9* 
Other fruits 26.2 38.3 12.1 25.0 35.5 10.5 1.6 
Butter/oil/fat 19.1 42.2 23.1 17.9 28.0 10.1 13.0* 
Supplementary 
Nutrition 
0.7 23.0 22.3 1.6 8.3 6.7 15.6* 
Total # of 
mothers 
1486 1518 1506 1472 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
# baseline data from Adequacy I data 
Nutritional status
Currently pregnant women and mothers of children 0-23 months had their weight and height 
measured at baseline and endline to examine whether the intervention had an effect on women’s 
nutritional status. Body Mass Index (BMI; Kg/m2
) was used as an indicator of nutritional status 
for both pregnant and non-pregnant women. It is recognized that BMI is not an optimal measure 
of nutritional status for pregnant women because of pregnancy weight gain. Nonetheless, it is an 
overall measure of nutritional status that is useful for comparing changes from baseline to 
endline in the intervention and comparison districts in each state. Malnutrition was defined as a 
BMI<18.5. Nutritional status data on pregnant women was stratified by trimester of pregnancy to 
control for confounding by pregnancy weight gain. 
Andhra Pradesh. Mean BMI among all pregnant women was lower in the intervention than in 
the comparison district at baseline, but increased by the same amount in both districts at endline 
(Table 5.17). Second and third trimester women had comparable increases in BMI at endline in 
both districts. However, there was an increase in the intervention district and a decrease in the 
comparison district among first trimester women at endline that resulted in a significant relative 
difference of 1.1. 
2/20/2008 60 
Table 5.17. Body Mass Index (BMI) among pregnant women stratified by trimester in Andhra Pradesh 
Mean BMI among pregnant women (Kg/m2
) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 19.1 19.6 0.5 20.2 19.6 -0.6 1.1* 
Second 
trimester 
20.3 20.7 0.4 20.6 21.4 0.8 -0.4 
Third trimester 22.0 22.5 0.5 22.2 22.4 0.2 0.3 
All pregnant 
women 
20.4 20.7 0.3 20.9 21.2 0.3 0 
Total # of 
pregnant 
women 
567 563 605 700 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The intervention also had a significant positive effect on nutritional status among mothers of 
children 6-23 months who were the thinnest of the 3 target groups of women in this study (Table 
5.18). Their BMI increased 0.6 units from baseline to endline relative to almost no change in the 
comparison district. By contrast, the mean BMI of mothers of children 0-5 months increased by 
approximately one-half a unit in both districts at endline. 
Table 5.18. Mean Body Mass Index among non-pregnant women in Andhra Pradesh 
Mean BMI among non-pregnant women (Kg/m2
) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-5 
months 
19.6 20.0 0.4 19.7 20.3 0.6 -0.2 
Mothers of 
children 6-23 
months 
18.5 19.1 0.6 19.3 19.4 0.1 0.5* 
Total # of non￾pregnant 
women 
2197 2426 2149 2407 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The malnutrition indicator followed the same trends as mean BMI in the AP districts. 
Malnutrition was more prevalent in the intervention than comparison district throughout the 
evaluation period. Among all pregnant women, the proportion who were malnourished decreased 
slightly in both districts at endline although the overall prevalence was ~30% higher in the 
intervention district (Table 5.19). By contrast, malnutrition prevalence among first trimester 
pregnant women decreased by 7 percentage points at endline in the intervention district relative 
to a comparable increase in the comparison district. There were modest decreases at endline in 
the proportion malnourished among second and third trimester women in both districts. 
2/20/2008 61 
Table 5.19. Malnutrition (BMI<18.5 Kg/m2
) among pregnant women stratified by trimester in Andhra 
Pradesh 
BMI<18.5 among pregnant women (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 43.9 36.6 -7.3 28.9 35.6 6.7 -14.0* 
Second 
trimester 
23.8 18.7 -5.1 15.6 12.3 -3.3 -1.8 
Third trimester 5.9 7 -1.1 8.4 4.0 -4.4 3.3 
All women 25.0 21.9 -3.1 18.4 15.7 -2.7 -0.4 
Total # of 
pregnant 
women 
567 563 605 700 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Over one-third of recently delivered mothers (of children 0-5 months) and over half of mothers 
of children 6-23 months were malnourished at baseline in the intervention district (Table 5.20). 
Prevalence was lower among women in the comparison district. Malnutrition prevalence 
decreased among both groups of women at endline, but the relative difference between the 
districts was statistically significant among mothers of children 6-23 months only. 
Table 5.20. Malnutrition (BMI<18.5 Kg/m2
) among non-pregnant women in Andhra Pradesh 
BMI<18.5 among non-pregnant women (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-5 
months 
37.8 29.8 -8.0 32.8 27.4 -5.4 -2.6 
Mothers of 
children 6-23 
months 
55.6 46.6 -9.0 43.0 43.4 0.4 -9.4* 
Total # of non￾pregnant 
women 
2197 2426 2149 2407 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
In summary, the intervention was effective at improving nutritional status and reducing 
malnutrition prevalence among non-pregnant women of children 6-23 months and women in the 
first trimester of pregnancy. Women later in pregnancy and during the early postpartum period 
also had improvements in nutritional status in the intervention district, but they were matched by 
comparable improvements in comparison district women. 
Uttar Pradesh. Surprisingly, overall nutritional status among pregnant and non-pregnant women 
was somewhat better in UP than AP in this evaluation. In contrast to the AP findings, mean BMI 
among first and second trimester pregnant women did not change much from baseline to endline 
in the intervention district (Table 5.21). By contrast, the mean BMI of third trimester pregnant 
women in the intervention district increased significantly by 0.9 from baseline to endline relative 
to no change in the comparison district. However, it should be noted that BMI of third trimester 
2/20/2008 62 
pregnant women was much higher in the comparison than the intervention district at baseline 
(22.4 vs. 21.0). 
Table 5.21. Mean Body Mass Index among pregnant women stratified by trimester in Uttar Pradesh 
Mean BMI among pregnant women (Kg/m2
) 
Intervention (Barabanki) Comparison (Unnao) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 20.0 19.9 -0.1 19.8 20.4 0.6 -0.7 
Second 
trimester 
21.0 21.1 0.1 21.4 21.3 -0.1 0.2 
Third trimester 21.0 21.9 0.9 22.4 22.4 0 0.9* 
All 20.8 21.1 0.3 21.2 21.4 0.2 0.1 
Total # of 
pregnant 
women 
481 653 605 604 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Mothers of children 6-23 months were the thinnest group of women in UP districts as was found 
in AP as well (Table 5.22). There were small increases in mean BMI at endline among non￾pregnant women in both UP districts, but no intervention effect. 
Table 5.22. Mean Body Mass Index among non-pregnant women in Uttar Pradesh 
Mean BMI among non-pregnant women (Kg/m2
) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 
children 0-5 
months 
19.7 20.0 0.3 19.8 20.4 0.6 -0.3 
Mothers of 
children 6-23 
months 
19.1 19.3 0.2 19.4 19.7 0.3 -0.1 
Total # of non￾pregnant 
women 
1619 2074 1953 1941 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The malnutrition prevalence among all pregnant women decreased slightly from baseline to 
endline in both districts, but this overall finding masked the variation observed in the trimester￾specific data (Table 5.23). Among first trimester women, malnutrition increased slightly in the 
intervention district relative to a large decrease in the comparison district. However, the relative 
difference was not statistically significant. By contrast, malnutrition among third trimester 
women was reduced by more than half at endline in the intervention district compared to a small 
increase in the comparison district. This relative difference was statistically significant. 
2/20/2008 63 
Table 5.23. Malnutrition (BMI<18.5 Kg/m2
) among pregnant women stratified by trimester in Uttar Pradesh 
BMI<18.5 Kg/m2
 among pregnant women (%) 
Intervention (Barabanki) Comparison (Unnao) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 26.1 29.3 3.2 32 21.9 -8.1 10.9 
Second 
trimester 
14.2 13.2 -1.0 8.3 10.4 2.1 -3.1 
Third trimester 12.6 5.1 -7.5 2.1 3.6 1.5 -9.0* 
All women 16.4 14.1 -2.3 13.2 11.9 -1.3 -1.0 
Total # of 
pregnant 
women 
481 653 605 604 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Malnutrition was 30% and 38-46% among mothers of children 0-5 and 6-23 months, respectively 
in both districts at baseline (Table 5.24). Malnutrition prevalence decreased at endline among all 
non-pregnant women, but the decrease was smaller in the intervention relative to the comparison 
district with no significant relative differences. 
Table 5.24. Malnutrition (BMI<18.5 Kg/m2
) among non-pregnant women in Uttar Pradesh 
BMI<18.5 Kg/m2
 among non-pregnant women (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Mothers of 0-5 
month olds 
30.5 26.3 -4.2 29.8 20.7 -8.9 4.7 
Mothers of 6-23 
month olds 
45.5 43.0 -2.5 37.6 30.7 -6.9 4.4 
Total non￾pregnant 
women 
1619 2074 1953 1941 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The positive intervention effect on women’s nutritional status observed late in pregnancy in UP 
state might reflect increased gestational weight gain as a result of increased food intake reported 
in the Dietary Intake section of this chapter (Table 5.14-16). However, this hypothesis is not 
supported by the fact that improvements in dietary intake were observed among non-pregnant 
women as well, but were not reflected in improved nutritional status at endline. 
Pregnancy Anemia
Andhra Pradesh. Mean hemoglobin (Hb) at baseline was 10.6 and 10.4 g/dL among pregnant 
women in the intervention and comparison districts, respectively, and changed by 0.1 g/dL only 
at endline (data not shown). Anemia prevalence among pregnant women (hemoglobin <11.0 
g/dL) was ~60% in both districts at baseline (Table 5.25). As expected, prevalence was higher in 
the second than in the first or third trimesters because of normal physiologic changes during 
pregnancy associated with hemodilution. There were small increases and decreases in anemia 
prevalence from baseline to endline in both districts, but none of the relative differences were 
statistically significant
2/20/2008 64 
Table 5.25. Anemia among pregnant women stratified by trimester in Andhra Pradesh 
Hemoglobin <11 g/dL among pregnant women (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 58.5 58.4 -0.1 49.0 47.0 -2 1.9 
Second 
trimester 
59.8 63.4 3.6 70.4 66.0 -4.4 8.0 
Third trimester 55.3 53.1 -2.2 64.3 64.2 -0.1 -2.1 
All women 58.2 60.1 1.9 61.5 61.0 -0.5 2.4 
Total # pregnant 
women 
567 554 605 676 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Moderate-severe anemia (hemoglobin<10.0 g/dL) prevalence at baseline was 34% and 41% in 
the intervention and comparison districts, respectively, and was virtually unchanged at endline 
(Table 5.26). The trend across trimesters matched that of overall anemia prevalence. Among 
women in the first and second trimesters, prevalence was increased slightly in the intervention 
district and decreased slightly in the comparison district. There was an 11 percentage point drop 
in prevalence at endline among third trimester women in the intervention relative to the 
comparison district, but the relative difference was not statistically significant. Severe anemia 
prevalence (<7.0 g/dL) was 1-2% at baseline and increased to 3-4% at endline, but did not differ 
between districts (data not shown). 
Table 5.26. Moderate-severe anemia among pregnant women stratified by trimester in Andhra Pradesh 
Hemoglobin <10 g/dL among pregnant women (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 32.2 32.9 0.7 31.3 26.5 -4.8 5.5 
Second 
trimester 
34.8 37.6 2.8 46.5 45.3 -1.2 4.0 
Third trimester 36.2 25.5 -10.7 44.2 44.2 0 -10.7 
All women 34.4 34.1 -0.3 40.7 40.5 -0.2 -0.1 
Total # pregnant 
women 
567 554 605 676 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Considering the low coverage of antenatal IFA supplementation (26% received 90+ 
supplements) and low levels of supplement intake (18%) in the AP intervention district, the lack 
of change in anemia prevalence among pregnant women reported here is not surprising. 
Uttar Pradesh. Mean hemoglobin at baseline was 10.8 and 10.7 g/dL among pregnant women in 
the intervention and comparison districts, respectively, and it did not change significantly at 
endline (data not shown). In contrast to AP state, hemoglobin decreased across the trimesters of 
pregnancy. Overall anemia prevalence was 56% in the intervention district and 52% in the 
comparison district (Table 5.27). At endline, anemia prevalence decreased in all trimesters in the 
intervention district relative to increases in the comparison district, but the relative differences 
2/20/2008 65 
were not statistically significant. The difference was largest among third trimester pregnant 
women (-11.1%). 
Table 5.27. Anemia among pregnant women stratified by trimester in Uttar Pradesh 
Hemoglobin <11 g/dL among pregnant women (%) 
Intervention (Barabanki) Comparison (Unnao) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 40.0 37.2 -2.8 40.4 42.6 2.2 -5.0 
Second 
trimester 
60.8 58.3 -2.5 53.4 59.3 5.9 -8.4 
Third trimester 61.6 56.8 -4.8 61.4 67.7 6.3 -11.1 
All women 56.1 53.2 -2.9 52.2 56.7 4.5 -7.4 
Total # pregnant 
women 
608 619 607 593 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
More than half of all anemia was moderate-severe anemia in the UP districts (Table 5.28). 
Baseline prevalence was 31-2% and it did not change demonstrably at endline in either district. 
The same trend of increasing prevalence across the 3 trimesters of pregnancy was observed. 
Although overall anemia prevalence among pregnant women was lower in UP than in AP, the 
severity of anemia among the most anemic women was worse. Severe anemia prevalence was 
4% at baseline and decreased by 0.5 to 1 percentage point only at endline in both districts. 
Table 5.28. Moderate-severe anemia among pregnant women stratified by trimester in Uttar Pradesh 
Hemoglobin <10 g/dL among pregnant women (%) 
Intervention (Barabanki) Comparison (Unnao) 
Trimester of 
pregnancy 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
First trimester 22.1 21.9 -0.2 22.8 21.6 -1.2 1.0 
Second 
trimester 
32.2 36.6 4.4 32.8 31.6 -1.2 5.6 
Third trimester 36.8 37.9 1.1 37.6 39.6 2.0 -0.9 
All women 31.3 33.8 2.5 31.5 30.9 -0.6 3.1 
Total # pregnant 
women 
608 619 607 593 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Summary
The INHP-II program was very successful at improving maternal nutrition from numerous 
perspectives. The program’s extensive efforts to support ICDS services in the transition from 
Title II commodities to locally available foods resulted in improved coverage of the 
supplementary feeding program and increased consumption of the SN by pregnant and 
postpartum women in both AP and UP states. Approximately half of women reported coverage 
during pregnancy or the early postpartum period at endline. INHP-II’s support for the SN 
program component of ICDS appeared early in the evaluation period, with the majority of 
improvements in coverage occurring during 2004. 
2/20/2008 66 
Maternal nutrition messages delivered by service providers to pregnant and postpartum women 
were effective at increasing consumption of at least 3 meals a day in both states. In AP, the vast 
majority of women already reported consumption of 3 or more meals a day, but the intervention 
was able to increase this proportion even more. However, snack consumption was not common 
and dropped significantly at endline in the intervention district. It may be that a snack was 
converted to an additional meal by the small proportion of women who were not already 
consuming 3 meals a day. In UP, only one-quarter of women reported eating at least 3 meals a 
day, but the intervention was effective at increasing this proportion by endline. Snack 
consumption was frequent already in UP (50-60%), but increased nonetheless at endline among 
postpartum women. 
Advice to increase the quantity and diversity of dietary intake during pregnancy and the early 
postpartum period led to improvements in diet in the intervention district of both states. Pregnant 
and postpartum women from the AP intervention district increased consumption of legumes, 
DGLV, and yellow-orange fruits at endline. Supplementary Nutrition was consumed more 
frequently among pregnant women only. In UP, legume intake was very frequent already and 
consumption of yellow-orange fruits and SN increased significantly at endline. Pregnant women 
and mothers of children 6-23 months significantly increased their weekly intake of meat, fish, 
chicken, and eggs while women who had recently delivered increased their dairy intake. In 
general, maternal diet advice emphasized in the INHP-II program resulted in widespread dietary 
improvements among women during pregnancy and the early postpartum period as well as 
mothers with young children. 
In light of the improvements in maternal diet achieved by the intervention, it is somewhat 
difficult to make sense of the findings on women’s nutritional status. Overall, the prevalence of 
malnutrition among women of reproductive age was astoundingly high, particularly in AP state 
where it was over 50% in the intervention district. The ~1 unit increase in BMI and decrease in 
malnutrition prevalence documented among first trimester pregnant women in AP and third 
trimester pregnant women in UP over a 2 year evaluation period is an exciting finding. 
Malnutrition prevalence decreased and BMI increased significantly among mothers of children 
6-23 months of age in the AP intervention district as well, but no comparable improvement was 
observed among this same group of women in UP. The increased use of SN and improvements in 
dietary intake during pregnancy and the postpartum may help explain these impressive results. 
However, given that improvements in dietary intake were observed among pregnant and 
postpartum women in both states, more widespread improvements in nutritional status would be 
expected for all women. Additional factors that could explain these discrepant results include 
provision of better quality foods or larger portion sizes to pregnant women, and reductions in 
workload during pregnancy. 
The prevalence of anemia among currently pregnant women was higher in AP (60%) than in UP 
(53%). These proportions are slightly higher than those reported in NFHS-2 (just under 50%). 
The intervention had no significant impact on mean hemoglobin or the prevalence of anemia 
among pregnant women in either state. The prevalence of moderate-severe anemia in late 
pregnancy had a relative decrease of 11% in both states, but this difference was not statistically 
significant. Any program benefits for pregnancy anemia would be most easily identified among 
women late in pregnancy after IFA supplements had been consumed for a number of months. In 
2/20/2008 67 
AP, given the continued low coverage and consumption of IFA supplements distributed to 
women as part of ANC visits, the lack of impact on anemia is not surprising. However, a 
reduction in anemia might be expected in the UP intervention district where 65% of women 
received and 36% consumed the IFA supplement program target of 90+ tablets at endline. 
2/20/2008 68 
CHAPTER 6: INFANT FEEDING 
Infant feeding is a crucial aspect of young child nutrition. Identifying and promoting optimal 
practices during this period may have considerable impact on improving the health and 
nutritional status of children ages 0-23 months. Data were collected at each survey round to 
generate key indicators of infant and young child feeding practices important for improved child 
nutrition. These indicators include initiation of and exclusive breastfeeding, timely introduction 
of complementary foods, and the content, quantity, and quality of children’s diets. Details of the 
INHP-II infant and child feeding messages conveyed to mothers as part of the intervention are 
given in Appendix I. 
Breastfeeding Initiation
Early initiation of breastfeeding
Significant improvement in initiation of breastfeeding within 1 hour of delivery was observed in 
the intervention relative to the comparison districts in both states (Table 6.1). However, the 
intervention effect was much more dramatic in UP districts with a relative increase of 46 
percentage points in the proportion of mothers reporting timely breastfeeding initiation. By 
endline, almost twice as many UP mothers as AP mothers from the intervention districts reported 
timely breastfeeding initiation. 
Table 6.1. Breastfeeding initiation among mothers of infants 0-5 months of age 
% Breastfeeding initiated ≤ 1 hour after delivery (n) 
Intervention Comparison 
State 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Andhra Pradesh 22.3 
(614) 
36.5 
(545) 
14.2 16.9 
(591) 
18.9 
(610) 
2.0 12.2* 
Uttar Pradesh 4.6 
(789) 
59.0 
(741) 
54.4 2.1 
(711) 
10.5 
(658) 
8.4 46.0* 
* Statistical test result comparing change from baseline to endline in intervention and comparison districts, 
significant at p<0.05
In UP state, approximately three-quarters of women in the endline sample delivered their infants 
at home. The dramatic improvement in early initiation of breastfeeding observed among 
intervention district women (Table 6.1) was comparable among women with home and 
institutional deliveries. However, the situation differed in AP state where only 20-30% of 
deliveries took place at home. Although the intervention was effective at increasing the 
proportion of women reporting early initiation of breastfeeding for both types of deliveries, the 
overall proportion was higher among women delivering at home (Figure 6.1). 
2/20/2008 69 
Figure 6.1. Breastfeeding initiation within 1 hour of delivery among mothers of infants 0-5 months in Andhra 
Pradesh by place of delivery 
36
17
56
17 18 17
32
20
0
10
20
30
40
50
60
70
80
90
100
AP intervention AP comparison
BF initiated within 1 h after delivery (%)
BL - home deliveries
EL - home deliveries
BL - inst. deliveries
EL - inst. deliveries
Figure 6.2 illustrates the pace of change over the evaluation period of this key indicator of early 
breastfeeding initiation. In the AP intervention district, changes in breastfeeding initiation 
practice did not take place until very late in the intervention period in late 2005 as little increase 
in this indicator was seen during the 2 adequacy surveys in early and mid-2005. By contrast, 
improvements in this practice appeared earlier in the UP intervention district, increasing 
progressively at each survey round. Nonetheless, the most dramatic increase was observed at the 
end of the intervention period in late 2005 when the proportion more than doubled from 
Adequacy II to endline. 
Figure 6.2. Breastfeeding initiation within 1 hour of delivery among mothers of infants 0-5 months across all 
survey rounds 
22
17
5 2
24
10
17
2
24
11
23
4
36
17
59
11
0
10
20
30
40
50
60
70
80
90
100
AP intervention AP comparison UP intervention UP comparison
BF initiated within 1 h of delivery (%)
Baseline
Adequacy I
Adequacy II
Endline
2/20/2008 70 
Prelacteal feeds
The proportion of mothers reporting prelacteal feeding of their newborn infants followed the 
same trend observed with timely initiation of breastfeeding (Table 6.2). The occurrence of this 
risky practice decreased significantly in the intervention relative to the comparison district in 
both states, but the intervention effect was much stronger in UP state. In the AP intervention 
district, only half of women reported this practice at endline compared to approximately two￾thirds at baseline. In UP, virtually all mothers reported giving prelacteal feeds at baseline, but 
fewer than 50% of mothers reported following this practice at endline. 
Table 6.2. Prelacteal feeds of newborn infants among mothers of infants 0-5 months 
% Prelacteal feeds (n) 
Intervention Comparison 
State 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Andhra Pradesh 61.7 
(614) 
49.2 
(545) 
-12.5 69.0 
(591) 
67.6 
(611) 
-1.4 -11.1* 
Uttar Pradesh 91.8 
(789) 
44.4 
(750) 
-47.4 96.1 
(711) 
93.2 
(661) 
-2.9 -44.5* 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
As with early initiation of breastfeeding, the data on prelacteal feeds were disaggregated by place 
of delivery in AP state because of the high proportion of institutional births there. The practice of 
giving prelacteal feeds to newborn infants was less common among mothers who delivered at 
home than in institutions in the intervention but not the comparison district (Figure 6.3). The 
INHP-II program was effective at reducing this practice among both groups of women in the 
intervention district, but the reduction was greater among those who delivered at home. 
Figure 6.3. Prelacteal feeds of newborn infants among mothers of infants 0-5 months in Andhra Pradesh by 
place of delivery 
45
73
28
75
68 66
55
64
0
10
20
30
40
50
60
70
80
90
100
AP intervention AP comparison
Prelacteal feeding (%)
BL - home deliveries
EL - home deliveries
BL - inst. deliveries
EL - inst. deliveries
2/20/2008 71 
The pace of change in the prelacteal feeds indicator across the 2-year evaluation period mirrored 
that of early initiation of breastfeeding (Figure 6.4). Prelacteal feeding remained at 
approximately the same level and did not decrease until late 2005 in the AP intervention district. 
By contrast, prelacteal feeding, almost universal at baseline, declined progressively across the 
evaluation period to less than 50% at endline in the UP intervention district. 
Figure 6.4. Prelacteal feeds of newborn infants among mothers of infants 0-5 months across all survey rounds 
62
69
92
96
64 63
81
99
66
71 71
97
49
68
44
93
0
10
20
30
40
50
60
70
80
90
100
AP intervention AP comparison UP intervention UP comparison
Prelacteal feeding (%)
Baseline
Adequacy I
Adequacy II
Endline
The INHP-II program messages about newborn feeding practices were to initiate breastfeeding 
within 1 hour of delivery and avoid giving prelacteal feeds. These breastfeeding messages were 
given to mothers by service providers during contacts in late pregnancy and at the time of 
delivery. As was reported in Chapter 3, the program was successful in increasing antenatal 
contacts for pregnancy advice with both ANM and AWW in both states. Recently delivered 
mothers (of children 0-5 months) were more likely to report having received breastfeeding 
advice during pregnancy from service providers in the intervention than comparison districts in 
AP and UP states. However, the proportion of women receiving any breastfeeding advice during 
pregnancy home visits was still quite low overall (AP: 9-11%; UP: 14-20%). Furthermore, the 
presence of ANM and AWW at home deliveries remained low throughout the intervention 
period. Therefore, the dramatic improvements in these two newborn feeding practice indicators, 
particularly in UP state, may not be explained completely by advice given during these increased 
service provider contacts. The larger effect in UP state may also be explained in part by the 
simultaneous implementation of the newborn care intervention developed as part of INHP-II to 
reduce neonatal mortality and the Newborn Care Evaluation Research study in the same district. 
Exclusive breastfeeding and early introduction of complementary foods
Mothers are advised to breastfeed their infants exclusively for the first 6 months of life. The 
ICDS program emphasizes this message during service provider contacts with pregnant and 
lactating women. The INHP-II intervention was expected to increase these contacts and increase 
the emphasis on infant feeding and other nutrition messages to improve exclusive breastfeeding 
in the first 6 months of life. 
2/20/2008 72 
Extensive data on infant feeding practices in the first 6 months of life were collected from 
recently delivered mothers at all survey rounds, including feeding of breastmilk, and 
complementary liquids and solids in the past 24 hours and since birth. Current infant feeding 
status (exclusive breastfeeding and early introduction of complementary foods) was defined as 
intake in the previous 24 hours for the purpose of this evaluation to make the indicators used 
comparable to NFHS data and those used by CARE in their own data collection. 
Andhra Pradesh. Over 75% of infants 0-5 months were exclusively breastfed at baseline in both 
districts (Table 6.3). The prevalence was close to 90% at 0-2 months, but dropped somewhat 
among infants 3-5 months. At endline, exclusive breastfeeding decreased in both AP districts, 
but more so in the comparison district, particularly at 3-5 months of age (Figure 6.5). However, 
the relative difference in these decreases in exclusive breastfeeding was not statistically 
significant. At endline, 71% of infants 0-5 months were breastfed exclusively in the intervention 
district. 
Figure 6.5. Exclusive breastfeeding among children 0-11 months of age in Andhra Pradesh
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10 11
Months
Exclusive breastfeeding (%)
AP intervention BL
AP intervention EL
AP comparison BL
AP comparison EL
Table 6.3. Exclusive breastfeeding among infants 0-5 months of age in Andhra Pradesh 
Exclusive breastfeeding (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Infant age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0-2 months 87.0 87.4 0.4 91.0 89.3 -1.7 2.1 
3-5 months 68.4 61.1 -7.3 84.6 69.8 -14.8 7.5 
0-5 months 77.0 70.7 -6.3 86.8 77.0 -9.8 3.5 
Total # of 
infants 0-5 mo 
614 549 591 616 
Early introduction of complementary foods was reported among <15% of infants 0-5 months of 
age at baseline in the AP districts, but the problem was more common among infants 3-5 months 
2/20/2008 73 
(Table 6.4). A significant decrease in the proportion of mothers reporting this practice at endline 
occurred in the intervention relative to the comparison district. The decrease was only 4 
percentage points in the intervention district, but it represented a ~30% decrease because of the 
low baseline prevalence. 
Table 6.4. Early introduction of complementary foods and liquids among infants 0-5 months of age in 
Andhra Pradesh 
Early introduction of complementary foods and liquids (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0-2 months 8.1 4.0 -4.1 4 4.9 0.9 -5.0 
3-5 months 21.6 14.9 -6.7 11.3 16.1 4.8 -11.5* 
0-5 months 15.3 10.9 -4.4 8.8 12.0 3.2 -7.6* 
Total # of 
infants 0-5 mo 
614 549 591 616 
Trends in the provision of breastmilk and water were examined to further investigate the decline 
in exclusive breastfeeding in both AP districts. Fewer than 4% of mothers at baseline reported 
feeding breastmilk and water only to their infants 0-5 months of age in both AP districts. 
However, this practice was much more common at endline, increasing by 16 and 8 percentage 
points in the intervention and comparison districts, respectively (difference of differences 
between intervention and comparison district = 4.7%, p<0.05). It is not known what accounts for 
this dramatic increase in the practice of giving young infants water along with breastmilk as one 
of the key infant feeding messages promoted by the ICDS and INHP-II programs is to feed 
infants breastmilk only for the first 6 months of life. It may be that mothers in AP perceive that 
their infants need additional fluids because of the hot weather and address this need by providing 
supplemental water. The messages they receive about exclusive breastfeeding may not be 
sufficiently explicit about the exclusion of even water from the diet of young infants. CARE may 
want to investigate this issue further with local ICDS officials. Information provided by service 
providers to mothers about exclusive breastfeeding could be modified to more strongly 
emphasize the exclusion of all other liquids and solids from the diet during the first 6 months of 
life. 
Uttar Pradesh. Approximately two-thirds of infants 0-5 months of age were breastfed 
exclusively in the UP districts at baseline (Table 6.5). Exclusive breastfeeding prevalence 
declined steadily across the first 6 months of life because of early introduction of complementary 
foods, particularly from 3 to 5 months of age (Figure 6.6). Unlike in the AP districts, exclusive 
breastfeeding increased from baseline to endline in both UP districts. The magnitude of the 
increase was twice as large in the intervention compared to the comparison district and this 
relative difference was statistically significant. Almost 80% of intervention district infants 0-5 
months were exclusively breastfed at endline. 
2/20/2008 74 
Figure 6.6. Exclusive breastfeeding among children 0-11 months of age in Uttar Pradesh
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10 11
Months
Exclusive breastfeeding (%)
UP intervention BL
UP intervention EL
UP comparison BL
UP comparison EL
Table 6.5. Exclusive breastfeeding among infants 0-5 months of age in Uttar Pradesh 
Exclusive breastfeeding (%) 
Intervention (Barabanki) Comparison (Unnao) 
Infant age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0-2 months 71.9 84.5 12.6 69.6 74.7 5.1 7.5 
3-5 months 55.0 73.1 18.1 57.5 67.4 9.9 8.2 
0-5 months 63.5 78.5 15.0 62.8 70.7 7.9 7.1* 
Total # of 
infants 0-5 mo 
789 758 712 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Early introduction of complementary foods before 6 months of age was much more common in 
UP than AP state at baseline, and was reported by approximately 30% of mothers in the UP 
districts (Table 6.6). The proportion of mothers reporting this practice decreased in both UP 
districts by endline, but the decrease of ~50% was statistically significantly greater in the 
intervention relative to the comparison district. 
Table 6.6. Early introduction of complementary foods and liquids among infants 0-5 months of age in Uttar 
Pradesh 
Early introduction of complementary foods and liquids (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0-2 months 23.4 12.1 -11.3 25.0 19.2 -5.8 -5.5 
3-5 months 39.4 21.1 -18.3 37.3 27.8 -9.5 -8.8 
0-5 months 31.3 16.9 -14.4 31.9 23.9 -8.0 -6.4* 
Total # of 
infants 0-5 mo 
789 758 712 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 75 
The pace of change of these two key infant feeding indicators was not consistent across the 2-
year evaluation period in the UP intervention district. Exclusive breastfeeding prevalence did not 
change demonstrably for the first 1 ½ years of the evaluation period, increasing significantly at 
the endline survey only (Figure 6.7). By contrast, the proportion of children given 
complementary foods and liquids before 6 months of age had decreased significantly by the 
Adequacy II survey in mid-2005. Given that these two early infant feeding messages were both 
supposed to be given to mothers of young infants, it is not clear why improvement was observed 
in one indicator before the other. It should be noted that these two indicators followed the same 
trend in the UP comparison district, suggesting that the infant feeding messages are being 
received by mothers through the regular ICDS services. It may be that success of the INHP-II 
program at increasing service provider contacts with women at home and elsewhere in 
communities has further increased exposure to these messages and/or the likelihood that mothers 
will act on this information to change their infant feeding behaviors. 
Figure 6.7. Exclusive breastfeeding and early introduction of complementary foods and liquids among 
children 0-5 months of age in Uttar Pradesh across all survey rounds
64
31
59
35
66
17
79
17
0
10
20
30
40
50
60
70
80
90
100
Exclusive breastfeeding Early complementary feeding
%
Baseline
Adequacy I
Adequacy II
Endline
Complementary Feeding
From six months of age, appropriate complementary feeding is recommended with continued 
breastfeeding up to 24 months. This involves important age-specific guidelines on food 
frequency, food quality (including energy and micronutrient levels), feeding during illness, 
hygienic practices, and methods of interaction and feeding the young child. 
Introduction of solids
Mothers are advised to introduce solids to their infants starting at 6 months of age. Both early 
and delayed introduction of solids were observed in AP and UP states. To assess the effect of the 
intervention on the delayed introduction of complementary foods, two types of indicators were 
constructed from the evaluation data. First, all mothers of children 6-23 months of age were 
asked to recall when they first introduced solids to their children to identify the proportion who 
did so at 6-8 months of age. This indicator is comparable to the one constructed from CARE’s 
RAPS data. The results for this indicator are reported below in Table 6.7. 
2/20/2008 76 
In the intervention districts of both states, <30% of mothers at baseline introduced solids to their 
infants at 6-8 months, increasing to ~50% at endline. In AP, this increase was statistically 
significant relative to the more modest increase in the comparison district. In UP, the increases 
observed in both districts were very large (30 percentage points, a ~100% increase), but 
comparable, yielding no significant improvement in the intervention relative to the comparison 
district. 
Table 6.7. Reported introduction of solids at 6-8 months among mothers of children 6-23 months of age 
Introduction of solids at 6-8 months among mothers of 6-23 months olds (%) 
Intervention Comparison 
State 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Andhra Pradesh 30.4 
(1746) 
49.5 
(2012) 
19.1 24.6 
(1831) 
35.7 
(2034) 
11.1 8.0* 
Uttar Pradesh 22.1 
(1643) 
52.8 
(1650) 
30.7 26.8 
(1666) 
56.5 
(1624) 
29.7 1.0 
* Statistical test result comparing change from baseline to endline in intervention and comparison districts, 
significant at p<0.05
The timing of introduction of solids can be assessed also through current dietary intake. This 
alternative approach avoids the problem of recall bias that can occur when persons have to recall 
an event that has occurred in the distant past. Dietary intake from 24-hour diet recall data can be 
used to determine the proportion of children 6-8 months of age who consumed any solids in the 
previous 24 hours. This indicator provides another measure of timely introduction of solids. 
These data are presented in the next section on dietary intake. 
Dietary intake
Caregivers were asked what kinds of foods their child consumed in the previous 24 hours. 
Portion size information also was ascertained. These data were used to describe the content of 
children’s diets as well as the quantity, quality, and frequency of feeds. 
Andhra Pradesh. At baseline, cereal intake at all ages was much higher in the intervention than 
in the comparison district (Table 6.8a-c). The reason for this difference in the intake of solids by 
young children is not known. Cereal intake among children 6-11 months increased at endline by 
approximately the same amount in both districts. Among children 12-23 months, cereal intake 
increased significantly less in the intervention than in the comparison district, but this may be 
explained by the fact that baseline intake was already high (76%) and increased almost as much 
as it could to close to 100%. Since cereal intake in the comparison district at baseline was 26 
percentage points lower, there was more room for improvement at endline (by 36 percentage 
points to 86%). Despite the fact that the evaluation analyses did not yield a positive intervention 
effect on cereal intake, it should be emphasized that cereal intake among children 6-23 months 
increased fairly consistently during the 2-year evaluation period to very high levels in the AP 
intervention district (Figure 6.8). 
2/20/2008 77 
Table 6.8. Foods consumption in the past 24 hours among breastfeeding children 6-23 months of age in 
Andhra Pradesh 
6.8a. 6-8 months 
Food consumed in the past 24 hours (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 56.8 70.6 13.8 36.3 48.0 11.7 2.1 
Legumes 2.3 4.6 2.3 2.8 1.0 -1.8 4.1* 
All fruits and 
vegetables 
12.2 9.5 -2.7 26.3 5.7 -20.6 17.9* 
Eggs, meat, 
chicken and fish 
2.9 0.6 -2.3 5.9 0.3 -5.6 3.3* 
Total # of 
children 6-8 
months 
345 347 388 402 
6.8b. 9-11 months 
Food consumed in the past 24 hours (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 67.4 85.7 18.3 44.1 66.9 22.8 -4.5 
Legumes 7.1 7.8 0.7 5.7 2.4 -3.3 4.0 
All fruits and 
vegetables 
17.7 15.3 -4.5 30.8 16.4 -14.4 9.9* 
Eggs, meat, 
chicken and fish 
4.8 4.2 -0.6 7.6 1.2 -6.4 5.8* 
Total # of 
children 9-11 
months 
294 307 263 335 
6.8c. 12-23 months 
Food consumed in the past 24 hours (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 76.3 94.3 18.0 49.7 87.7 38.0 -20.0* 
Legumes 13.2 13.4 0.2 7.3 9.4 2.1 -1.9 
All fruits and 
vegetables 
23.7 21.6 -2.1 30 21.5 -8.5 6.4* 
Eggs, meat, 
chicken and fish 
4.6 4.4 -0.2 3.9 2.3 -1.6 1.4 
Total # of 
children 12-23 
months 
990 1240 1060 1166 
*Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 78 
Figure 6.8. Cereal consumption during the past 24 hours among children 6-23 months in Andhra Pradesh 
intervention district (Karimnagar) 
57
67
76
49
83
93
68
74
90
71
86
94
0
10
20
30
40
50
60
70
80
90
100
6-8 mo 9-11 mo 12-23 mo
Cereal consumption in the past 24 h (%)
Baseline
Adequacy I
Adequacy II
Endline
In contrast to cereal intake, fruit and vegetable intake was lower in the intervention (12-24%) 
than the comparison (26-31%) district at baseline (Table 6.8a-c). Unfortunately, fruit and 
vegetable intake decreased among all ages at endline, but the decreases were small in the 
intervention district and much larger in the comparison district, yielding a statistically significant 
relative difference between the 2 districts. Consumption of eggs and meat was low (<8%) in all 
age groups at baseline, and decreased in all but one group (children 12-23 months in the 
intervention district) at endline. However, the decreases followed the same trend as fruit and 
vegetable intake with small decreases in the intervention district and larger decreases in the 
comparison district yielding statistically significant relative differences. 
Legume intake was low at baseline (2-13% across all age groups), and remained low at endline 
in both AP districts (Table 6.8). A review of 7-day food frequency data on legume consumption 
confirms the low daily intake, but indicates an increase in the proportion of children consuming 
legumes 2-6 times in the past week in the intervention relative to the comparison district (Figure 
6.9). It should be noted that idli and dosa, two cereal-based foods commonly consumed in AP 
state, are a mixture of cereal grains and lentils (a legume). They were classified as “cereals” in 
these data, and this may have contributed somewhat to the low consumption levels of legumes 
reported. Therefore, if increased legume consumption in the AP intervention district involved 
increased intake of idli and dosa only, then this intervention effect may have been missed in 
these analyses. However, one of the key feeding messages given for introducing and feeding 
solids at 6-8 months of age was to mash legumes with rice or bread for an energy-dense 
complementary food. The AP data suggest that this feeding practice was not widely received, 
understood or practiced in this study population. 
2/20/2008 79 
Figure 6.9. Legume consumption during the past week among children 6-23 months in Andhra Pradesh 
intervention district (Karimnagar) 
3
35
14
48
4
57
8
31
0
10
20
30
40
50
60
70
80
90
100
Daily 2-6 days/w k 1 day/w k Not eaten
Legumes eaten in the past week
(%)
Baseline
Endline
Data on liquid intake in the past 24 hours were collected as well. Although the majority of 
women breastfeed their children until 2 years of age, it is common practice to supplement 
breastmilk intake with animal or powdered milk or infant formula. In AP districts, 20-30% of 
children 6-11 months consumed animal/powdered milk or infant formula at baseline (Table 6.9). 
This practice was much less common among older children 12-23 months with only 5-7% 
reporting animal/powdered milk or infant formula intake in the past 24 hours. Among children 6-
8 months, this practice decreased by ~50% in the intervention district relative to a very small 
decrease in the comparison district. The changes were minimal in the older age groups. 
Table 6.9. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding children 6-23 
months of age in Andhra Pradesh 
Animal/powdered milk or formula consumed in the past 24 hours (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Child age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 29.6 15.0 -14.6 21.9 18.4 -3.5 -11.1* 
9-11 months 27.2 26.4 -0.8 25.1 22.1 -3.0 2.2 
12-23 months 5.5 6.9 1.4 6.6 4.2 -2.4 3.8* 
Total # of 
children 6-23 
months 
1629 1894 1711 1903 
Uttar Pradesh. Intake of solids was extraordinarily low at baseline (8-57%), particularly among 
children <1 year of age (Table 6.10a-c). Consumption of cereals and legumes increased 
dramatically at endline in both districts, but the increases were significantly greater in the 
intervention than in the comparison district for all age groups. Cereal intake in the previous 24 
hours in the intervention district at endline had improved by 32 to 54 percentage points to 44%, 
77%, and 89% among children ages 6-8, 9-11, and 12-23 months, respectively. The pace of 
improvement in cereal intake in the intervention district was fairly consistent across the 
evaluation period (Figure 6.10). Large increases were observed for children 9-23 months at the 
Adequacy I survey in early 2005. Improvements in cereal intake among children 6-8 months was 
2/20/2008 80 
more modest at that time, but had increased dramatically by the time of the Adequacy II survey 
in mid-2005. 
Table 6.10. Foods consumption in the past 24 hours among breastfeeding children 6-8 months of age in Uttar 
Pradesh 
6.10a. 6-8 months 
Food consumed in the past 24 hours (%) 
Intervention (Barabanki) Comparison (Unnao) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 7.9 44.3 36.4 11.9 35.0 23.1 13.3* 
Legumes 6.9 30.8 23.9 2.3 12.8 10.5 13.4* 
All fruits and 
vegetables 
4.0 16.9 12.9 3.2 16.3 13.1 -0.2 
Eggs, meat, 
chicken and fish 
0 0.8 0.8 0.3 0.7 0.4 0.4 
Total # of 
children 6-23 
months 
277 237 346 306 
6.10b. 9-11 months 
Food consumed in the past 24 hours (%) 
Intervention (Barabanki) Comparison (Unnao) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 23.4 77.0 53.6 37.8 73.3 35.5 18.1* 
Legumes 13.5 51.2 37.7 13.3 24.6 11.3 26.4* 
All fruits and 
vegetables 
12.0 31.6 19.6 17.4 43.1 25.7 -6.1 
Eggs, meat, 
chicken and fish 
0.5 0.5 0 0 1.0 1.0 -1.0 
Total # of 
children 6-23 
months 
192 209 196 195 
6.10c. 12-23 months 
Food consumed in the past 24 hours (%) 
Intervention (Barabanki) Comparison (Unnao) 
Food type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All cereals 56.6 88.7 32.1 71.7 89.5 17.8 13.3* 
Legumes 31.1 65.2 34.1 25.1 31.4 6.3 27.8* 
All fruits and 
vegetables 
27.7 45.8 18.1 47.4 63.7 16.3 1.8 
Eggs, meat, 
chicken and fish 
0.7 2.3 1.6 1.0 1.3 0.3 1.3 
Total # of 
children 6-23 
months 
995 1035 969 975 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 81 
Figure 6.10. Cereal consumption during the past 24 hours among children 6-23 months in Uttar Pradesh 
intervention district (Barabanki) 
8
23
57
15
53
82
42
68
82
44
77
89
0
10
20
30
40
50
60
70
80
90
100
6-8 mo 9-11 mo 12-23 mo
Cereal consumption in the past 24 h (%)
Baseline
Adequacy I
Adequacy II
Endline
Large increases of 13 to 20 percentage points were observed for fruit and vegetable intake in the 
intervention district, but they were matched or exceeded by comparable increases in the 
comparison district, so these increases cannot be attributed to the intervention (Table 6.10a-c). 
Intake of eggs and meat was virtually nonexistent at baseline and remained so at endline in both 
UP districts. 
Increased consumption of legumes in the daily diet was particularly dramatic in the UP districts 
(Table 6.10a-c). Although legume intake increased among children 6-23 months in both districts, 
the increases in the intervention district were much more pronounced, ranging from 24 to 38 
percentage points. By endline, one third of children 6-8 months, one half of children 9-11 
months, and two-thirds of children 12-23 months consumed legumes in the previous 24 hours. 
The 7-day food frequency data provide further evidence of this strong shift to more frequent 
inclusion of legumes in the diets of young children (Figure 6.11). These data suggest that the 
feeding message to mash legumes with rice or bread for an energy-dense complementary food 
was effective in this study population. 
2/20/2008 82 
Figure 6.11. Legume consumption during the past week among children 6-23 months in Uttar Pradesh 
intervention district (Barabanki) 
24
19
7
51 55
29
3
15
0
10
20
30
40
50
60
70
80
90
100
Daily 2-6 days/w k 1 day/w k Not eaten
Legumes eaten in the past week
(%)
Baseline
Endline
Intake of animal and powdered milk and infant formula was slightly more common in UP than 
AP with close to one-third of children 6-23 months of age reporting intake during the previous 
24 hours (Table 6.11). Unlike AP, consumption remained at the same level in the second year of 
life. Endline proportions were unchanged from baseline in all age groups in both districts, 
indicating that the practice of supplementing breastmilk with animal milk or infant formula 
intake was not influenced by the infant feeding counseling received by mothers. 
Table 6.11. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding children 6-
23 months of age in Uttar Pradesh 
Animal/powdered milk or formula consumed in the past 24 hours (%) 
Intervention (Barabanki) Comparison (Unnao) 
Child age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 33.2 32.1 -1.1 27.2 25.8 1.4 -2.5 
9-11 months 34.4 34.5 0.1 30.1 32.3 2.2 -2.1 
12-23 months 29.9 32.2 2.3 31.1 29.1 -2 4.3 
Total # of 
children 6-23 
months 
1464 1481 1511 1476 
Quantity of solids consumed
The World Health Organization (WHO) recommends a minimum of 200, 300 and 400 grams of 
solids per day for children 6-8, 9-11 and 12-23 months, respectively. Twenty-four hour recall 
data were used to calculate the quantity of solids per day consumed by children 6-23 months. 
Solids were defined as rice, bread and other grain products. Because the quantities of solids 
consumed in the study areas were very low, the indicator used for evaluation purposes was the 
proportion of children consuming at least half the recommended quantity of solids per day. 
Andhra Pradesh. The proportion of children consuming at least half the recommended quantity 
of solids at baseline was higher in the intervention than comparison district for all ages (Table 
6.12a). This proportion increased quite significantly by 12-25 percentage points at endline in all 
age groups, but the increase was comparable in the two districts, indicating that the intervention 
was not responsible for the improvement of this feeding practice. 
2/20/2008 83 
Table 6.12. Consumption of at least half the recommended quantity of solids among children 6-23 months of 
age 
6.12a. Andhra Pradesh 
Half the recommended quantity of solids (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 
(100+ g) 
26.7 43.2 16.5 16.0 27.6 11.6 4.9 
9-11 months 
(150+ g) 
18.0 36.6 18.6 8.3 23.2 13.9 4.7 
12-23 months 
(200+ g) 
20.1 44.8 24.7 10.3 35.7 25.4 0.7 
Total # of 
children 6-23 
months 
1746 2012 1831 2034 
6.12b. Uttar Pradesh 
Half the recommended quantity of solids (%) 
Intervention (Barabanki) Comparison (Unnao) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 
(100+ g) 
1.0 6.0 5.0 0.8 1.3 0.5 4.5* 
9-11 months 
(150+ g) 
0.5 6.0 5.5 2.4 1.0 -1.4 6.9* 
12-23 months 
(200+ g) 
1.5 6.9 5.4 2.1 3.2 1.1 4.3* 
Total # of 
children 6-23 
months 
1643 1650 1666 1624 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. Virtually all children in the UP districts failed to consume at least half the 
recommended quantity of solids at baseline (Table 6.12b). There was a significant increase in 
this proportion at endline for all age groups in the intervention relative to the comparison district. 
However, the endline proportion did not exceed 6% for any category, suggesting that 
consumption of an insufficient quantity of solids constitutes a major problem for complementary 
feeding interventions in UP districts. 
Frequency of solids consumed
The World Health Organization (WHO) recommends a minimum of 2 and 3 solid feeds per day 
for breastfeeding children 6-8 and 9-23 months, respectively. As with the quantity indicator, 24-
hour recall data were used to calculate the number of solid feeds per day among children 6-23 
months. A solid feed was defined as intake of rice, bread or other grain products. When children 
were reported to have consumed a given food more than once in the previous 24 hours, these 
feeds were recorded as separate episodes in the data collection instrument. 
Andhra Pradesh. The proportion of children consuming the recommended minimum frequency 
of solids at baseline followed the same trends by age and district as the quantity indicator (Table 
6.13a). The baseline proportions were higher in the intervention than comparison district and 
among the youngest children. Large increases of 15-30 percentage points were observed among 
2/20/2008 84 
all age groups at endline, but these increases were comparable between the two districts so there 
were no significant intervention effects. At endline, over 50% of children 6-8 months were 
receiving at least 2 feeds per day, but the proportion was <30% for children 9-11 months who are 
expected to consume at least 3 feeds per day to meet their nutritional needs. Mothers were more 
successful at increasing the number of feeds to 3 per day for children in the second year of life, 
reaching over 40% in the intervention district. 
Table 6.13. Minimum recommended frequency of solids consumed per day among breastfeeding children 6-
23 months of age 
6.13a. Andhra Pradesh 
Recommended complementary feeding frequency (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 
(2+ times /d) 
34.5 56.1 21.6 11.4 35.5 24.1 -3.5 
9-11 months 
(3+ times/d) 
12.9 28.3 15.4 7.6 17.3 9.7 5.7 
12-23 months 
(3+ times/d) 
15.5 42.4 26.9 9.6 40.9 31.3 -4.4 
Total # of 
mothers 
1620 1893 1702 1901 
6.13b. Uttar Pradesh 
Recommended complementary feeding frequency (%) 
Intervention (Barabanki) Comparison (Unnao) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 
(2+ times /d) 
2.3 25.4 23.1 3.2 18.2 15.0 8.1* 
9-11 months 
(3+ times/d) 
3.1 23.9 20.8 4.6 11.3 6.7 14.1* 
12-23 months 
(3+ times/d) 
8.8 37.8 29.0 15.9 31.3 15.4 13.6* 
Total # of 
mothers 
1447 1476 1445 1473 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. The proportion of children consuming the minimum recommended frequency of 
solid feeds was low at baseline (2-16%), but was higher at older ages unlike in the AP districts 
(Table 6.13b). However, the proportion of children receiving the recommended number of feeds 
increased significantly at endline in all three age groups in the intervention relative to the 
comparison district. At least 24% of children 6-11 months and almost 40% of children 12-23 
months consumed the minimum recommended frequency of solids at endline. Although these 
proportions remain too low, they represent a major improvement over the baseline and 
comparison district proportions. 
Dietary diversity
In addition to the quantity and frequency of complementary foods consumed, the quality of the 
diet is an important characteristic to consider in complementary feeding interventions. CARE’s 
child nutrition messages include recommendations to incorporate vegetables, fruits, and protein 
2/20/2008 85 
sources into the weaning diet along with staple foods. Dietary diversity was characterized by the 
number of different food groups fed to a child in the previous 24 hours. These groups include 1) 
cereals, 2) legumes, 3) meat, chicken, fish and eggs, 4) vitamin A-rich fruits and vegetables, 5) 
other fruits and vegetables, 6) dairy (cheese, curd, animal milk, infant formula), and 7) butter, 
oil, and fat. Adequate dietary diversity was defined as intake of 3 or more of these food groups in 
the 24-hour recall data. 
 
Andhra Pradesh. Few children had diverse diets at baseline, particularly those <12 months of age 
(Table 6.14a). The proportion of children 6-11 months with diverse diets increased by very little 
at endline in both districts. Among children 12-23 months of age, the proportion with diverse 
diets increased significantly less at endline in the intervention relative to the comparison district. 
Only 22% of intervention district children 12-23 months had a diverse diet at endline. 
Table 6.14. Consumption of ≥3 food groups in the past 24 hours among children 6-23 months of age 
6.14a. Andhra Pradesh 
Consumption of ≥3 food groups in the past 24 hours (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 5.1 8.2 3.1 0.8 3.9 3.1 0 
9-11 months 8.2 14.7 6.5 2.9 6.6 3.7 2.8 
12-23 months 15 22.1 7.1 4.6 16.3 11.7 -4.6* 
Total # of 
mothers 
1746 2012 1831 2034 
6.14b. Uttar Pradesh 
Consumption of ≥3 food groups in the past 24 hours (%) 
Intervention (Barabanki) Comparison (Unnao) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 2.4 22.4 20 2.8 10.9 8.1 11.9* 
9-11 months 9.4 39.2 29.8 14.4 32.9 18.5 11.3* 
12-23 months 25.6 53.4 27.8 31.1 43.7 12.6 15.2* 
Total # of 
mothers 
1643 1650 1666 1624 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Uttar Pradesh. At baseline, diverse diets were uncommon among children <12 months, but were 
reported among approximately one-quarter of children 12-23 months of age (Table 6.14b). The 
proportion of children with diverse diets increased at endline among all age groups in both 
districts, but the increases were significantly larger in the intervention (20-30 percentage points) 
relative to the comparison (8-18 percentage points) district. Over half of intervention district 
children 12-23 months had a diverse diet at endline. 
The dietary diversity proportions in AP and UP pose an interesting contrast to the other child 
feeding indicators (quantity, frequency) and the 24-hour recall data. At baseline, children in AP 
districts were more likely than those in UP districts to consume solids (cereals), fruits, 
vegetables, eggs, and meat, and to consume food in recommended quantities and frequency. 
Nonetheless, the diversity in children’s diets in AP was lower than in UP. This disparity became 
more apparent in the endline data where the success of the INHP-II intervention at improving the 
2/20/2008 86 
diversity and frequency of dietary intake in the UP intervention district resulted in a vastly 
improved picture of children’s diets at endline. However, the low dietary diversity among 
children in the AP districts remains difficult to explain. It appears that high proportions of AP 
children 6-23 months are receiving daily solid feeds of cereal grains, but very few other types of 
food are being added into their diets. The quantity and frequency of these solid feeds have 
improved in both districts, but remain inadequate for preventing malnutrition. In AP, the 
decrease in intake of other important dietary components such as fruits, vegetables, eggs, and 
meat is of concern as these are important sources of micronutrients. 
Consumption of oil
Complementary feeding messages promoted by CARE in its training of service providers 
included a recommendation to add a small amount of oil to solid feeds for children 6-23 months. 
To assess this practice in the evaluation, an indicator was created to identify children who 
reported consumption of oil and ≥1 solid feed in the previous 24 hours. 
Andhra Pradesh. Consumption of oil added to solids was <10% at baseline in the intervention 
district but only 1% in the comparison district (Table 6.15a). There were very small increases (0-
4%) in this practice at endline in both districts. 
Table 6.15. Consumption of oil added to solids among breastfeeding children 6-23 months 
6.15a. Andhra Pradesh 
Oil added to solids (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Child age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 8.9 10.1 1.2 0 3.5 3.5 -2.3 
9-11 months 5.1 7.8 2.7 1.1 4.8 3.7 -1.0 
12-23 months 3.1 5.5 2.4 0.9 3.9 3.0 -0.6 
Total # of 
mothers 
1620 1677 1702 1656 
6.15b. Uttar Pradesh 
Oil added to solids (%) 
Intervention (Barabanki) Comparison (Unnao) 
Child age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 0 4.2 4.2 0 1.0 1.0 3.2* 
9-11 months 1.7 10.1 8.4 0 5.1 5.1 3.3 
12-23 months 1.4 7.4 6.0 1.6 7.6 6.0 0 
Total # of 
mothers 
1447 1476 1445 1473 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Uttar Pradesh. There was virtually no consumption of oil added to solids at baseline in the UP 
districts (Table 6.15b). Oil intake with solids increased at endline by 1 to 8 percentage points in 
both districts, so that 4-8% of children reported oil intake in the intervention district. However, a 
significant increase in the intervention relative to the comparison district was observed among 
children 6-8 months only. 
2/20/2008 87 
It appears that the feeding message from service providers to mothers about the addition of oil to 
solid feeds for young children was not received, understood, or followed in either state. A better 
understanding of how often this advice was given and how it was understood by mothers may 
come from examination of the questions on infant and child feeding advice in the cohort data. 
Supplementary Nutrition
INHP-II supports the distribution of supplementary nutrition commodities to ICDS beneficiaries 
including children 6 months to 6 years of age. This evaluation examined receipt and 
consumption of SN among children 6-23 months of age. Unfortunately, these data were collected 
only at the Adequacy II survey in mid-2005 and the endline survey in early 2006. Therefore, it is 
most illustrative to compare the proportion of children who received and consumed SN at 
endline to determine whether the INHP-II program increased the effectiveness of this component 
of the ICDS. However, the intervention effect between the Adequacy II and endline surveys (~6 
months) is presented also to examine short-term trends in program implementation. 
Andhra Pradesh. Receipt and consumption of SN among children 6-23 months of age was 
significantly higher in the intervention relative to the comparison district at both the Adequacy II 
and endline surveys (Table 6.16). In the intervention district, almost three-quarters of children 
had ever received SN, almost half had received it at least once in the past 3 months, and more 
than one-third had consumed it at least once in the same time period. However, there were 
significant decreases in the proportion of children who received or consumed SN 6 months later 
at endline in the intervention relative to the comparison district. The cause of this decline is not 
clear as the period in question at the endline survey did not include the transition period from 
CSB to locally available foods in late 2002 when SN availability was an issue. 
Table 6.16. Receipt and consumption of Supplementary Nutrition by children 6-23 months of age in Andhra 
Pradesh 
Supplementary Nutrition receipt/consumption among children 6-23 months (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Adequacy
II 
Endline Δ Adequacy 
II 
Endline Δ
Difference 
of Δ
Ever received 
SN 
72.9 46.5 -26.4 29.6 33.3 3.7 -30.1* 
Received SN in 
the past 3 mo 
45.6 38.3 -7.3 23.8 26.3 2.5 -9.8* 
Consumed SN 
in the past 3 mo 
38.4 33.8 -4.6 17.9 21.3 3.4 -8.0* 
Total # children 
6-23 months 
380 2012 453 2034 
* Statistical test result comparing change from baseline to endline in intervention and control districts, 
significant at p<0.05 
Despite reports that the state government in Andhra Pradesh had switched its SN distribution to 
spot feeding at AWC rather than providing SN to take home, data from both survey rounds 
suggest that most SN is still distributed as take-home rations in the program. Among children 
who ever received SN, 76-80% received it as take-home rations only. 
Uttar Pradesh. As in AP state, receipt and consumption of SN among children 6-23 months was 
significantly higher in the intervention than comparison district at the Adequacy II and endline 
2/20/2008 88 
survey rounds (Table 6.17). The proportion of children who received SN (ever or in the past 3 
months) was more than twice as high in the intervention (EL: 46-50%) compared to the 
comparison (19-20%) district at endline. The proportion of children who had consumed SN in 
the past 3 months was also more than double in the intervention district at endline. Within the ~6 
month period between the Adequacy II and endline surveys, the proportion of children who had 
recently received or consumed SN increased significantly in the intervention relative to the 
comparison district. Virtually all SN distributed in the UP districts (98%) was in the form of 
take-home rations. 
Table 6.17. Receipt and consumption of Supplementary Nutrition by children 6-23 months of age in Uttar 
Pradesh 
Supplementary Nutrition receipt/consumption among children 6-23 months (%) 
Intervention (Barabanki) Comparison (Unnao) 
Adequacy
II 
Endline Δ Adequacy 
II 
Endline Δ
Difference 
of Δ
Ever received 
SN 
46.2 49.7 3.5 17.7 20.3 2.6 0.9 
Received SN in 
the past 3 mo 
32.7 45.8 13.1 13.2 19.3 6.1 7.0* 
Consumed SN 
in the past 3 mo 
18.8 32.5 13.7 7.8 12.6 4.8 8.9* 
Total # children 
6-23 months 
437 1650 408 1624 
* Statistical test result comparing change from baseline to endline in intervention and control districts, 
significant at p<0.05 
Caregiver feeding practices
Caregiving practices involving feeding are an essential component to appropriate complementary 
feeding to improve the nutritional status of young children. Such practices are intended to ensure 
that young children consume adequate amounts of the foods offered at a meal. For example, 
plate sharing with other household members may decrease the likelihood that children receive 
sufficient quantities of food. Different caregivers assisting the child to eat may also influence the 
diet. There is some evidence that children fed by their mothers are more likely to consume 
adequate nutrient intakes. 
Caregiver feeding
Andhra Pradesh. In the AP baseline data, 50-75% of children were usually fed by their mothers 
and this proportion did not vary greatly by age (Table 6.18a). Feeding by mothers increased in all 
groups and in both districts at endline. The increase in this proportion among children 9-23 
months was significantly larger in the intervention relative to the comparison district. The 
negative relative difference among children 6-8 months in the intervention district is explained 
by the higher baseline proportion in that group that leaves less room for improvement. 
2/20/2008 89 
Table 6.18. Child feeding by mothers among children 6-23 months of age 
6.18a. Andhra Pradesh 
Usually fed by mother (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 66.8 86.1 19.3 51.7 78.1 26.4 -7.1* 
9-11 months 71.2 85.4 14.2 73.9 82.2 8.3 5.9* 
12-23 months 64.3 78.8 14.5 63.5 69.0 5.5 9.0* 
Total # children 
6-23 months 
1746 2012 1831 2034 
6.18b. Uttar Pradesh 
Usually fed by mother (%) 
Intervention (Barabanki) Comparison (Unnao) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 14.0 56.8 42.8 17.1 44.9 27.8 15.0* 
9-11 months 30.7 71.4 40.7 34.6 64.8 30.2 10.5* 
12-23 months 32.7 51.7 19.0 37.6 45.1 7.5 11.5* 
Total # children 
6-23 months 
1643 1650 1666 1624 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Uttar Pradesh. Children of all ages in UP were less likely to be fed by their mothers than in AP. 
Only 14-17% of children 6-8 months and 30-38% of children 9-23 months were fed usually by 
their mothers at baseline (Table 6.18b). However, these proportions increased dramatically at 
endline, particularly among children 6-11 months, and the increases were significantly larger in 
the intervention relative to the comparison district. By endline, over 50% of intervention district 
children were reported to be fed usually by their mothers. 
Plate sharing
Andhra Pradesh. Plate sharing was relatively uncommon in AP at baseline as the majority of 
children were fed from a separate plate (Table 6.19a). There were small positive increases in the 
proportion of children eating from their own plate at endline in both districts, so there was not an 
intervention effect on this feeding practice. However, the relative difference between the 
intervention and comparison districts was negative for children 6-8 months possibly because the 
baseline proportion in the comparison district was half of that in the intervention district (61 vs 
38%). 
2/20/2008 90 
Table 6.19. Feeding from a separate plate among children 6-23 months 
6.19a. Andhra Pradesh 
Fed from a separate plate (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 61.1 67.6 6.5 38.2 52.2 14.0 -7.5* 
9-11 months 75.5 81.9 6.4 63.0 69.9 6.9 -0.5 
12-23 months 77.8 81.4 3.6 70.4 74.4 4 -0.4 
Total # children 
6-23 months 
1746 2012 1831 2034 
6.19b. Uttar Pradesh 
Fed from a separate plate (%) 
Intervention (Barabanki) Comparison (Unnao) 
Age 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
6-8 months 5.1 41.2 36.1 7.4 21.8 14.4 21.7* 
9-11 months 19.8 44.7 24.9 19.2 26.7 7.5 17.4* 
12-23 months 37.0 57.8 20.8 45.4 49.9 4.5 16.3* 
Total # children 
6-23 months 
1643 1650 1666 1624 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Uttar Pradesh. In UP districts, eating from a separate plate was quite rare among the youngest 
children at baseline, but was more common among older children, particularly those 12-23 
months of age (Table 6.19b). These low numbers most likely reflect the small proportion of 
children 6-11 months who consumed solids and other foods in the previous 24 hours at baseline 
(see Table 5.10). However, the proportion of children feeding from a separate plate increased 
dramatically at endline in the intervention relative to the comparison district. Among children 6-
11 months, approximately twice as many fed from a separate plate in the intervention (41-45%) 
relative to the comparison (22-27%) district. These findings suggest that infant and child feeding 
advice by service providers most likely included messages about caregiver feeding behaviors in 
addition to recommendations about dietary intake quantity, frequency, and quality. 
Summary 
Newborn feeding practices associated with initiation of breastfeeding improved significantly in 
both districts. The intervention effect on these practices was particularly dramatic in UP state. 
Improvements in breastfeeding practices during the first 6 months of life were also observed. 
Exclusive breastfeeding among children 0-5 months of age was >70% in both states at endline. 
However, the trends differed in the 2 states. In UP, exclusive breastfeeding increased 
significantly in the intervention relative to the comparison district, but in AP, the proportion 
decreased in both districts. Nonetheless, the intervention was modestly effective in both states at 
reducing the practice of early introduction of complementary foods and liquids before 6 months 
of age. 
Evaluation of the intervention effect on the timely introduction of complementary foods varies 
depending on the indicator used. Using the indicator based on maternal recall of child age at the 
2/20/2008 91 
time of introduction, the intervention was modestly effective in AP, but not in UP, where large 
increases in timely introduction of solids were observed in both districts. At endline, ~50% of 
mothers reported introducing solids to their infants at 6-8 months of age. By contrast, the 24-
hour recall data yielded a somewhat different picture. Among children 6-8 months of age, cereal 
intake increased by ~10 percentage points in AP districts and 22 to 36 percentage points in UP 
districts. However, the relative increase in the intervention district was significant in UP only. 
The trends in cereal intake among older children mirror those observed among the children 6-8 
months of age. In UP, the intervention was effective at increasing intake of cereals and legumes 
among all children. The increased consumption of legumes in the intervention district was 
particularly impressive. However, increased consumption of a lesser magnitude occurred in the 
comparison district as well, suggesting positive effects of the ICDS program on child feeding 
practices. In AP, cereal intake increased equally in both districts and legume intake remained 
very low among all age groups. For fruit and vegetable intake, there was no intervention effect in 
either state, but the trends by state diverged also. Fruit and vegetable intake increased in UP 
districts, but decreased in AP districts. 
CARE’s INHP-II program was effective at improving distribution and consumption of SN 
among children 6-23 months of age in both states. However, data from the Adequacy II and 
endline surveys revealed a decline in the coverage and utilization of this program component in 
the last 6 months of the evaluation period in AP intervention district. The reasons for this recent 
decline are unknown, but do not appear to include a switch from the use of THR to spot feeding. 
Complementary feeding indicators of diet quantity, quality and feeding frequency improved over 
the intervention period in all study districts. However, there were no intervention effects in AP 
whereas the increases for all 3 indicators in UP were significantly larger in the intervention 
relative to the comparison district. Surprisingly, dietary diversity was much less of a problem in 
UP than AP, where fewer than half as many children in each age group had diverse diets. 
However, it is important to note that consumption of adequate quantity and quality of 
complementary foods and adequate frequency of feeds remain major problems in both states, but 
particularly in UP. Fewer than 6% of children in UP were consuming at least half the 
recommended minimum quantity of complementary foods and <25% of children 6-11 months 
received the minimum recommended frequency of solid feeds at endline. These figures indicate 
that improving the quantity and frequency of solids consumed must be a major priority of future 
complementary feeding interventions if they hope to demonstrate any effect on nutritional status. 
Improving the micronutrient content of weaning diets also remains a concern as intake of 
micronutrient-rich foods did not improve significantly as a result of the intervention. The added 
contribution of these nutrients to children’s diets may be important for improved physical growth 
as well.
2/20/2008 92 
CHAPTER 7: MICRONUTRIENT SUPPLEMENTATION AMONG CHILDREN
Vitamin A supplementation
Micronutrient supplementation was evaluated among children 12-23 months of age at each 
survey round. Children should receive their first vitamin A dose at 9 months of age and 
subsequent doses at 6 month intervals thereafter. Receipt of at least one dose of vitamin A was 
compiled from both immunization cards (if available) and by mother’s report. 
The INHP II program was effective at increasing the vitamin A supplementation coverage in 
both states, but particularly in UP where coverage increased approximately four-fold to 69% in 
the intervention district relative to a slight decrease in the comparison district (Figure 7.1). In AP 
state, coverage was ~50% at baseline, and increased by approximately half in the intervention 
district to 79% at endline. 
Figure 7.1. Vitamin A supplementation among children 12-23 months of age
55
18
79
69
47
36
61
27
0
10
20
30
40
50
60
70
80
90
100
Andhra Pradesh Uttar Pradesh
Received vitamin A syrup (%)
Intervention baseline
Intervention endline
Comparison baseline
Comparison endline
*Difference of differences comparing change from baseline to endline in intervention and control districts was 60% 
in UP and 10% in AP, significant at p<0.05 for both states 
The effect of the intervention on vitamin A supplementation coverage was fairly consistent 
across the 2-year intervention period in both states (Figure 7.2). In AP, the coverage increased 
most dramatically in the first year and then continued to increase more modestly in the second 
year of the evaluation period. In UP, where baseline coverage was only a third of AP, the initial 
increase in coverage was slow after the first year, but shot up thereafter, doubling between early 
and mid-2005, and then increasing by another 18 percentage points by early 2006. It appears that 
the post-MTR strategy was particularly beneficial for this component of ICDS services in the UP 
intervention district. 
2/20/2008 93 
Figure 7.2. Vitamin A supplementation among children 12-23 months of age by survey round in Andhra 
Pradesh and Uttar Pradesh intervention districts
55
18
71
25
76
51
79
69
0
10
20
30
40
50
60
70
80
90
100
AP intervention UP intervention
Received vitamin A syrup (%)
Baseline
Adequacy I
Adequacy II
Endline
Pediatric iron supplementation 
The GOI provides daily pediatric iron-folic acid (IFA) supplements for children beginning at 12 
months up to 5 years of age. The pediatric IFA supplements provide a daily dose of 20 mg of 
iron and the program target is to provide all eligible children at least 100 tablets per year. These 
supplements are supposed to be distributed by the ANM and AWW in communities. 
Coverage at baseline was very low and remained so in the comparison districts at endline in both 
states (Figure 7.3). As was observed with pediatric vitamin A supplementation, the intervention 
significantly increased the proportion of children who received IFA supplements in both states. 
However, the increase was dramatic in UP (a seven-fold increase) and more modest in AP. 
Figure 7.3. Pediatric iron-folic acid (IFA) supplementation among children 12-23 months of age 
21
9
30
64
12
7
11
2
0
10
20
30
40
50
60
70
80
90
100
Andhra Pradesh Uttar Pradesh
Received IFA supplements (%)
Intervention baseline
Intervention endline
Comparison baseline
Comparison endline
*Difference of differences comparing change from baseline to endline in intervention and control districts was 60% 
in UP and 10% in AP, significant at p<0.05 for both states 
2/20/2008 94 
The pace of the intervention for increasing the coverage of pediatric IFA supplement distribution 
varied in the two states. In AP, the coverage increased very modestly across the entire evaluation 
period, increasing by ~50% from 21% at baseline in early 2004 to 30% at endline in early 2006 
(Figure 7.4). The pace of implementation of this program component was quite different in UP 
state. The coverage remained very low at 5% at the Adequacy I survey after the first year of the 
evaluation period, but increased dramatically during the year 2005, increasing six-fold to 29% by 
the Adequacy II survey and more than doubling again to 64% by the endline survey in early 
2006. 
Figure 7.4. Pediatric iron-folic acid (IFA) supplementation among children 12-23 months of age by survey 
round in Andhra Pradesh and Uttar Pradesh intervention districts 
21
9
26
5
28 30 29
64
0
10
20
30
40
50
60
70
80
90
100
AP intervention UP intervention
Received IFA supplements (%)
Baseline
Adequacy I
Adequacy II
Endline
Source of IFA supplements
Andhra Pradesh. Children in the intervention district were most likely to receive IFA 
supplements from the AWW, but the ANM and private doctors and clinics were also significant 
sources for these supplements (Figure 7.5). In the comparison district, over 90% of children 
given any IFA supplements received them from the private sector. It is not known whether the 
supplements sold by the private sector were comparable to the pediatric IFA supplements 
distributed by the government services. The difference in the pattern of service providers as a 
source for pediatric IFA supplements suggests that the INHP-II program was effective at 
stimulating the ANM and AWW to distribute the supplements to eligible children in accordance 
with GOI policy. 
2/20/2008 95 
Figure 7.5. Iron-folic acid supplementation by type of service provider among children 12-23 months of age 
who received IFA supplements in Andhra Pradesh1
21
30
47
5 9
70
00
23
0.3
0
20
40
60
80
100
ANM/government
doctor
Private doctor/clinic AWW Change Agent Chemist shop/other
Children receiving IFA supplements (%)
AP intervention
AP comparison
1
Number of children who received pediatric IFA supplements, AP intervention: n=403, AP comparison: n=142. 
Uttar Pradesh. The pattern of distribution source for pediatric IFA supplements in the 
intervention district was very distinct with three-quarters of children 12-23 months receiving 
their supplements from the AWW in their communities (Figure 7.6). Another 20% received 
supplements from the ANM. Efforts by CARE to train AWW in the community-based 
distribution of these supplements were clearly effective. Only 17 children (2% of all eligible 
children) in the comparison district received pediatric IFA supplements at endline, so it is not 
very meaningful to interpret the service provider distribution statistics in this case. 
Figure 7.6. Iron-folic acid supplementation by type of service provider among children 12-23 months of age 
who received IFA supplements in Uttar Pradesh1
21
2
76
3 1
77
0
12
0
6
0
20
40
60
80
100
ANM/government
doctor
Private
doctor/clinic
AWW Change Agent Chemist
shop/other
Children receiving IFA supplements (%)
UP intervention
UP comparison
1
Number of children who received pediatric IFA supplements, UP intervention: n=752, AP comparison: n=17. 
Receipt and consumption of IFA supplements
Children in the intervention district in both states received a greater quantity of IFA supplements 
than those in the comparison district (Table 7.1). At least 30 supplements were received by 39% 
2/20/2008 96 
and 68% of intervention district children who received any supplements in AP and UP states, 
respectively. In UP, at least 60 supplements were received by 28% of children who received any 
IFA supplements. Among mothers who reported their children 12-23 months had received any 
IFA supplements, large proportions in both AP districts (AP intervention = 29.8%, AP 
comparison=88.7%) could not estimate the number of supplements received. The reasons for this 
are unclear as a similar problem was not observed in the UP survey. 
Table 7.1. Number of IFA supplements received by children 12-23 months of age at endline
IFA supplements received (%) 
Andhra Pradesh Uttar Pradesh 
Number of 
supplements 
Intervention Comparison Intervention Comparison 
<10 3.2 4.9 4.0 29.4 
10-29 27.5 4.2 24.7 41.2 
30-59 33.1 0.7 38.0 17.7 
≥60 6.2 1.4 28.1 5.9 
Don’t know 29.8 88.7 4.8 5.9 
Total # of children 
12-23 mo who 
ever received IFA 
403 142 748 17 
The recommended dosage for pediatric IFA supplements is one supplement per day. Information 
was collected on the number of days IFA supplements were consumed. The extent of supplement 
intake was lower than the number of supplements received in all districts (Table 7.2). This could 
reflect incomplete compliance with IFA supplementation or the fact that many children received 
their supplements recently and still have a supply for consumption. Overall intake was quite low 
in the intervention district of both states. Pediatric IFA supplements were consumed for at least 
30 days by 15% and 21% of children who received supplements in the intervention district of AP 
and UP, respectively. The low total amount of iron consumed most likely explains why there was 
little decrease in the prevalence of anemia as a result of the intervention despite the large 
increase in IFA supplement coverage, particularly in UP. 
Table 7.2. Number of days IFA supplements consumed by children 12-23 months of age at endline
IFA supplements consumed (%) 
Andhra Pradesh Uttar Pradesh 
Number of days 
Intervention Comparison Intervention Comparison 
0 6.2 0.7 10.6 17.7 
<10 23.1 4.2 34.4 41.2 
10-29 25.4 2.1 25.1 23.6 
≥30 14.7 0.7 20.9 11.8 
Don’t know 30.8 92.3 8.9 5.9 
Total # of children 
12-23 mo who 
ever received IFA 
403 142 752 17 
Mothers were asked whether their children 12-23 months were currently consuming pediatric 
IFA supplements at the time of the endline survey. Table 7.3 shows that approximately one￾quarter of children receiving supplements in the AP intervention district and one half in the UP 
intervention district continue to consume IFA supplements. Whether these low levels of 
compliance are a result of inadequate supplies, poor distribution systems, or compliance issues is 
2/20/2008 97 
unknown. To have a significant impact on the high prevalence of child anemia, further 
investigation of these aspects of the pediatric IFA supplementation program is needed. 
Table 7.3. Current consumption of IFA supplements among children 12-23 months of age at endline
IFA supplements currently consumed (%) 
Andhra Pradesh Uttar Pradesh 
Number of days 
Intervention Comparison Intervention Comparison 
Yes 26.8 0.7 47.3 17.7 
No 42.4 5.6 51.7 76.5 
Don’t know 30.8 93.7 0.9 5.9 
Total # of children 
12-23 mo who 
ever received IFA 
403a
 142 a 752a
 17a
a
 Did not report current consumption of IFA supplements: AP (intervention)=31%, AP (comparison)=94%, UP 
(intervention)= 1%, UP (comparison)=6%. 
Summary
CARE’s INHP-II program was clearly effective at increasing coverage of both vitamin A and 
IFA supplementation distribution offered as part of the ICDS program. Distribution of vitamin A 
syrup to children starting at 9 months of age is coordinated with routine immunization activities 
and receipt is recorded on immunization cards given each child. Reflective of the coverage of 
immunization services in the 2 states, more than one half of children 12-23 months in the AP 
intervention district, but only one-fifth of children in the UP intervention district had received at 
least 1 dose of vitamin A syrup at baseline. The coverage increased significantly in both states 
over the 2-year evaluation period mirroring simultaneous improvements in immunization 
coverage as well. However, the increases in vitamin A supplementation in UP were particularly 
dramatic (a four-fold increase) and even exceeded the coverage of the immunization program to 
which it is linked. 
The impact of the INHP-II intervention on pediatric IFA supplementation followed the same 
trends observed with vitamin A supplementation in the two states; the coverage increased more 
modestly in the AP intervention district and more dramatically in the UP intervention district. 
However, the delayed increase in implementation of this program component, particularly in UP 
state, helps to explain the small proportion of children who received and consumed more than a 
small number of supplements. In both states, only 15-20% of children receiving any IFA 
supplements consumed at least one month’s supply. These details of program coverage help 
explain the lack of program impact on anemia among this target group (see chapter 7 results). 
Discussions with CARE staff about implementation of INHP-II support activities for the 
pediatric IFA supplementation distribution help with interpretation of the observed changes in 
the program. In general, CARE reported that their initial efforts during the evaluation period 
were more focused on established ICDS program elements, such as utilization of antenatal care 
services, promotion of breastfeeding, and immunization. Within ICDS, the pediatric IFA 
supplementation component had received little attention as evidenced by the very low coverage 
reported at baseline in early 2004. CARE reported that pediatric IFA supplies were available 
generally, but there was a lack of knowledge among service providers in both the health system 
and ICDS about the appropriate distribution system and target groups for the supplements. 
Distribution of the supplements was meant to be coordinated between the ANM and AWW. 
2/20/2008 98 
There were many anecdotal reports that service providers were aware of pediatric IFA 
distribution for older children (3-6 years) only. 
In the UP intervention district, CARE initially focused on improving coverage by discussing the 
pediatric IFA supplementation distribution with local doctors, ANM, and ICDS supervisors. It 
was envisioned that supervisors would disseminate this information to AWW. One important 
area of clarification was that the program’s target group included children starting at 1 year of 
age. After the terrible pediatric IFA supplement coverage reported from the Adequacy I survey 
in early 2005 and RACHNA’s mid-term review (MTR), CARE shifted its strategy to a sector￾strengthening approach where CARE’s district team and local NGOs met with local service 
providers to deliver information on the supplement distribution program directly at monthly PHC 
and sector meetings. At the PHC meetings, ANM were encouraged to coordinate with the AWW 
in their assigned communities to distribute IFA supplements to children 12-23 months of age. At 
ICDS sector meetings, INHP-II staff worked with AWW to list out children eligible for pediatric 
IFA supplements and to brainstorm about strategies for encouraging consumption. Individual 
AWW progress with pediatric IFA supplementation was tracked at subsequent meetings to 
identify barriers to program implementation. These recent programmatic efforts resulted in 
dramatically increased coverage by early 2006. 
In the AP intervention district, initial coverage of supplement distribution was not as low, but 
only modest progress was evident after one year into the evaluation period. The same post-MTR 
strategy shift described above for the UP intervention district took place in the AP intervention 
district as well in mid-2005. In September 2005, the District Medical Health Officer (DMHO) 
issued an order to all medical officers to review pediatric IFA and vitamin A distribution with 
ANM at monthly PHC meetings. Although some further progress in coverage was observed in 
the endline data, the overall coverage remained low overall, reaching less than one-third of 
eligible children in the district. 
The INHP-II intervention made appreciable progress in focusing attention and encouraging 
action on the pediatric IFA supplementation distribution, particularly during the last year of the 
evaluation period. However, the data presented in this report indicate that further efforts are 
required to ensure that more children receive and consume daily IFA supplements starting at 12 
months of age in an effort to lower the very high prevalence of anemia among young children. 
2/20/2008 99 
CHAPTER 8: CHILD NUTRITIONAL STATUS AND ANEMIA
Child Nutritional Status 
Children 0-23 months of age were weighed and their height measured as length lying down at the 
time of the survey interview at baseline and endline. Child nutritional status was characterized 
using three indices: weight-for-age (WAZ), height-for-age (HAZ) and weight-for-height (WHZ). 
Weight-for-age is a composite measure of acute and chronic malnutrition, and reflects the extent 
to which a child is underweight. The height-for-age index is a measure of stunting, also 
reflecting chronic malnutrition, while the weight-for-height index reflects wasting as a measure 
of acute malnutrition. The 3 nutritional status indices are expressed as Z-scores relative to the 
median of a standard reference population (WHO, 1979) reflecting normal size at a given age. 
Children whose heights and weights were more than 2 standard deviations from the median 
reference population are considered to be undernourished. Those who were more than 3 standard 
deviations from the reference population are considered to be severely undernourished. All 
values outside the range -5 to 5 standard deviations of each z-score were excluded from these 
analyses. 
Underweight
Andhra Pradesh. The mean WAZ, and the proportion of underweight and severe underweight 
among all children did not change significantly in the intervention relative to the comparison 
district (Tables 8.1-8.3). There were no significant differences by child sex also, although there 
was a worsening trend observed among boys, but not girls. The strong trend of underweight 
prevalence increasing dramatically with age remained at endline, but there were no differences 
between districts. However, severe underweight among children 6-17 months and mean WAZ 
among children 12-17 months were significantly worse in the intervention compared to the 
comparison district. Nevertheless, this age-specific trend did not carry through to the oldest age 
group of children 18-23 months where mean WAZ and underweight prevalence were unchanged 
between baseline and endline in the intervention district and slightly worse in the comparison 
district. 
2/20/2008 100 
Table 8.1. Mean weight-for-age Z-score (WAZ) among children 0-23 months by sex and age in Andhra 
Pradesh 
Mean WAZ 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -1.20 -1.34 -0.14 -1.25 -1.31 -0.06 -0.08 
Sex 
 
All boys -1.27 -1.42 -0.15 -1.31 -1.34 -0.03 -0.12 
All girls -1.13 -1.26 -0.13 -1.20 -1.27 -0.07 -0.06 
Age groups 
 
0-5 months -0.28 -0.43 -0.15 -0.16 -0.18 -0.02 -0.13 
6-11 months -1.14 -1.15 -0.01 -1.14 -1.22 -0.08 0.07 
12-17 months -1.66 -1.82 -0.16 -1.88 -1.85 0.03 -0.19* 
18-23 months -1.87 -1.82 0.05 -1.92 -1.96 -0.04 0.09 
 
Total # of 
children 
2232 2556 2280 2633 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table 8.2. Underweight (<-2SD WAZ) among children 0-23 months by sex and age in Andhra Pradesh 
WAZ <-2SD (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 25.7 29.2 3.5 28.5 30.0 1.5 2.0 
Sex 
 
All boys 27.5 32.6 5.1 31.0 31.6 0.6 4.5 
All girls 23.9 25.6 1.7 25.8 28.4 2.6 -0.9 
Age groups 
 
0-5 months 4.1 7.2 3.1 4.0 4.9 0.9 2.2 
6-11 months 21.5 20.0 -1.5 21.3 20.4 -0.9 -0.6 
12-17 months 37 42.1 5.1 44.3 45.7 1.4 3.7 
18-23 months 44.7 43.5 -1.2 47.3 50.6 2.8 4.0 
 
Total # of 
children 
2232 2556 2280 2633 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 101 
Table 8.3: Severe underweight (<-3SD WAZ) among children 0-23 months by sex and age in Andhra Pradesh 
WAZ <-3SD 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 4.4 6.5 2.1 5.5 6.6 1.1 1 
Sex 
 
All boys 5.2 7.8 2.6 5.9 6.9 1 1.6 
All girls 3.4 5.2 1.8 5.1 6.3 1.2 0.6 
Age groups 
 
0-5 months 1.2 1.1 -0.1 0.7 1.6 0.9 -1 
6-11 months 2.3 5.7 3.4 3.2 3.6 -0.4 4* 
12-17 months 6.4 10.2 3.8 10.7 9.8 -0.9 4.7* 
18-23 months 8.6 8.1 -0.5 7.9 11.9 4 -4.5 
 
Total # of 
children 
2232 2556 2280 2633 
*Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Although there was a small, non-significant increase in underweight prevalence overall in the AP 
intervention district, a disaggregation of the baseline and endline data by geographical block 
reveals considerable heterogeneity. Four blocks had an increased prevalence, four were 
essentially unchanged, and one had a dramatically decreased prevalence. Clearly, the decrease in 
Mahadevpoor served to somewhat offset the large increases in underweight prevalence in 
Sulthanabad, Gangadhara, and Manthani Blocks. The reasons for this variability in findings by 
geographical area are unknown, but it could be that the samples in the poorest performing blocks 
had more boys or children 12-17 months of age at endline, i.e. groups that tended to be more 
underweight. Alternatively, local disease outbreaks, food shortages, or specific local socio￾political events that were not documented in the evaluation surveys may have contributed to this 
variability. 
2/20/2008 102 
Figure 8.1. Underweight (<-2 SD WAZ) among children 0-23 months by geographical block in Andhra 
Pradesh intervention district (Karimnagar) 
0
10
20
30
40
50
60
Jagital
Sircilla
Peddapalli
Sulthanabad
Gangadhara
Mallial
B. D. Palli
Manthani
Mahadevpoor
ALL BLOCKS
WAZ <-2SD (%)
Baseline
Endline
Uttar Pradesh. As in AP district, there was no significant change in the proportion of 
underweight and severe underweight among all children in the intervention relative to the 
comparison district (Tables 8.4-8.6). However, underweight and mean WAZ were significantly 
worse among girls, but not boys. Underweight prevalence among children <1 year of age did not 
differ between districts, but was worse among children 12-23 months. The mean WAZ among 
children 12-23 months dropped significantly in the intervention district.
Table 8.4. Mean weight-for-age Z-score (WAZ) among children 0-23 months by sex and age in Uttar Pradesh 
Mean WAZ 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -1.49 -1.58 -0.09 -1.49 -1.57 -0.08 -0.01 
Sex 
 
All boys -1.54 -1.53 0.01 -1.49 -1.62 -0.13 0.14 
All girls -1.44 -1.63 -0.19 -1.49 -1.51 -0.02 -0.17* 
Age groups 
 
0-5 months -0.54 -0.57 -0.03 -0.57 -0.73 -0.16 0.13 
6-11 months -1.56 -1.60 -0.04 -1.53 -1.63 -0.1 0.06 
12-17 months -2.00 -2.20 -0.2 -2.09 -2.10 -0.01 -0.19 * 
18-23 months -2.08 -2.18 -0.1 -2.13 -1.96 0.17 -0.27 * 
 
Total # of 
children 
1738 2137 2067 2059 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 103 
Table 8.5. Underweight (<-2SD WAZ) among children 0-23 months by sex and age in Uttar Pradesh 
WAZ <-2SD (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 36.0 41.0 5.0 35.8 38.4 2.6 2.4 
Sex 
 
All boys 37.8 39.7 1.9 35.1 39.1 4 -2.1 
All girls 34.1 42.3 8.2 36.5 37.6 1.1 7.1 * 
Age groups 
 
0-5 months 10.1 12.0 1.9 10.5 14.3 3.8 -1.9 
6-11 months 34.1 36.4 2.3 31.2 35.4 4.2 -1.9 
12-17 months 51.7 60.9 9.2 53.5 55.7 2.2 7.0 
18-23 months 54.5 60.0 5.5 58.7 53.3 -5.4 10.9 * 
 
Total # of 
children 
1738 2124 2067 2039 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table 8.6. Severe underweight (<-3SD WAZ) among children 0-23 months by sex and age in Uttar Pradesh 
WAZ <-3SD (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 9.2 12.4 3.2 8.0 12.5 4.5 -1.3 
Sex 
 
All boys 9.9 13.0 3.1 7.5 13.9 6.4 -3.3 
All girls 8.3 11.8 3.5 8.5 10.9 2.4 1.1 
Age groups 
 
0-5 months 0.6 2.9 2.3 1.6 3.8 2.2 0.1 
6-11 months 7.7 10.5 2.8 6.3 11.9 5.6 -2.8 
12-17 months 17.1 19.4 2.3 13.9 18.5 4.6 -2.3 
18-23 months 12.1 18.6 6.5 12.7 17.3 4.6 1.9 
 
Total # of 
children 
1738 2124 2067 2039 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Disaggregation of the underweight data by geographic block in the UP intervention district 
revealed that the small increase in overall prevalence from baseline to endline matched the trend 
observed in 8 of the 10 blocks. In fact, 4 of those 8 blocks had increases in underweight 
prevalence that exceeded the increase reported for the entire district. The other 2 blocks, 
Ramnagar and Trivediganj, reported very small decreases in underweight prevalence. Therefore, 
the aggregated data reported above appears to reflect the general trend in underweight across the 
10 blocks surveyed in the district. 
2/20/2008 104 
Figure 8.2. Underweight (<-2 SD WAZ) among children 0-23 months by geographical block in Uttar Pradesh 
intervention district (Barabanki)
0
10
20
30
40
50
60
Ramnagar
Masauli
Dewa
Trivediganj
Nindura
Haidergarh
Banki
Fatepur
Banikodar
Harakh
ALL BLOCKS
WAZ <-2 SD (%)
Baseline
Endline
Stunting
Andhra Pradesh. The prevalence of stunting and severe stunting among all children and stratified 
by age did not vary from baseline to endline between the 2 districts (Tables 8.8-8.9). However, 
intervention effects differed by sex with significantly increased stunting among boys, but not 
girls. The mean HAZ was significantly worse overall and among children 12-17 months in the 
intervention relative to the control district (Table 8.7). 
Table 8.7. Mean height-for-age Z-score (HAZ) among children 0-23 months by sex and age in Andhra 
Pradesh 
Mean HAZ 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -1.04 -1.17 -0.13 -1.13 -1.12 0.01 -0.14* 
Sex 
 
All boys -1.07 -1.20 -0.13 -1.19 -1.16 0.03 -0.16* 
All girls -1.00 -1.13 -0.13 -1.07 -1.08 -0.01 -0.12 
Age groups 
 
0-5 months -0.36 -0.58 -0.22 -0.25 -0.33 -0.08 -0.14 
6-11 months -0.90 -0.90 0 -0.93 -0.94 -0.01 0.01 
12-17 months -1.32 -1.42 -0.10 -1.58 -1.44 0.14 -0.24* 
18-23 months -1.73 -1.68 0.05 -1.88 -1.84 0.04 0.01 
 
Total # of 
children 
2192 2513 2267 2605 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 105 
Table 8.8. Stunting (<-2SD HAZ) among children 0-23 months by sex and age in Andhra Pradesh 
HAZ <-2SD (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 22.8 23.3 0.5 25.8 23.2 -2.4 2.9 
Sex 
 
All boys 22.9 25.0 2.1 27.4 24.8 -2.6 4.7 
All girls 22.7 21.5 -1.2 24.2 21.5 -2.7 1.5 
Age groups 
 
0-5 months 9.7 12.8 3.1 9.3 8.1 -1.2 4.3 
6-11 months 20.1 13.4 -6.7 18.4 14.6 -3.8 -2.9 
12-17 months 27.2 29.8 2.6 35.3 32.1 -3.2 5.8 
18-23 months 38.1 35.7 -2.4 43.7 40.4 -3.3 0.9 
 
Total # of 
children 
2192 2513 2267 2605 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table 8.9. Severe stunting (<-3SD HAZ) among children 0-23 months by sex and age in Andhra Pradesh 
HAZ <-3 SD (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 6.6 7.0 0.4 7.5 6.2 -1.3 1.7 
Sex 
 
All boys 7.4 7.7 0.3 7.7 6.7 -1 1.3 
All girls 5.7 6.2 0.5 7.2 5.6 -1.6 2.1 
Age groups 
 
0-5 months 3.1 3.6 0.5 3.3 2.3 -1 1.5 
6-11 months 4 4.1 0.1 2.7 2.3 -0.4 0.5 
12-17 months 7.5 7.6 0.1 11.2 7.4 -3.8 3.9 
18-23 months 13.7 12.1 -1.6 14.1 14.5 0.4 -2 
 
Total # of 
children 
2192 2513 2267 2605 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Disaggregation of the AP intervention district stunting data revealed a considerable amount of 
heterogeneity in the prevalence by geographical block (Figure 8.3). The overall prevalence of 
stunting was essentially unchanged from baseline to endline, but only one of the 9 blocks, 
Sulthanabad, followed that same pattern. Stunting prevalence increased in 4 blocks and 
decreased in another 4 blocks. In 2 of the 4 blocks with a decreasing trend, Mallial and 
Mahadevpoor, stunting prevalence decreased by approximately half. Sircilla and Peddapalli had 
2/20/2008 106 
the largest increases in stunting. Mahadevpoor block was the same block where underweight 
prevalence decreased significantly. The same possible reasons for the heterogeneity of 
underweight prevalence may also explain the variation in stunting observed in the AP 
intervention district. 
Figure 8.3. Stunting (<-2 SD HAZ) among children 0-23 months by geographical block in Andhra Pradesh 
intervention district (Karimnagar) 
0
10
20
30
40
50
60
Jagital
Sircilla
Peddapalli
Sulthanabad
Gangadhara
Mallial
B. D. Palli
Manthani
Mahadevpoor
ALL BLOCKS
HAZ <-2 SD (%)
Baseline
Endline
Uttar Pradesh. Stunting and severe stunting prevalence did not differ overall or by sex between 
the districts (Tables 8.11-8.12). The only age-specific differences were among children 18-23 
months where stunting prevalence was unchanged in the intervention district, but was reduced 
significantly in the comparison district. Mean HAZ followed this same trend among children in 
the second year of life (Table 8.10). Mean HAZ was significantly worse among girls only in the 
comparison relative to the comparison district. 
Table 8.10. Mean height-for-age Z-score (HAZ) among children 0-23 months by sex and age in Uttar Pradesh 
Mean HAZ 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -1.75 -1.74 0.01 -1.71 -1.65 0.06 -0.05 
Sex 
 
All boys -1.83 -1.72 0.11 -1.72 -1.70 0.02 0.09 
All girls -1.66 -1.76 -0.1 -1.69 -1.60 0.09 -0.19 * 
Age groups 
 
0-5 months -0.74 -0.76 -0.02 -0.71 -0.74 -0.03 0.01 
6-11 months -1.68 -1.62 0.06 -1.60 -1.46 0.14 -0.08 
12-17 months -2.30 -2.30 0 -2.28 -2.13 0.15 -0.15 
18-23 months -2.58 -2.55 0.03 -2.71 -2.51 0.2 -0.17 
Total # of 
children 
1603 2056 2030 1957 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 107 
Table 8.11. Stunting (<-2SD HAZ) among children 0-23 months by sex and age in Uttar Pradesh 
HAZ <-2SD (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 43.4 43.9 0.5 43.6 42.5 -1.1 1.6 
Sex 
 
All boys 45.0 43.1 -1.9 45.3 43.9 -1.4 -0.5 
All girls 41.5 44.7 3.2 41.9 41.0 -0.9 4.1 
Age groups 
 
0-5 months 12.7 16.2 3.5 13.6 18.9 5.3 -1.8 
6-11 months 39.5 35.4 -4.1 34.6 33.5 -1.1 -3.0 
12-17 months 59.8 60.8 1.0 62.0 55.8 -6.2 7.2 
18-23 months 70.9 70.5 -0.4 79.4 68.3 -11.1 10.7 * 
 
Total # of 
children 
1603 2056 2030 1957 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table 8.12. Severe stunting (<-3SD HAZ) among children 0-23 months by sex and age in Uttar Pradesh 
HAZ <-3SD 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 19.0 19.1 0.1 17.6 19.2 1.6 -1.5 
Sex 
 
All boys 20.0 19.7 -0.3 17.9 21.1 3.2 -3.5 
All girls 18.0 18.4 0.4 17.3 17.0 -0.3 0.7 
Age groups 
 
0-5 months 3.9 5.1 1.2 3.0 7.6 4.6 -3.4 
6-11 months 13.1 13.7 0.6 11.3 12.9 1.6 -1.0 
12-17 months 28.1 27.0 -1.1 25.1 24.6 -0.5 -0.6 
18-23 months 36.5 34.6 -1.9 39.2 35.4 -3.8 1.9 
 
Total # of 
children 
1603 2056 2030 1957 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Although the overall prevalence of stunting in the UP intervention district was unchanged from 
baseline to endline, changes in prevalence at the block level did not match this trend. Stunting 
prevalence decreased modestly in 4 blocks and increased in another 6 blocks (Figure 8.4). The 
largest increases in stunting were observed in Banki, Fatepur, and Harakh blocks. Banki and 
Fatepur blocks also reported large increases in underweight prevalence from baseline to endline. 
It is not known what trends in the data or events may have contributed to these local increases. 
2/20/2008 108 
The mean HAZ score was significantly lower among girls and the prevalence of stunting 
increased among children <6 months in the intervention district, so larger proportions of either of 
these sub-groups in a block level sample could have contributed to the observed results. 
Figure 8.4. Stunting (<-2 SD HAZ) among children 0-23 months by geographical block in Uttar Pradesh 
intervention district (Barabanki) 
0
10
20
30
40
50
60
Ramnagar
Masauli
Dewa
Trivediganj
Nindura
Haidergarh
Banki
Fatepur
Banikodar
Harakh
ALL BLOCKS
HAZ <-2 SD (%)
Baseline
Endline
Wasting 
Andhra Pradesh. Wasting prevalence overall and by sex increased by a similar magnitude from 
baseline to endline in the intervention and comparison districts. This trend was observed among 
children of all ages except 0-5 months where wasting prevalence doubled in the intervention 
district relative to a ~50% decrease in the comparison district. Boys were more likely to be 
wasted than girls in the sample. By contrast, severe wasting was significantly worse among all 
children, boys, and children 6-17 months in the intervention relative to the control district. Mean 
WHZ did not differ between districts. 
2/20/2008 109 
Table 8.13. Mean weight-for-height Z-score (WHZ) among children 0-23 months by sex and age in Andhra 
Pradesh 
Mean WHZ 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -0.70 -0.78 -0.08 -0.72 -0.79 -0.07 -0.01 
Sex 
 
All boys -0.78 -0.86 -0.08 -0.77 -0.81 -0.04 -0.04 
All girls -0.62 -0.69 -0.07 -0.68 -0.78 -0.10 0.03 
Age groups 
 
0-5 months -0.08 -0.08 0 -0.03 -0.05 -0.02 0.02 
6-11 months -0.53 -0.58 -0.05 -0.53 -0.62 -0.09 0.04 
12-17 months -1.04 -1.13 -0.09 -1.15 -1.20 -0.05 -0.04 
18-23 months -1.29 -1.17 0.12 -1.25 -1.30 -0.05 0.17 
 
Total # of 
children 
2160 2498 2245 2584 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table 8.14. Wasting (<-2SD WHZ) among children 0-23 months by sex and age in Andhra Pradesh 
WHZ <-2SD (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 12.1 16.0 4.1 10.9 13.1 2.2 1.9 
Sex 
 
All boys 13.3 17.3 4.0 12.5 13.7 1.2 2.8 
All girls 10.8 14.6 3.8 9.3 12.6 3.3 0.5 
Age groups 
 
0-5 months 3.8 8.5 4.7 3.5 2.2 -1.3 6.0* 
6-11 months 8.8 11.4 2.6 6.5 8.2 1.7 0.9 
12-17 months 16.8 19.1 2.3 15.7 19.1 3.4 -1.1 
18-23 months 20.8 23.4 2.6 19.6 23.9 4.3 -1.7 
 
Total # of 
children 
2160 2498 2245 2584 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 110 
Table 8.15. Severe wasting (<-3SD WHZ) among children 0-23 months by sex and age in Andhra Pradesh 
WHZ <-3SD (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 2.1 3.0 0.9 2.1 1.7 -0.4 1.3* 
Sex 
 
All boys 2.2 3.5 1.3 2.1 1.9 -0.2 1.5 
All girls 1.9 2.4 0.5 2 1.6 -0.4 0.9 
Age groups 
 
0-5 months 1.6 2.5 0.9 0.6 1.0 0.4 0.5 
6-11 months 1.2 3.0 1.8 1.9 1.2 -0.7 2.5* 
12-17 months 2.1 3.4 1.3 3.4 2.4 -1.0 2.3 
18-23 months 3.9 2.8 -0.9 2.3 2.4 0.1 -1.0 
 
Total # of 
children 
2160 2498 2245 2584 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
The disaggregated wasting data from the AP intervention district were more homogenous than 
was seen with underweight and stunting prevalence in the same blocks. The overall prevalence 
of wasting increased by ~30% (4 percentage points) from baseline to endline, and 7 of 9 blocks 
followed this same pattern (Figure 8.5). Wasting prevalence increased most dramatically (by 
~50%) in Sulthanabad and Manthani blocks. However, 2 blocks, Peddapalli and Mahadevpoor, 
went against the dominant trend with very large decreases in prevalence at endline. The latter 
block was the only one in the AP intervention district to report decreases in all 3 child 
malnutrition indicators. 
2/20/2008 111 
Figure 8.5. Wasting (<-2 SD WHZ) among children 0-23 months by geographical block in Andhra Pradesh 
intervention district (Karimnagar) 
0
5
10
15
20
25
30
Jagital
Sircilla
Peddapalli
Sulthanabad
Gangadhara
Mallial
B. D. Palli
Manthani
Mahadevpoor
ALL BLOCKS
WHZ <-2 SD (%)
Baseline
Endline
Uttar Pradesh. As in AP districts, the prevalence of wasting and severe wasting increased from 
baseline to endline in both districts (Tables 8.17-8.18). However, the increase among all children 
and by sex was significantly greater in the comparison relative to the intervention district. This 
same trend was observed for all ages except children 17-23 months. Mean WHZ did not differ 
significantly between districts (Table 8.16). 
Table 8.16. Mean weight-for-height Z-score (WHZ) among children 0-23 months by sex and age in Uttar 
Pradesh 
Mean WHZ 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children -0.43 -0.55 -0.12 -0.51 -0.63 -0.12 0 
Sex 
 
All boys -0.46 -0.55 -0.09 -0.52 -0.68 -0.16 0.07 
All girls -0.40 -0.56 -0.16 -0.51 -0.58 -0.07 -0.09 
Age groups 
 
0-5 months 0.04 -0.06 -0.1 -0.10 -0.37 -0.27 0.17 
6-11 months -0.31 -0.37 -0.06 -0.37 -0.56 -0.19 0.13 
12-17 months -0.69 -0.88 -0.19 -0.85 -0.86 -0.01 -0.18 
18-23 months -0.92 -0.94 -0.02 -0.89 -0.72 0.17 -0.19 
 
Total # of 
children 
1581 2042 1983 1931 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 112 
Table 8.17. Wasting (<-2SD WHZ) among children 0-23 months by sex and age in Uttar Pradesh 
WHZ <-2SD (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 4.9 8.5 3.6 6.7 13.8 7.1 -3.5* 
Sex 
 
All boys 5.4 8.4 3.0 7.1 14.3 7.2 -4.2* 
All girls 4.5 8.5 4.0 6.3 13.3 7.0 -3.0 
Age groups 
 
0-5 months 2.0 3.9 1.9 3.0 9.8 6.8 -4.9* 
6-11 months 4.6 4.6 0 4.8 10.9 6.1 -6.1* 
12-17 months 6.5 10.4 3.9 9.6 17.9 8.3 -4.4 
18-23 months 7.5 16.0 8.5 11.1 16.5 5.4 3.1 
 
Total # of 
children 
1581 2042 1983 1931 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table 8.18. Severe wasting (<-3SD WHZ) among children 0-23 months by sex and age in Uttar Pradesh 
WHZ <-3SD (%) 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
All children 0.6 1.2 0.6 1.2 4.3 3.1 -2.5* 
Sex 
 
All boys 0.6 1.5 0.9 1.2 4.1 2.9 -2.0* 
All girls 0.7 0.8 0.1 1.1 4.6 3.5 -3.4* 
Age groups 
 
0-5 months 0.7 1.4 0.7 1.0 4.7 3.7 -3.0* 
6-11 months 1.3 0.7 -0.6 1.0 4.2 3.2 -3.8* 
12-17 months 0.4 1.0 0.6 1.2 4.3 3.1 -2.5* 
18-23 months 0 1.7 1.7 1.5 3.9 2.4 -0.7 
 
Total # of 
children 
1581 2042 1983 1931 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Wasting prevalence in 8 of 10 blocks in the UP intervention district followed the overall trend of 
increasing from baseline to endline. Dewa and Fatepur blocks reported the largest increases. In 
one block (Ramnagar), the prevalence remained unchanged, while it decreased modestly in 
another (Banki). Nonetheless, the trend in the disaggregated data indicates that it is represented 
quite well by the summary measure among all blocks. 
2/20/2008 113 
Figure 8.6. Wasting (<-2 SD WHZ) among children 0-23 months by geographical block in Uttar Pradesh 
intervention district (Barabanki)
0
5
10
15
20
25
30
Ramnagar
Masauli
Dewa
Trivediganj
Nindura
Haidergarh
Banki
Fatepur
Banikodar
Harakh
ALL BLOCKS
WHZ <-2 SD (%)
Baseline
Endline
Anemia
Anemia is defined as a low hemoglobin concentration in blood. The most common cause of 
anemia is iron deficiency, although other micronutrient deficiencies and infections such as 
malaria also contribute to the problem. Infants are born with iron stores, but they are depleted by 
rapid growth in the first few months of life and the hematological changes that accompany the 
transition to the extrauterine environment. Breast milk provides sufficient iron during the first 6 
months of life, but additional iron is required from 6-24 months of age to prevent the 
development of iron deficiency anemia. Iron deficiency anemia during the first 2 years of life 
leads to deficits in cognitive and motor development. Additional iron can be obtained from iron￾fortified complementary foods or iron supplements. The Government of India policy is that all 
children 12 months of age and older should receive daily pediatric iron supplements. Anemia 
among children <2 years is defined as hemoglobin concentration < 11 g/dL. Mild, moderate and 
severe anemia are defined as 10.0-10.9, 7.0-9.9, and <7.0 g/dL, respectively. 
Hemoglobin concentration was assessed in all children 12-23 months of age from a single drop 
of blood collected by fingerstick and assessed on site using the Hemocue method. Mothers were 
informed of their child’s hemoglobin measurement, and in the case of severe anemia, were 
referred to the local health center. 
Andhra Pradesh. Anemia prevalence among children 12-23 months dropped slightly from 
baseline to endline, but remained at ~85% in both districts (Table 8.19). Mean hemoglobin 
increased and the severity of anemia was reduced at endline, but the changes were comparable in 
the two districts. 
2/20/2008 114 
Table 8.19. Anemia among children 12-23 months of age in Andhra Pradesh 
Child anemia (%), mean hemoglobin 
Intervention (Karimnagar) Comparison (Rangareddy) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any anemia 
(Hb<11 g/dL) 
88.0 85.8 -2.2 92.3 86.6 -5.7 3.5 
Moderate￾severe anemia 
(Hb<10 g/dL) 
68.2 61.6 -6.6 73.4 63.3 -10.1 3.5 
Severe anemia 
(Hb<7 g/dL) 
7.1 3.7 -3.4 8.9 5.6 -3.3 -0.1 
Mean 
Hemoglobin 
(g/dL) 
9.21 9.51 0.30 9.00 9.43 0.43 -0.13 
Total # of 
children 
996 1288 1052 1171 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Uttar Pradesh. Anemia prevalence among children 12-23 months was significantly reduced in 
the intervention relative to the comparison district due to increased prevalence at endline in the 
latter area. However, child anemia prevalence remained alarmingly high at 87% overall. In 
contrast to AP districts, the severity of anemia increased and mean hemoglobin concentration 
decreased approximately equally in both UP districts. 
Table 8.20. Anemia among children 12-23 months of age in Uttar Pradesh 
Child anemia (%), mean hemoglobin 
Intervention (Barabanki) Comparison (Unnao) 
Proportion 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any anemia 
(Hb<11 g/dL) 
88.6 86.7 -1.9 83.1 87.4 4.3 -6.2 * 
Moderate￾severe anemia 
(Hb<10 g/dL) 
65.5 67.1 1.6 53.1 58.3 5.2 -3.6 
Severe anemia 
(Hb<7 g/dL) 
4.0 8.3 4.3 3.6 5.5 1.9 2.4 
Mean 
Hemoglobin 
(g/dL) 
9.30 9.12 -0.18 9.74 9.39 -0.35 0.17 
Total # of 
children 
699 867 755 867 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Summary
Child nutritional status
The lack of a positive intervention effect on malnutrition is not entirely surprising in light of the 
changes observed in infant and child feeding practices in both states. In AP, cereal intake was 
somewhat prevalent, but increases in the intervention district were matched or exceeded in the 
2/20/2008 115 
comparison district. Indicators of the quantity and frequency of dietary intake also increased by 
similar proportions in both AP districts. Intake of legumes, fruits, and vegetables remained low 
or decreased over the intervention period and dietary diversity was poor, suggesting that 
children’s diets contained inadequate amounts of micronutrients that might inhibit growth. 
However, these trends in feeding practices in AP do not explain the decline in the mean HAZ 
score and severe wasting overall or the increased stunting among boys in the intervention relative 
to the comparison district. Multivariate adjustment for a variety of sociodemographic 
characteristics did not alter the results significantly. It is possible that atypical local events such 
as disease outbreaks may have occurred in the program district during the intervention period 
that may have adversely affected children’s nutritional status. 
In UP, intervention effects on many complementary feeding practices were evident from the 
endline data, including significantly increased intake of cereals and legumes, and increased 
quantity, frequency, and diversity of dietary intake. However, fruit and vegetable intake did not 
improve, and the prevalence of complementary feeding indicators remained quite low in absolute 
terms, particularly for intake of solids in the past 24 hours and quantity and frequency of dietary 
intake. The demonstrated improvements in feeding practices in the UP intervention district might 
be expected to yield reductions in malnutrition prevalence. However, underweight prevalence 
and mean HAZ score were worse among girls and among children 12-23 months in the 
intervention relative to the comparison district. This suggests that dietary improvements 
demonstrated thus far in the UP intervention district were insufficient to yield improvements in 
nutritional status. 
Child anemia
The GOI policy of daily iron supplementation for children starting at 12 months of age is meant 
to address the imbalance between high iron requirements and low dietary iron intake of children 
6-24 months. Clearly, the extraordinarily high prevalence of child anemia in both states indicates 
that initiating this supplementation protocol at 12 months of age is too late in these populations. 
Data on receipt of pediatric IFA supplements demonstrates that the intervention was effective at 
increasing supplement distribution, particularly in UP state. However, the detailed data collected 
at endline on IFA supplementation revealed that most children exposed to this component of the 
program intervention received and consumed an inadequate quantity of IFA supplements. In AP 
state, this is explained by a low coverage level of only 30% by endline. In UP, two-thirds of 
eligible children were receiving IFA supplements by endline, but the increase in coverage 
occurred in the last months of the intervention period only, so exposure to this intervention 
component was most likely insufficient to impact iron status. Therefore, the lack of an 
intervention effect on anemia prevalence is not surprising. 
The dramatic increase in the prevalence of severe anemia in UP is a cause for concern. If 
increased program coverage was accompanied by increased intensity of the intervention (i.e. 
more IFA supplements consumed per child), one could expect a reduction in the severity of 
anemia at a minimum. However, the current prevention protocol may be insufficient to impact 
the high prevalence and severity of anemia in this population. Additional strategies that address 
treatment of severe anemia should be considered. 
2/20/2008 116 
SECTION II: COHORT STUDY IN UTTAR PRADESH 
STATE 
2/20/2008 117 
CHAPTER 9: METHODOLOGY AND DATA COLLECTION FOR THE COHORT 
STUDY 
Rationale for the Cohort Study Design 
A longitudinal, cohort study was conducted in conjunction with the cross-sectional surveys to 
further assess INHP-II program effectiveness to improve infant feeding practices and reduce 
malnutrition among children less than 2 years of age in Uttar Pradesh state. Pregnant women 
were enrolled in the cohort and they and their infants were followed from delivery through 18 
months of age. The cohort study design allowed for collection of more detailed data on 
intervention exposures and dietary intakes, and repeated follow-up of children to assess 
nutritional outcomes, including child growth, over time. However, the cohort sample size was 
not designed to provide the primary means of assessing the program’s impact on child nutritional 
status and growth as this was provided by the cross-sectional evaluation. 
Cohort Study Design 
The cohort study was conducted in Uttar Pradesh state only. An INHP-II district was selected as 
the intervention area (Barabanki) and a similar non-INHP-II (with an ICDS program) district was 
chosen as the comparison area (Unnao). Although the cross-sectional and cohort components of 
the study were both conducted in the same UP districts, there was limited overlap between the 
women and their children recruited into the cohort and those enrolled in the cross-sectional 
evaluation because of the multi-stage sampling design. Enrollment information and 
measurements were collected from pregnant mothers during the third trimester (6-9 mo 
gestation) on socio-demographic factors, antenatal advice and services received, birth 
preparedness, diet, rest and workload, and anthropometry. Additional information and 
measurements were collected subsequently on utilization of health services and advice received 
that comprised the program intervention, as well as maternal and infant diet, infant feeding 
practices, and nutritional status at 7 days, and 3, 6, 9, 12, 15, and 18 months postpartum. 
Cohort Sampling Design 
Sample size 
Sample size was calculated to detect differences in infant dietary intake outcomes between the 
intervention and comparison districts. The statistical power used for the calculations was 80% to 
detect a difference of 50 kcal/day in energy intakes from complementary foods at 7-10 months 
and 12-15 months of age at a significance level of 95% (alpha = 0.05). This sample size would 
allow us to detect differences in nutritional status of about 0.2 Z between groups in cross￾sectional comparisons, with smaller differences in growth (about 0.1 Z) detectable given the 
longitudinal nature of this evaluation component. Primary outcomes were breastfeeding 
initiation and duration, and infant feeding practices. Secondary outcomes included energy intake 
from complementary foods and growth. Thus, it was estimated that approximately 400 
mother/child pairs in each district were required to detect the differences expected. Using data 
from the study’s baseline cross-sectional survey, it was estimated that there were 12.2 pregnant 
women per AWC area, of whom 5 (~40%) would be in their third trimester of pregnancy. Thus, 
to enroll 400 pregnant women in each district, women from 81 AWC areas were recruited. 
2/20/2008 118 
Cohort Sample Selection
A multi-stage sampling strategy similar to the cross-sectional evaluation was used in the cohort 
component also. The sample of pregnant women was enrolled from AWCs chosen from selected 
sectors within the blocks selected for the cohort in each district. Each rural block in the districts 
contained 5-7 sectors, an area with 15-25 AWC and an estimated population of 20,000-25,000 
persons. The sampling universe for the intervention district sample selection included only those 
sectors and AWC areas that were participating in CARE’s INHP-II program as replication sites. 
All sectors in the selected blocks had AWC areas participating in the intervention. In the 
comparison district, the sampling universe included all areas with a functioning AWC from all 
sectors of the selected blocks in the district.
Three rural blocks from each district were purposively selected for the cohort: Nindura, Banki, 
and Fatehpur blocks in Barabanki District; and Nawabganj, Hasanganj, and Asoha blocks in 
Unnao District. The block selection criteria for both districts were: 1) accessibility of the district 
town; 2) physical proximity to other selected blocks; and 3) absence of other known MCH 
nutrition and health projects. Block population size and number of AWC were also considered 
in selecting comparable blocks in the 2 districts. 
Four sectors were selected randomly from each of the 3 blocks for both districts. There were 135 
and 192 AWCs in the selected blocks from the intervention and comparison districts, 
respectively. Within each sector, 6-7 AWCs were selected randomly to yield a total of 81 AWCs 
per district. This number of AWCs was expected to yield at least 400 pregnancies per district. 
Cohort Data Collection 
The cohort study was conducted from May 2004 to June 2006. Enrollment data were collected 
from pregnant women at 6-9 months gestation, and follow-up of women and their infants was 
conducted at 7 days, and 3, 6, 9, 12, 15, and 18 months postpartum. These time points coincided 
with critical periods when services were to be provided and advice delivered. The data 
collection activities were carried out by staff employed locally by Johns Hopkins. 
Personnel
Eight field interviewers were employed by Johns Hopkins for data collection for the duration of 
the cohort study. The interview teams consisted of 2 interviewers; the four teams worked in both 
Barabanki and Unnao Districts. Field teams were coordinated and supervised by the Nutrition 
Project Coordinator with assistance from the Field Coordinator. Dr. VK Srivastava and his 
colleagues at KGMU provided technical assistance to the Project Coordinator and the cohort 
field team. 
Household listing and eligibility criteria
A household listing operation was conducted in all the sampled AWC areas. The purpose of the 
household listing was to identify households with eligible pregnant women. Women were 
considered eligible if: 
(a) They were at least 24 weeks (6 months completed) pregnant on the day of household 
listing. Gestational age was assessed by recall of the date of the last menstrual period 
(LMP) and by inquiring about the number of completed months of pregnancy. 
2/20/2008 119 
(b) They intended to remain in the program area (same village) for the duration of the study. 
Women visiting their parents’ home were excluded unless it was determined that their 
permanent residence was also within the cohort sample area. Women reporting that they 
intended to visit their parents’ home for delivery or during the postpartum period were 
enrolled if the visit was not greater than 3 months and there was a willing family member 
reporting on the birth regardless of the woman’s location. 
(c) They were permanently residing (greater than one year) with their parents. 
Enrollment
Once eligibility of a pregnant woman was determined, informed consent was taken and she was 
invited to enroll in the study. All eligible women from sampled AWC areas were recruited for 
the cohort. If more than one eligible pregnant woman was identified in a household, both were 
enrolled. Thus, actual cohort sample recruitment was not limited by the target sample size, but 
included all eligible, consenting pregnant women in the sampled areas.
Cohort follow-up
Enrolled pregnant women were asked to inform the study office within 24 hours of delivery. A 
calendar with project office phone numbers was provided to each study participant to facilitate 
this process. Follow-up visits were scheduled for critical periods immediately after programmatic 
exposure occurred. The first postpartum visit was scheduled for 7 days after delivery, and 
subsequent visits were scheduled for 3, 6, 9, 12, 15 and 18 months postpartum. 
Data collected
During the pregnancy visit (6-9 months gestation) at the time of enrollment, mothers were 
interviewed about socio-demographic information and birth preparedness, and anthropometric 
measures of weight and height were taken. Mothers were asked during this interview and again 
at 7 days postpartum about the health advice received and health service utilization throughout 
pregnancy. Information on current maternal diet and rest and workload was collected during the 
pregnancy visit, and again at 3, 6, and 9 months postpartum. At the 7-day postpartum visit, data 
were collected on the delivery experience and breastfeeding initiation, and in the case of the 
latter repeated in month 3. During post-partum visits, infant anthropometric measures of weight 
and length/height were taken, and mothers were asked about immunizations, infant feeding 
practices, feeding and care during illness, and the health advice received. A 24-hr dietary recall 
was conducted on a sub-sample of 10% of the mothers (n=96) during in months 12 and 18 to 
estimate within-subject variation in dietary intakes. 
Cohort Data Management
All data collection instruments for enrollment and follow-up visits were edited in the field and 
brought to the main project office in Lucknow for further editing and data entry by local Johns 
Hopkins staff. The quality of data was checked and validated for completion, range check, skip￾pattern matching, heaping and consistency. In the case of discrepancies, data were cross-checked 
with the paper questionnaires for immediate resolution. Statistical analyses were conducted at 
JHU in Baltimore, Maryland. 
Cohort Sample Coverage 
2/20/2008 120 
Nearly all eligible pregnant women were enrolled in the cohort study (Figure 9.1). Of the 2,547 
pregnant women identified in the sampled areas in both districts, 980 were considered eligible 
for the study. Enrollment coverage was greater than 95% in both districts (Barabanki = 98.6%, 
Unnao=96.6%). The main reason for non-response was failure to find women at home despite 3 
re-visits for all eligible survey respondents. 
Figure 9.1. Cohort Study Profile 
Statistical Methods 
After assessing the data for quality, frequency distributions of the sample characteristics were 
examined with descriptive statistics. We compared the similarity in the characteristics of sample 
distribution between intervention and comparison arms with Student’s t-test and Pearson’s chi￾square for metric (continuous) and non-metric (categorical) variables, respectively. 
Statistical comparisons of intervention and comparison arms
For cross-sectional outcomes examined among cohort infants at specific ages from 0 to 20 
months, indicators were compared between the intervention and comparison districts using the 
Student’s t-test for continuous and Pearson’s chi-square for categorical variables. 
Analysis of the intervention effect on physical growth was based on a latent growth model 
(Diggle, 1994) in which the outcome examined was the change in nutritional status over a one 
month period in combination with or in biostatistical terms, in interaction with, the district of 
Eligible pregnant 
women in BBK 
n = 512 
Eligible pregnant 
women in Unnao 
n = 468 
Enrolled pregnant 
women in BBK 
n = 505 
Enrolled pregnant 
women in Unnao 
n = 452 
Total pregnant women in UP 
(Barabanki & Unnao) 
n = 2,547
Mothers and infants 
followed through 18 
mo postpartum 
n = 406 
Mothers and infants 
followed through 18 
mo postpartum 
n = 356
2/20/2008 121 
residence. Longitudinal regression modeling was carried out to investigate the effect of the 
intervention over the entire follow-up period - adjusting for intra-subject correlations in 
measured outcomes at each time point. The mixed-effect multiple linear regression model 
applied in the analyses specified the correlation structure among subjects with use of a random 
effects assumption. As well, regression modeling permitted investigators to consider 
confounding factors among the sociodemographic and economic characteristics of the sample. 
Age was modeled with quadratic and cubic terms to take into consideration the declining rate of 
growth from 6 to 18 months followed by an increase thereafter from 18-24 months of age. 
Estimates of the intervention effect summed the interaction terms and multiplied by 12 to obtain 
an average one-year effect: [(age * district) + (age2 * district)] * 12. 
All analyses were conducted with STATA 8.0 (Stata Corp, TX). 
2/20/2008 122 
CHAPTER 10: SOCIODEMOGRAPHIC CHARACTERISTICS IN THE COHORT 
STUDY
The sociodemographic and economic characteristics of the cohort sample are presented in this 
chapter. To investigate for potential confounding factors in the intervention effect, the 
distribution of these characteristics were compared by district. We also present the baseline 
sociodemographic data collected from the currently pregnant women in the cross sectional study 
for the purpose of making overall comparisons across the two studies in Uttar Pradesh. Again, as 
stated in Chapter 3, the large sample size increased the likelihood of statistical significance in 
even very small differences among characteristics. This should not necessarily be construed as a 
meaningful difference for the evaluation results. 
Maternal Age
The mean age of mothers at enrollment into the cohort study in the intervention district was 25.2 
(±5.9) yr, significantly higher than the mean age in the comparison district, 24.1 yr (±5.3) 
(P=0.003). In both districts, greater than one-third of the mothers fell into the 20-24 year old 
range (Table 10.1). There was a difference of 4.6% points in the percentage of adolescent 
mothers between the districts, with the comparison group showing the higher value. Generally, 
mothers tended to fall in the older age ranges for the intervention. In the cross-sectional study, 
trends were similar with higher percentages of younger mothers observed in the comparison 
district; these differences, albeit small, were statistically significant. 
Table 10.1. Age distribution among pregnant women in the cohort and cross sectional studies, Uttar Pradesh 
* Statistical test comparing distributions between districts, significant at p<0.05
Cohort study Cross sectional study 
Intervention Comparison * Intervention Comparison * 
< 20 15.9 20.5 23.9 25.5 
20-24 36.3 37.1 34.1 36.5 
25-29 24.6 26.3 22.7 23.2 
30-34 14.1 11.9 12.1 11.2 
35+ 8.4 4.2 7.4 3.5 
Total no. of 
pregnant women 453 504 746 706 
2/20/2008 123 
Religion 
The majority of respondents in both districts were Hindu, though the proportion in various 
categories did differ significantly (Table 10.2). The intervention district had a larger percentage 
(23.8%) of mothers reporting Muslim religion compared to the comparison district (5.5%). 
These differences by district were again reduced in the cross sectional study where the 
percentage of those reporting adherence to the Hindu religion in the intervention district was 
nearly 7% percentage points higher. 
Table 10.2. Religion among pregnant women in the cohort and cross sectional studies, Uttar Pradesh 
Cohort study Cross sectional study 
Religion Intervention Comparison * Intervention Comparison * 
Hindu 76.2 94.5 83.1 91.4 
Muslim 23.8 5.5 16.9 8.1 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts, significant at p<0.05
Caste 
One half of the cohort respondents fell into the other backward caste category in the intervention 
district, followed by approximately one-third in the scheduled caste (Table 10.3). There was a 
more even distribution in the comparison district among these two categories. Few women from 
either district reported belonging to a scheduled tribe. There were comparable differences 
between the districts in the cross-sectional study that also demonstrated statistical significance. 
Table 10.3. Caste among pregnant women in the cohort and cross sectional studies, Uttar Pradesh 
Cohort study Cross sectional study 
Caste groups Intervention Comparison * Intervention Comparison * 
Scheduled caste 36.3 42.6 33.1 37.3 
Scheduled tribe 0.6 0.4 0.1 0.4 
Other backward 
caste 50.0 40.6 37.0 41.2 
Other castes 13.1 15.0 12.9 13.0 
Not applicable#
 0 1.3 16.9 8.1 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts, significant at p<0.05 
#
Includes women who do not belong to any caste or belong to a religion with no caste system. 
2/20/2008 124 
Education 
The number of school years completed by mothers using the categories from the NFHS II survey 
was considered the education level attained (Table 10.4). There were no statistical differences 
by district in education levels among the pregnant women in the cohort study. Illiteracy was 
highly prevalent among the women, with greater than two-thirds reporting no education in both 
intervention and comparison districts. Only 7.7% and 6.0% in the intervention and comparison, 
respectively, completed any high school education. In the cross sectional study among the 
currently pregnant women, there was slightly more variation among categories between the 
districts that reached statistical significance, though percentages followed similar trends to the 
cohort study and showed the predominant education level to be illiteracy. 
Table 10.4. Education level among pregnant women in the cohort and cross sectional studies, Uttar Pradesh 
Education level Cohort study Cross sectional study 
(years) Intervention Comparison Intervention Comparison * 
No 
education/illiterate 67.9 69.0 68.0 61.5 
1-7 15.3 13.3 15.6 19.0 
8-9 9.1 11.3 8.9 12.6 
10+ 7.7 6.0 7.6 6.9 
Total no. of pregnant 
women 504 453 746 706 
* Statistical test between districts, significant at p<0.05
Parity 
Parity is often found to be an important covariate in modeling growth outcomes. Differences in 
parity were not apparent between the cohort and cross-sectional pregnancy samples or between 
the intervention and comparison districts in each sample (Table 10.5). The highest proportion of 
live births fell in the 2-3 range in all districts. 
2/20/2008 125 
Table 10.5. Number of live births (parity) among pregnant women in the cohort and cross sectional studies, 
Uttar Pradesh 
 
Number of live Cohort study Cross sectional study 
births Intervention Comparison Intervention Comparison 
0 20.8 19.9 22.9 23.5 
1 16.9 19.9 19.8 20.4 
2-3 32.1 35.3 30.6 32.4 
4-5 19.6 14.6 18.0 14.3 
6+ 10.5 10.4 8.7 9.4 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts, significant at p<0.05
Inter-Birth Interval 
Women were asked to report the number of months between the birth of their last child and the 
current birth. Approximately 20% of mothers in both cohort districts reported an interval of less 
than 24 months, but there were small but significant differences in the categories 24-47 and ≥48 
months between cohort intervention and comparison mothers (Table 10.6). The distribution of 
inter-birth intervals was comparable in the cross-sectional study and did not differ between 
districts. 
Table 10.6. Inter-pregnancy birth intervals among pregnant mothers in the cohort and cross sectional studies, 
Uttar Pradesh 
Cohort study Cross sectional study 
Inter-Birth 
interval 
(months) Intervention Comparison * Intervention Comparison 
<24 months 19.1 20.1 18.6 17.4 
24-47 months 36.1 30.5 29.5 30.6 
48+ months 6.9 9.5 8.7 7.1 
N/A 37.9 40.0 43.2 44.9 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts within a state, significant at p<0.05 
Household Toilet Facilities 
Most women in both intervention and comparison districts reported having no toilet facility; in 
the comparison, this was overwhelmingly the case, in 94.3% of mothers. Flush toilets were more 
2/20/2008 126 
common in the intervention district than in the comparison district as were open, shared, or 
public toilets. Statistically significant differences across these categories were found among 
districts for both the cohort and cross sectional studies. 
Table 10.7. Household toilet facilities among pregnant women in the cohort and cross sectional studies, 
Uttar Pradesh 
Cohort study Cross sectional study 
Type of toilet facility Intervention Comparison * Intervention Comparison * 
Flush toilet 7.7 4.4 7.8 6.4 
Open/shared/public 
Toilet 4.6 1.1 1.9 0.7 
No facility 85.9 94.3 89.3 92.8 
Other 1.6 0.2 1.1 0.1 
Total no. of pregnant 
women 504 453 746 706 
* Statistical test between districts, significant at p<0.05 
Drinking Water Source 
With the exception of the ~10% difference by district in mothers reporting household water 
coming from a public well, similar percentages were observed across other categories of drinking 
water source (Table 10.8). Greater than 80% of the pregnant women in the cohort study had 
access to piped water, either from a public tap or within their residence. Differences between 
intervention and comparison districts were significant in both the cohort and cross sectional 
studies. 
Other variables related to sanitation and hygiene beyond drinking water source and toilet type 
were considered in comparing the study samples. For example, the number of persons living in 
households differed significantly between districts –7.0 (±4.0) in the intervention district versus 
6.2 (±3.0) in the comparison district (P=0.001), though there was large variation in both districts. 
2/20/2008 127 
Table 10.8. Household drinking water source among pregnant women in the cohort and cross sectional 
studies, Uttar Pradesh 
Drinking water Cohort study Cross sectional study 
source type Intervention Comparison * Intervention Comparison * 
Piped water in 
residence 46.8 39.1 40.6 33.7 
Piped water in 
public tap 44.8 42.2 48.4 50.6 
Well water in 
residence 1.8 1.3 0.8 1.0 
Public Well 6.4 16.3 10.2 14.5 
Natural 
water/other 0 1.1 0 0.3 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts, significant at p<0.05 
Child Care for Mothers who Work Outside the Home 
At least two-thirds of pregnant women sampled in either study did not work outside the home, 
and this question did not have relevance for them (Table 10.9). Among the minority of women 
who did work outside the home, similar proportions in the cohort districts reported that their 
older children, mothers-in-law, or other relatives or neighbors took care of their youngest child 
for them. However, women in the comparison district were twice as likely to take their child to 
work with them compared to those in the intervention district. The trends in child care 
arrangements were similar in the cross-sectional study. Overall, the child care arrangements for 
women working outside the home were found to differ significantly between intervention and 
comparison in both the cohort and cross sectional groups, although the differences were not 
dramatic. 
2/20/2008 128 
Table 10.9. Child care arrangements among pregnant women in the cohort and cross sectional studies, 
Uttar Pradesh 
Child care Cohort study Cross sectional study 
arrangements Intervention Comparison * Intervention Comparison * 
Take child to 
work 5.0 11.9 6.7 11.5 
Husband 0.2 0.2 0.3 0 
Mother in law 7.1 5.5 6.3 5.7 
Older children 
13.3 13.0 10.9 7.5 
Other relatives/ 
neighbor 3.6 3.8 2.4 4.7 
No children yet 
70.8 65.6 73.5 70.7 
Total no. of 
pregnant women 504 453 746 706 
* Statistical test between districts, significant at p<0.05 
Socioeconomic Status Index 
As described for the cross-sectional study, the socioeconomic status index was derived using a 
principal components analysis (PCA) from data collected on household assets and possessions. 
This index may be a more optimal marker of the economic viability of households in developing 
countries than income data that is often difficult to collect accurately and can vary greatly over 
time. Monthly income from the past 6 months was ascertained in the cohort study, however, 
with no significant differences noted between districts (P=0.14). Respondents were stratified 
into terciles according to the SES index, and the distributions compared between intervention 
and comparison districts. Again, the lowest tercile indicates the poorest group and the highest 
tercile indicates the richest group. 
In the cohort sample, a lower proportion (~6 percentage points) of pregnant women were 
classified in the low index level and a higher proportion in the high index level in the 
intervention relative to the comparison district (Table 10.10). This difference in the distribution 
of the index between the districts was statistically significant. A similar difference in the 
distributions of the cross-sectional sample across the index was observed for the medium and 
high index levels, but the difference was not statistically significant in this case. 
2/20/2008 129 
Table 10.10. Socioeconomic Status (SES) Index among pregnant women in the cohort and 
cross sectional studies, Uttar Pradesh 
Cohort study Cross sectional study 
SES Index level Intervention Comparison * Intervention Comparison 
Low 30.4 36.6 33.4 33.3 
Medium 32.8 34.0 31.0 36.0 
High 36.8 29.4 35.7 30.7 
Total no. of 
pregnant women 453 504 746 706 
* Statistical test between districts, significant at p<0.05 
Summary
Although the distributions of numerous sociodemographic characteristics were statistically 
significantly different between the intervention and comparison districts in the cohort sample, 
most of these differences were quite small in fact. The large sample sizes led to even small 
differences yielding statistical significance. Evaluation analyses in the cohort data were run with 
and without multivariate adjustment for potential confounders. However, because this adjustment 
was not found to change the results, unadjusted statistical results are presented in the two 
chapters that follow. 
This chapter also provided a comparison of the sociodemographic characteristics of the cohort 
and cross-sectional samples of currently pregnant women from Uttar Pradesh. The results 
presented here demonstrate that the more geographically limited cohort sample is comparable to 
the more comprehensive community sample selected for the cross-sectional evaluation. 
Therefore, any results reported from the cohort sample should not be discounted because that 
sample is less representative of the population from which it was drawn. 
2/20/2008 130 
CHAPTER 11: INFANT FEEDING IN THE COHORT STUDY 
Breastfeeding 
Early initiation of breastfeeding
Mothers are advised to initiate breastfeeding within one hour of birth to protect against infection 
and hypothermia, but as well, to ensure early bonding and latching on for the newborn. This 
message is communicated during service provider contacts with mothers during the 3rd trimester 
of pregnancy and at delivery within the INHP-II basic package of nutrition interventions. 
Despite the fact that this exposure occurred early in the project’s intervention period for the 
cohort sample, there was significant improvement in timely initiation of breastfeeding. Mothers 
of infants less than one month of age were asked about the timing of initiation of breastfeeding 
relative to birth. There was a significant positive difference of 9 percentage points in mothers 
initiating breastfeeding within one hour, and a 37% difference for breastfeeding initiation within 
one day of birth in the intervention district compared to the comparison district (Figure 11.1). 
Figure 11.1. Breastfeeding initiation among mothers of children in the cohort sample, by district 
10
42
1 5
0
20
40
60
80
100
Within 1 hr Within 24 hr
Timing of initiation
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
Prelacteal feeds
Related to the timely initiation of breastfeeding, mothers were also advised against giving 
prelacteal feeds to their newborns, a common and risky practice in resource-poor settings. 
Mothers of infants less than one month were asked whether their infants were given any other 
liquids or solids before initiation of breastfeeding. Mothers were significantly less likely to give 
their infants prelacteal feeds in the intervention (79.8%) than in the comparison (98.0%) district, 
where this practice was nearly universal in the sampled population. 
Current breastfeeding 
No differences were found between districts for percentages of mothers reporting any 
breastfeeding at each follow-up visit in the cohort. During the first 12 months of life, ≥95% of 
mothers reported currently breastfeeding their infant. This proportion dropped to 90% among 
infants 12-17 months of age in both districts. By 18 months, 82.4% of mothers in the 
* *
2/20/2008 131 
intervention district and 85.3% of mothers in the comparison district reported still breastfeeding, 
with no statistically significant differences by districts noted. 
Exclusive breastfeeding and early introduction of complementary foods
Table 11.1 provides details on current breastfeeding practices by age and district. For infants at 
less than one month of age, mothers were asked about breastfeeding practices since the time of 
birth; at all subsequent follow-up visits, the information presented is based on recall of feeding 
practices in the previous 24 hours. Exclusive breastfeeding is recommended until 6 months of 
age, a behavior that could have been encouraged and reinforced through the INHP-II package in 
6 or more contacts throughout this period. At 3-5 months of age, there was a significantly higher 
percentage of exclusively breastfed infants in the intervention compared to the comparison 
district. Also striking was the high numbers of infants in both districts who were still receiving 
only breast milk at 6-8 months of age (to be discussed below). 
Table 11. 1 Current breastfeeding status among children in the cohort sample 0-20 months of age, by district 
Exclusive breastfeeding 
Breastfeeding 
& water 
Breastfeeding 
& other liquids/solids 
No 
Breastfeeding 
Age1 Inter. Comp. Inter. Comp. Inter. Comp. Inter. Comp. 
< 1 mo 70.3 65.0 5.5 8.8 23.7 25.3 0.6 1.0 
3-5 mo 58.1 46.8* 8.8 13.3 31.2 37.2 1.9 2.7 
6-8 mo 22.5 18.2 14.4 20.9* 59.6 58.0 3.5 3.0 
9-11 mo 0.3 0.6 15.2 15.0 79.5 81.2 5.1 3.2 
12-14 mo 0.0 0.8 8.8 4.9 84.5 86.9 6.7 7.4 
15-17 mo 0.0 0.5 4.5 2.3 87.2 87.4 8.3 9.9 
18-20 mo 0.3 0.0 2.4 0.8 79.7 84.5 17.6 14.7 1
Based on 24 hr recall for at all ages, except at <1 mo when mothers reported breastfeeding status since birth. 
*Statistical test for differences by district, significant at p<0.05 
At 0-6 months, there were no significant differences in the percentages of infants receiving 
water, or other liquids and solids in addition to breast milk. However, there was a trend towards 
smaller percentages of mothers reporting these feeding practices in the intervention district. 
Early introduction of complementary foods results in increased risks of infectious disease and 
growth faltering during this the first half of infancy. During the first month of life, approximately 
one-quarter of the infants were reported to have already received other liquids and solids along 
with breast milk in both districts. This proportion increased to approximately one-third of infants 
at 3-5 months of age. Other cultural and economic barriers to the acceptance of the exclusive 
breastfeeding behavior in this population should be explored and incorporated into programming 
for improved nutrition in early infancy. 
2/20/2008 132 
Complementary Feeding 
At least one contact with mothers was expected in the INHP-II program for delivering nutrition 
interventions in each of the infancy periods of 6-8 months and 9-11 months, and two or more 
were expected during the period when children were 12-24 months of age. Messages based on 
age-specific complementary feeding guidelines were conveyed concerning quantity, frequency, 
variety, and feeding behaviors. Children ages 6-24 months living in resource-poor settings are 
most vulnerable to diarrheal and other infectious diseases morbidities together with the growth 
deficits associated with inadequate complementary feeding practices. It is a critically important 
period that has tremendous complexities for programming based on differences across settings 
for cultural practices, food availabilities, and economic conditions. 
Introduction of solids
At 6 months, breastmilk is insufficient in the infant diet in terms of both energy and particular 
nutrients to fulfill daily requirements for the breastfeeding baby – thus necessitating the 
introduction of complementary foods. The timing of this introduction was evaluated in two ways 
in our data: by report of current intake of solids at the 6 month follow-up visit and by recall at the 
subsequent 9-month visit. Among the infants at 6-8 months, 42.0% in the intervention and 37.1% 
in the comparison district reported that their infant had received solids in the previous day. By 
contrast, 69.4% of mothers in the intervention and 72.2% in the comparison district reported at 
the 9-month visit that they introduced solids to their infant between 6 and 8 months of age. This 
discrepancy between recall of the timing of introduction of solids and reported current intake was 
observed in the cross-sectional survey data as well. 
Dietary intakes by 24 hour recall 
Information was collected on the dietary intakes of cohort infants using 24-hr recall methods 
beginning at the 9-month and all subsequent follow-up visits. Mothers were asked to report what 
their child had consumed in the previous one-day period. The type of food consumed, as well as 
the frequency of its consumption and portion sizes were recorded. 
At 9-11 months of age, only 40-50% of breastfed infants had consumed cereals (rice, bread, or 
other grains) in the previous 24 hours, and the proportion was significantly smaller in the 
intervention than the comparison district (Figure 11.2). By one year of age, three-quarters of 
infants had consumed cereal in the previous 24-hour period, and by 15-17 months, cereal intake 
was approximately 90% in both districts. 
2/20/2008 133 
Figure 11.2. Cereal consumption in the past 24 hours among breastfeeding children in the cohort sample 
from 9 to 20 months, by district 
42
76
91 92
51
75
90 93
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
The INHP-II promoted the addition and increase of legumes in the diets of infants and young 
children to improve both energy and nutrition content of primarily cereal-based diets. The 
cohort data suggest that this infant feeding message was received and followed by mothers in the 
intervention district. There was an increasing disparity in the proportion of children consuming 
legumes between the districts through 18-20 months of age. At 9-11 months, a greater percentage 
of the intervention infants than the comparison infants had received legumes with this difference 
growing and becoming more significant by 12-14 months and continuing through 18-20 months. 
Figure 11.3. Legume consumption in the past 24 hours among breastfeeding children in the cohort sample 
from 9 to 20 months, by district 
42
73
83 84
35
53 50 53
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
Consumption of fruits and vegetables is encouraged to improve the quality of diets in terms of 
micronutrients such as vitamin A. At all four follow-up visits from 9-18 months of age when 24-
hour dietary recall data were collected, intake of fruits and vegetables was significantly greater 
*
* * *
2/20/2008 134 
among children in the comparison district (Figure 11.4). It is unclear whether this difference 
reflects food availability or dietary preferences. 
Figure 11.4. Fruit and vegetable consumption in the past 24 hours among breastfeeding children in the cohort 
sample from 9 to 20 months, by district 
13
27 33
56
23
40
52
67
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
A very small number of children from both districts were reported to have consumed eggs, meat, 
chicken, or fish (Figure 11.5). By 15-17 months, the percentage reached 1% in both districts and 
by 18-20 months, this percentage declined again slightly in the intervention district and increased 
to 3.7% in the comparison district, attaining a significant difference. Breastmilk continues to be 
an important source of both energy and micronutrients from 6-20 months, but particular nutrients 
have been singled out as problem nutrients due to their low availability in breastmilk and local 
diets. Iron and zinc, for example, are both found in higher concentrations and in more 
bioavailable forms in meats and fish when compared to other plant-based foods. 
Figure 11.5. Consumption of eggs, meat, chicken or fish in the previous 24 hours among breastfeeding 
children in the cohort sample from 9 to 20 months, district 
0.3 00 0.7 1 2 0.8 4
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
*
* *
*
*
2/20/2008 135 
Although most children are breastfed for the first 2 years of life, it is a common practice to 
supplement breastmilk with animal or powdered milk or infant formula. Mothers received advice 
from service providers to continue breastfeeding and limit this practice. Approximately 30% of 
breastfed children consumed other milks at 9-11 months of age in both districts (Figure 11.6). 
However, this proportion decreased among children in the intervention at subsequent visits in the 
second year of life and was statistically lower than the proportion among children in the 
comparison district at these visits. 
Figure 11.6. Animal/powdered milk and formula intake in the past 24 hours among breastfeeding children in 
the cohort sample 9 to 20 months of age, by district 
29
20 19
28 33 30 33 38
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
 * Statistical test for differences by district, significant at p<0.05 
Quantity of solids consumed
WHO provides an estimated minimum quantity of solids recommended for the breastfeeding 
young child as follows: 200 g/d for infants 6-8 mo; 300 g/d for infants 9-11 mo; and 400 g/d for 
children 12-23 months of age. Analyses for the cohort study, similar to the cross-sectional 
analysis, defined solids as rice, bread, and grain products – observed to be the primary sources of 
energy in the infant and young child diet. Because so few children were found to consume 
adequate amounts, the definition was refined to be the “proportion who consumed at least half 
the recommended quantity of solids per day.” Table 2 gives the percentage of breastfeeding 
children meeting the requirements of 150 or more grams per day for 9-11 months, and 200 or 
more grams per day for those in the second year of life – one-half of requirements. Less than 3% 
of children across all age categories were found to meet this very minimal level of solids 
consumption. There were no differences between intervention and comparison groups. 
* * *
2/20/2008 136 
Table 11.2. Consumption of at least half the recommended quantity of solids among children in the cohort 
sample at follow-up from 9 to 20 months of age, by district 
Half the recommended quantity of solids (%) 
Age (months) Intervention Comparison 
9-11 (150+ g) 0.8 1.0 
12-14 (200+ g) 0.3 0 
15-17 (200+ g) 1.7 1.4 
18-20 (200+ g) 2.4 2.8 
Frequency of solids consumed
The WHO recommendation on feeding frequency was also examined in the cohort study. Using 
the 24-hour recall, the number of times solids – similarly defined to include rice, bread, and grain 
products – was calculated and summarized as the percentages of breastfeeding children meeting 
the WHO recommendations (Figure 7). A minimum of 2 solid feeds per day is specified for 
children 6-8 months and 3 solid feeds per day for children 9-23 months. The intervention district 
was shown to be significantly higher than the comparison district across all ages, significant after 
12 months. 
Figure 11.7. Minimum recommended frequency of solids consumed per day among breastfeeding children in 
the cohort sample at follow-up from 9 to 20 months of age, by district 
9
25
66
8
13
27
46
46
0 20 40 60 80
9-11 mo
12-14 mo
15-17 mo
18-20 mo
Round
Percent (%)
Comparison
Intervention
 * Statistical test for differences by district, significant at p<0.05 
Dietary diversity
Increasing the variety of foods in an infant diet is generally recommended to improve the quality 
or micronutrient content. Dietary diversity was described by first defining food groups and 
quantifying the number of groups a child consumed in the previous 24 hours. Foods were 
grouped as: cereals; legumes; meat and eggs; vitamin-A rich fruits and vegetables; other fruit 
and vegetables; dairy foods; and butter, oil, and fats. Adequate dietary diversity was defined as 3 
*
*
*
2/20/2008 137 
or more food groups consumed in the previous 24 hours. While the percent of breastfeeding 
children found to have adequate dietary diversity was higher in the comparison district at all age 
categories, no significant differences between intervention and comparison groups were found. 
Figure 11.8. Consumption of ≥3 food groups in the past 24 hours among children in the cohort sample at 
follow-up from 9 to 20 months of age, by district 
17
29 31
41
22
32 37
45
0
10
20
30
40
50
60
70
80
90
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Age groups
Intervention
Comparison
To increase energy density of foods and the bioavailability of particular micronutrients, 
consumption of oil with solids is recommended in young children. The percentages of 
breastfeeding children reported receiving oil with one or more solid foods in this study was low 
in both intervention and comparison districts (Table 11.3). In the 18-20 month age group, a 
statistically significant difference of 10 percentage points was found between the intervention 
and comparison district, with a greater proportion of children receiving oil with solids in the 
latter. 
Table 11.3. Consumption of oil added to solids among breastfeeding children in the cohort sample at follow￾up from 9 to 20 months of age, by district 
Oil added to solids (%) 
Age (months) Intervention Comparison 
9-11 1.2 0.3 
12-14 0.3 1.3 
15-17 3.0 6.5 
18-20 1.7 11.7* 
 * Statistical test for differences by district, significant at p<0.05 
2/20/2008 138 
Supplementary Nutrition 
There were large differences between intervention and comparison districts noted in the 
reporting of supplementary nutrition, AWC food, ever received (Table 11.4). Those in the 
intervention district were far more likely ever to have received supplementary nutrition in the 
past than in the comparison district. There appeared to be greater proportions receiving the 
supplementary nutrition earlier in the second year of life in the intervention district - with smaller 
incremental changes in the 15-17 months and 18-20 months periods. In the comparison district, 
a larger increase was observed later in the 18-20 age group. 
Table 11.4. Receipt of Supplementary Nutrition among children in the cohort sample at follow-up from 12 to 
20 months of age, by district 
Receipt of Supplementary Nutrition (%) 
Age (months) Intervention Comparison 
12-14 45.0 5.3* 
15-17 51.2 6.4* 
18-20 56.1 23.9* 
 * Statistical test for differences by district, significant at p<0.05 
Caregiver feeding practices 
Caregiver feeding
Caregiving practices are considered important determinants in infant and young child growth. 
Variability in nutritional outcomes might be expected based on who is actually feeding the child. 
For this study, the percentage of children fed by mother, as a positive practice for promoting 
growth, was investigated (Figure 11.9). The proportion of children usually fed by the mother 
was greater in the intervention than the comparison district at follow-up visits from 9 to 18 
months. By contrast, higher percentages of children in the comparison district had no feeding 
assistance at all, considered a riskier practice for promoting growth (Figure 11.10). 
Figure 11.9. Child feeding by mothers among children in the cohort sample at follow-up from 9 to 20 months 
of age, by district 
65 64 70
60
51 50
41 35
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
* Statistical test for differences by district, Pearson chi-square significant at p<0.05 
* * *
2/20/2008 139 
Figure 11.10. Child feeding without assistance among children in the cohort sample at follow-up from 9 to 20 
months of age, by district 
14
5 4
24
18
9 14
41
0
20
40
60
80
100
9-11 mo 12-14 mo 15-17 mo 18-20 mo
Round
Intervention
Comparison
* Statistical test for differences by district, Pearson chi-square significant at p<0.05 
Plate sharing
Children fed from a separate platter may be protected against inadequate food intakes resulting 
from competition for foods with older siblings or caregivers. Infants and young children 
receiving food from a shared platter may be incapable of taking adequate quantities or potentially 
benefit because of the attendance of a caregiver to its needs while the family eats. Interestingly, 
greater percentages of infants were fed from a separate platter in the comparison district 
compared to the intervention district despite the messages communicated by INHP-II during 
contacts from 6-24 months to feed the child from a separate plate (Table 11.5). Except in the 9-
11 month age group, significantly lower proportions of mothers in the intervention district 
reported following this practice of feeding from a separate platter. 
Table 11.5. Plate sharing practices among children in the cohort sample at follow-up from 9 to 20 months, by 
district 
Fed from a separate plate (%) Fed from a shared plate (%) 
Age (months) Intervention Comparison Intervention Comparison 
9-11 18.7 13.6 68.5 71.7 
12-14 7.6 26.5* 90.2 70.6* 
15-17 9.3 30.9* 90.7 69.1* 
18-20 30.5 53.2* 69.5 46.8* 
* Statistical test for differences by district, significant at p<0.05 
* *
2/20/2008 140 
Summary
Despite the occurrence of cohort deliveries relatively early on in the 2-year INHP-II intervention 
period, advice about desired newborn feeding practices appears to have been moderately 
effective at improving indicators of these practices in the cohort intervention district. Both an 
increase in early initiation of breastfeeding and a decrease in use of prelacteal feeds were 
observed among intervention district mothers. 
Improvements in breastfeeding practices during the first 6 months of life were limited. Exclusive 
breastfeeding was significantly more prevalent among children at 3-5 months of age in the 
intervention than comparison district. However, there was no reduction in early introduction of 
complementary foods that was reported in both districts by approximately one-third of mothers 
when their children were 3-5 months of age. 
There were limited improvements in dietary intake and complementary feeding indicators in the 
cohort sample. Timely introduction of solids did not differ in the two districts when assessed by 
recall or by current intake. Consumption of cereals, fruits and vegetables increased from 9 to 18 
months, but the increases were comparable in the intervention and comparison districts. This 
finding suggests that the ICDS program’s efforts to improve child feeding practices may be 
having some success. By contrast, legume intake approximately doubled from 9 to 15 months in 
the intervention district, and was significantly greater than in the comparison district at 12 
through 18 months of age. Further, intake of powdered/animal milk and infant formula reduced 
from 9 to 12 months in the intervention district and was significantly lower than in the 
comparison district from 12 to 18 months of age. These positive changes in specific child feeding 
practices indicate that the INHP-II intervention has likely been successful at communicating 
clear, specific messages about increasing legume intake and reducing animal/powdered milk 
intake during the weaning period. 
Among indicators of complementary feeding, there was significant improvement only in the 
proportion of children consuming the minimum recommended number of solid feeds per day. 
There was an 8-fold increase in this proportion among intervention district children relative to a 
less dramatic increase among comparison district children from 9 to 18 months of age. 
Nonetheless, <25% of children at 12 months and <50% at 15 months received the minimum 
recommended number of feeds. Indicators of adequate quantity and variety of intake did not 
differ between the districts, and the proportion consuming at least half the recommended 
minimum quantity was less than 3% at all ages in the study cohort. Although the positive 
intervention effects on child feeding practices observed in the cohort sample are encouraging, the 
low proportions of children who consume adequate quantity and quality of complementary foods 
with sufficient frequency in this population indicate that further dietary improvements are needed 
before an impact on malnutrition prevalence can be expected. 
2/20/2008 141 
CHAPTER 12: CHILD NUTRITIONAL STATUS AND PHYSICAL GROWTH IN THE 
COHORT STUDY 
Children were weighed and measured prostrate for length at each follow-up visit from 7 days 
postpartum through 18 months of age in the cohort study. Anthropometric data from cohort 
children reported for each age represents that age plus and minus one month, e.g. 3 month data 
represents data from children at 2-4 months of age. The exception is data from the initial 
postpartum visit that includes data from infants <1 month of age only. Data on weight and 
length were used to calculate the indices of weight-for-age (WAZ), height-for-age (HAZ), and 
weight-for-height (WHZ). The Z-scores represent standardized distances from a reference 
median (WHO, 1979). While length measures may be referred to as length-for-age or weight￾for-length, for consistency purposes, they are hereafter referred to as height. Undernutrition is 
defined as WAZ<-2 SD, stunting as HAZ<-2 SD, and wasting as WHZ<-2 SD. Children with Z 
scores less than -3 SD for each of the indices were considered severely underweight, stunted, and 
wasted respectively. Those with Z scores less than -5 SD were excluded from these analyses. 
Underweight 
Weight-for-age Z score reflects a combined effect of both short-term, acute undernutrition and 
longer-term, chronic undernutrition in infants and young children. This index does not 
incorporate consideration of height and may produce bias if not also used in combination with 
the indices of HAZ and WHZ. In population surveys, WAZ, HAZ, and WHZ generally follow a 
normal distribution. This was the case with the children in the cohort sample, and therefore, the 
Z scores are presented as mean values. 
A significantly higher mean WAZ was found in the intervention district relative to the 
comparison district at 6 and 9 months of age (Table 12.1). In all other rounds, children from the 
intervention district showed a non-significant trend for higher Z scores than the comparison 
district. When stratified by sex, the same trend was observed among both girls and boys, but the 
WAZ differed significantly between districts for boys only at 9 months of age. In the cross￾sectional study, boys from the intervention district also appeared to have a slight advantage in 
maintaining WAZ compared to boys from the comparison district and girls in both districts. 
2/20/2008 142 
Table 12.1. Mean weight-for-age Z score (WAZ) among children in the cohort sample at follow-up from 0-18 
months, by district 
Mean WAZ 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 - 0.73 -0.75 -0.78 -0.87 -0.66 -0.61 
3 -0.70 -0.73 -0.73 -0.69 -0.66 -0.77 
6 -1.37* -1.56* -1.57 -1.39 -1.35 -1.55 
9 -2.17* -2.37* -2.24* -2.46* -2.10 -2.25 
12 -2.52 -2.62 -2.62 -2.66 -2.42 -2.57 
15 -2.32 -2.34 -2.34 -2.36 -2.30 -2.31 
18 -2.33 -2.34 -2.35 -2.34 -2.31 -2.34 
 * Student t test comparing means by district, significant at p<0.05. 
Figure 12.1 illustrates the divergence of mean WAZ between the districts. A very small increase 
in WAZ was observed in both districts from the initial postpartum visit to the subsequent visit at 
3 months. From 3 to 12 months, mean WAZ in the comparison district declined slightly more 
precipitously than in the intervention district. From 12 to 15 months of age, the trend reverses 
direction, giving slightly improved WAZ in both districts. By the final follow-up visit at 18 
months of age, mean WAZ were nearly equal in the 2 districts. It is important to note that the 
mean WAZ did not rise above -0.5 in either district for the entire cohort study period. 
Figure 12.1. Mean WAZ among children in the cohort sample at follow-up from 0-18 months, by district 
-3
-2.5
-2
-1.5
-1
-0.5
0
< 1 3 6 9 12 15 18
Age (months)
Weight-for-age Z score
Intervention
Comparison
* Student t test comparing means by district, significant at p<0.05.
*
*
2/20/2008 143 
Table 12.2 gives the proportion of children who were classified as underweight (WAZ <-2 SD). 
During the first 6 months postpartum, the prevalence of underweight was more modest, but more 
than half of the children in both districts were underweight after 6 months of age. The proportion 
underweight was lower in the intervention than the comparison district from 3 through 12 
months, but the difference was statistically significant at 6 months only. This same trend was 
observed for both boys (p<0.05 at 6 mo) and girls (p<0.05 at 12 mo). Underweight prevalence 
peaked at nearly 80% at 12 months and then dropped back to ~65% at 15-18 months in the 
comparison district. 
Table 12.2. Underweight (% <-2SD WAZ) among children in the cohort sample at follow-up from 0-18 
months, by district 
WAZ < -2SD (%) 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 12.3 13.6 15.7 20.4 8.4 5.1 
3 6.0 8.9 4.9 8.7 7.3 9.0 
6 21.1 29.5* 21.0 30.1* 21.3 28.7 
9 62.1 65.1 66.7 68.3 57.3 61.2 
12 72.7 78.0 77.9 77.3 67.6 78.9* 
15 68.8 65.2 71.0 67.3 66.4 62.8 
18 63.6 63.0 66.9 62.9 60.1 63.1 
 * Pearson Χ2
, significant at p<0.05 
The 6-12 month period - when complementary foods are needed to optimize growth and 
infectious diseases associated with malnutrition (e.g. diarrhea) are highly prevalent - is known to 
be a particularly vulnerable phase for growth faltering. In both districts of the cohort study, there 
were very large increases in the proportion underweight from 6 to 12 months worth noting. 
Nonetheless, there was some evidence of a protective effect of the intervention on underweight 
during this period. 
Stunting 
The height-for-age index is used as a marker of linear growth and thought to reflect longer-term 
nutritional exposures. Similar to the findings from the cross sectional analysis, little or no 
intervention impact was demonstrated for improving height or diminishing losses in linear 
growth (Table 12.3). However, at the first postpartum visit during the neonatal period, children 
in the intervention district were significantly shorter than those in the comparison district. The 
reason for this initial difference is unclear, but it was erased by the subsequent follow-up visit at 
3 months of age and no further height differences were observed for the duration of the study 
follow-up. 
2/20/2008 144 
Table 12.3. Mean height-for-age Z score (HAZ) among children in the cohort sample at follow-up from 0-18 
months, by district 
Mean HAZ 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 -0.40* -0.06* -0.38* -0.14* -0.42* 0.04* 
3 -0.81 -0.78 -0.82 -0.79 0.80 0.78 
6 -1.35 -1.35 -1.39 -1.47 -1.30 -1.25 
9 -1.86 -1.86 -1.92 -2.01 -1.79 -1.67 
12 -2.25 -2.18 -2.35 -2.28 -2.14 -2.05 
15 -2.34 -2.33 -2.37 -2.34 -2.31 -2.31 
18 -2.32 -2.30 -2.28 -2.27 -2.32 -2.38 
 * Student t test comparing means by district, significant at p<0.05. 
As viewed in Figure 12.2, the mean HAZ in the 2 districts converged at 3 months of age and 
remained almost identical for the remainder of the follow-up visits. If the difference in mean 
HAZ between the 2 districts at the initial postpartum visit is taken as real, then the decline in the 
intervention district is less than that observed in the comparison district. Around 12 months, 
there is a leveling off of HAZ at approximately -2.3 in both districts. Again, the HAZ did not 
rise above the reference median at any age from birth to 18 months. 
Figure 12.2. Mean height-for-age Z score (HAZ) among children in the cohort sample at follow-up from 0-18 
months, by district 
-2.5
-2
-1.5
-1
-0.5
0
< 1 3 6 9 12 15 18
Age (months)
Height-for-age Z score
Intervention
Comparison
* Student t test comparing means by district, significant at p<0.05.
*
2/20/2008 145 
No statistically significant differences between the intervention and comparison districts were 
observed for prevalence of stunting (HAZ < -2 SD) in the cohort study among all children. An 
approximately 40-50 percentage point rise in the proportion stunted in both districts occurred 
between 3 and 12 months during the vulnerable complementary feeding period. Among all 
children, stunting prevalence was highest at 15-18 months. Disaggregating the sample by sex 
did not produce any significant differences between intervention and comparison groups for 
stunting prevalence with the exception of girls at the initial postpartum visit. The apparent 
length advantage in the comparison district appears to be driven by girls according to stunting 
prevalence. 
Table 12.4. Stunting (% <-2SD HAZ) among children in the cohort sample at follow-up from 0-18 months, 
by district 
HAZ < -2SD (%) 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 6.5 4.0 5.6 5.7 7.5* 1.9* 
3 12.0 11.7 10.7 15.1 13.4 7.6 
6 21.2 26.4 22.9 30.8 19.3 20.7 
9 42.5 42.4 46.6 50.3 38.0 32.8 
12 59.8 55.7 63.5 60.6 56.1 49.5 
15 62.0 63.6 65.2 63.7 58.7 63.4 
18 60.9 61.4 61.0 59.3 60.8 64.0 
 * Pearson Χ2
, significant at p<0.05 
Wasting 
In contrast to the other two indices of WAZ and HAZ, the WHZ does not use chronological age 
to compare with the reference population. Instead, weight or body mass is examined relative to 
the height of individuals. This index, used to signify wasting or thinness, may be more 
responsive to short-term interventions or environmental factors influencing nutritional status. 
The cross-sectional analysis found the increases in wasting from baseline to endline were 
significantly reduced in the intervention versus comparison districts of Uttar Pradesh. 
At 4 of the 7 follow-up visits, mean WHZ was significantly higher in the intervention relative to 
the comparison district, and a trend for higher average WHZ in the other rounds was found as 
well (Table 12.5). Both boys and girls in the intervention district had significantly higher 
average WHZ at the initial postpartum visit and at 9 months. Additionally, girls in the 
intervention district had statistically higher mean WHZ at 6 and 12 months as well. Nonetheless, 
these improvements in the intervention district disappeared for all children by 15 months of age. 
2/20/2008 146 
Table 12.5. Mean weight-for-height Z score (WHZ) among children in the cohort sample at follow-up from 
0-18 months, by district 
Mean WHZ 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 -0.74 -1.00* -0.69 -0.90* -0.81 -1.12* 
3 -0.14 -0.20 -0.19 -0.18 -0.07 -0.23 
6 -0.43 -0.66* -0.66 -0.51 0.34 -0.66* 
9 -0.94 -1.23* -1.05 -1.26* -0.84 -1.19* 
12 -1.24 -1.47* -1.32 -1.46 -1.17 -1.49* 
15 -1.20 -1.25 -1.21 -1.28 -1.18 -1.22 
18 -1.47 -1.49 -1.50 -1.54 -1.43 -1.43 
 * Student t test comparing means by district, significant at p<0.05. 
In contrast to Figures 12.1 and 12,2 for mean WAZ and HAZ, respectively, WHZ values 
increased in the first 3 months of life before following the typical declining trend until 12 months 
(Figure 12.3). Between 12 and 15 months, average WHZ improved slightly, only to falter again 
between 15 and 18 months. The intervention district WHZ remained above that of the 
comparison district throughout the study period – with greatest distances between intervention 
and comparison observed at 9 and 12 months. 
Figure 12.3. Mean weight-for-height Z score (WHZ) among children in the cohort sample at follow-up from 
0-18 months, by district 
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
< 1 3 6 9 12 15 18
Age (months)
Weight-for-height Z score
Intervention
Comparison
* Student t test comparing means by district, significant at p<0.05. 
*
*
*
2/20/2008 147 
Figures 12.4 and 12.5 show average WHZ for boys and girls, respectively. The most apparent 
difference in these two figures is that average WHZ in the girls from the intervention district was 
above those observed in the girls from the comparison district throughout the study period, with 
a sizeable disparity observed at 6 months. At that age, the WHZ indicator among girls in the 
intervention district actually rose above the reference median only to fall precipitously at 9 
months by 1.18 SD; WHZ of girls in the comparison district fell by 0.53 SD between those two 
ages. The differences by district observed for girls in the WHZ may be in part explained by 
lower average HAZ among girls observed from the initial postpartum visit to 12 months of age in 
the intervention district compared to the comparison district. Boys showed fewer differences by 
district in the cohort analysis of mean WHZ. 
Figure 12.4. Mean weight-for-height Z score (WHZ) among boys in the cohort sample at follow-up from 0-18 
months, by district 
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
< 1 3 6 9 12 15 18
Age (months)
Weight-for-height Z score
Intervention
Comparison
* Student t test comparing means by district, significant at p<0.05. 
*
*
2/20/2008 148 
Figure 12.5. Mean weight-for-height Z score (WHZ) among girls in the cohort sample at follow-up from 0-18 
months, by district 
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
< 1 3 6 9 12 15 18
Age (months)
Weight-for-height Z score
Intervention
Comparison
* Student t test comparing means by district, significant at p<0.05. 
The proportion of children classified as wasted in the cohort study was very high in the non￾emergency context of both districts. In the neonatal period, there were relatively high rates of 
wasting at the initial postpartum visit that were reduced at 3 and 6 months of the breastfeeding 
period (Table 12.6). From 9 to 18 months, prevalence of wasting was above 10% and by 18 
months, over one-fourth of the children were wasted. The intervention district had significantly 
lower proportions wasted at 6 and 12 months. No differences were observed among the boys by 
district, and among the girls, there was significantly lower prevalence of wasting in round 6 
months for the intervention district compared to the comparison district. 
*
*
*
*
2/20/2008 149 
Table 12.6. Wasting (% <-2SD WHZ) among children in the cohort sample at follow-up from 0-18 months, 
by district 
WHZ < -2SD (%) 
All children Boys Girls 
Age (mo) Intervention Comparison Intervention Comparison Intervention Comparison 
<1 5.2 8.5 3.7 6.8 7.0 10.6 
3 1.6 2.5 2.0 1.7 1.1 3.5 
6 2.6 6.5* 3.4 6.4 1.7 6.7* 
9 10.7 15.3 18.0 14.2 7.0 11.9 
12 16.6 24.3* 18.2 25.8 14.9 22.3 
15 15.1 15.1 15.9 16.8 14.3 12.9 
18 27.3 28.1 31.9 31.5 23.1 23.4 
 * Pearson Χ2
, significant at p<0.05 
Longitudinal modeling of physical growth 
The previous sections have presented standardized measures of weight and height, centered in 
each age group. Growth patterns will now be considered longitudinally to examine the impact of 
the intervention on physical growth. A cubic polynomial equation has been applied to account 
for a declining rate of growth from 0-6 months that plateaus between 6 and 18 months, and 
begins to increase thereafter from 18-24 months. Interaction terms for district and age represent 
the growth differences between intervention and comparison districts. 
Among all children, infants from the intervention district showed a 0.017 kg increase in weight 
in one month’s time compared to infants in the comparison district (p<0.06), adjusting for 
confounding factors. This translates into a 0.199 kg increase over 12 months time. After 
stratifying by sex, the increase in weight among boys from the intervention district compared to 
the comparison district remained (0.011 kg/m), but was not significant. In the girls from the 
intervention district, however, there was a 0.026 kg/m increase in weight compared to the girls 
from the comparison district (P<0.05), and 0.307 kg for the entire year observed. After 
consideration of a significant negative quadratic age term – suggesting a declining rate of weight 
increases over 12 months - in the longitudinal model for girls, the total intervention effect for 
weight was 0.025 kg/m and 0.294 kg/yr. 
A significant increase in height was found for all infants from the intervention district compared 
to the comparison district, 0.052 cm/months (P=0.05). Even with a 12-month accumulation 
calculated at 0.619 cm/yr, the difference in height for the children was minimal, albeit 
statistically significant. When longitudinal modeling was carried out for boys and girls 
separately, those from the intervention district still showed greater increases in height – 0.043 
cm/m for boys and 0.069 cm/m for girls – though the effect was no longer statistically significant 
when disaggregated by sex. Similar to weight, girls from the intervention district demonstrated 
2/20/2008 150 
the greater growth rate advantage compared to the increase noted for boys. Over one year’s 
time, the heights of girls and boys from the intervention district were 0.82 and 0.51 cm greater, 
respectively, than those from the comparison district. 
Thus, after controlling for confounding variables maternal age and education, longitudinal 
modeling showed a small, positive intervention effect for height (p=0.05) and a trend toward a 
positive intervention effect for weight (p=0.06). Only weight gain in girls retained significance 
when modeling growth separately by sex. 
Summary 
High rates of childhood malnutrition were found in both districts of the cohort study conducted 
in Uttar Pradesh. For all nutritional status indicators, there was a clear trend toward dramatically 
worsening status between 3 and 12 months of age. However, there was some limited evidence 
that nutritional status and physical growth among children less than 2 years of age were 
improved significantly in the intervention relative to the comparison district. Mean WHZ and 
wasting showed the greatest differences between districts throughout the study period. Children 
from the intervention district had consistently larger average WHZ than those in the comparison 
district, with significant differences at four of the seven follow-up visits. There was an important 
clustering of significant differences in WHZ and wasting at 6 through 12 months of age - a 
period when children are considered at high-risk for growth faltering. After stratifying by sex, 
the largest improvements attributable to the intervention during the latter half of infancy were 
observed among girls. 
There was a consistent trend for higher average WAZ among children in the intervention relative 
to the comparison district during the first year of life that was statistically significant at 6 and 9 
months only. Underweight was significantly lower among intervention district children at 6 
months only. These trends were similar for boys and girls except that the intervention benefits 
appeared to extend to 12 months among girls only. With the exception of an uninterpretable 
difference at the initial postpartum visit during the neonatal period, no differences in HAZ and 
stunting prevalence were observed between the intervention and comparison districts. 
At 6 months of age, it is recommended that infants are fed solid foods in addition to breast milk 
to increase the caloric and micronutrient density of the diets. This complementary feeding 
period, usually lasting from 6 to 24 months, may be a particularly vulnerable period for 
malnutrition and growth faltering. The risks arise from inadequate feeding practices, among 
other causes (e.g. infectious diseases), that were targeted by the CARE program. Some program 
effectiveness in the cohort analysis may be surmised from the response documented in the WHZ 
index, though additional investigation of program interventions in relation to outcomes is 
needed. 
The longitudinal modeling of physical growth outcomes took into consideration confounding 
factors and assessed differences in the rate of weight and height gains between the districts. 
Small positive intervention effects were found, with a slight advantage demonstrated for girls. 
Infants from the intervention district grew just over one-half of a cm more over 12 months (0.619 
cm/yr; P=0.05) compared to those in the comparison district. Weight gain in one year was about 
2/20/2008 151 
one-fifth of a kilogram (0.199 kg; P=0.06) higher in the intervention district compared to the 
comparison district. This effect was statistically significant for weight gain in girls only. 
Despite the positive intervention impacts, rates of underweight, stunting, and wasting reached 
unacceptable levels in these populations. Rapid ascents in malnutrition prevalences were 
especially observed in the 6-12 month period - with probable lifetime consequences. Wasting 
that exceeds 5% in a population signifies a potential crisis situation; in this sample of children, 
wasting reached and surpassed 25%. Over 60% of the children by 18 months of age were 
classified as stunted and underweight. Additional program and policy attention is obviously 
deserved. 
2/20/2008 152 
CHAPTER 13: SUMMARY FINDINGS AND RECOMMENDATIONS FOR FURTHER 
ANALYSES 
Summary findings 
Antenatal care
• Receipt of ANC visits was universal in AP. INHP-II significantly increased ANC 
coverage in UP so that 50% of women reported at least one visit and 25% reported ≥3 
visits during the most recent pregnancy. 
• INHP-II was highly effective in both states at increasing outreach by service providers to 
provide ANC services and health and nutrition advice to pregnant women through home 
visits and contacts in the community. Approximately 50% of AP women and 60% of UP 
women reported at least one home visit by a service provider during the most recent 
pregnancy. Home visits by AWW increased most dramatically in both states. However, 
ANM and CA also increased their outreach significantly in UP. Service provider contacts 
at the Anganwadi Centers and Nutrition Health Days increased significantly also. 
• Improvements in coverage and consumption of antenatal IFA supplementation were 
limited in AP. However, UP state recorded significant increases in the proportion of 
women who received (65%) and consumed (36%) ≥90 tablets during their most recent 
pregnancy. In both states, the proportion of women who consumed all the IFA tablets 
they received more than doubled by endline. 
• Recent attendance at NHD increased significantly among pregnant women and mothers 
of children 0-23 months in both states. Coverage during the past 3 months was only 
~15% in AP, but 34-39% in UP at endline. However, the proportion of women in both 
states who remain unaware of NHD is still too high. 
Maternal nutrition
• Receipt of SN among pregnant and postpartum women increased significantly to ~50% in 
both states. At endline, consumption of SN was more common among intervention 
district than comparison district women also. In the recall of dietary intake during the 
past 7 days, SN intake was significantly increased among all groups of women in UP and 
among pregnant women only in AP. 
• Consumption of at least 3 meals a day increased but snack consumption decreased in AP 
state where the vast majority of women already consumed 3 meals a day. In UP state, 
only 25% of women ate 3 meals a day at baseline, but both meal and snack consumption 
increased at endline. In both states, these changes in meal and snack consumption 
occurred among pregnant and recently delivered women as well as mothers of children 6-
23 months of age. 
• There were multiple improvements in women’s diets attributable to INHP-II in both 
states. Recent consumption of legumes, dark green leafy vegetables, and yellow-orange 
fruits was increased among all groups of women in AP. In UP, only yellow-orange fruit 
2/20/2008 153 
intake increased significantly among all women, but pregnant women and mothers of 
children 6-23 months increased their weekly intake of meat, fish, chicken and eggs and 
recently delivered women increased their dairy intake. Dietary messages delivered to 
women during pregnancy and the early postpartum period appeared to have a positive 
effect on all women’s diets in many cases. 
• Malnutrition was very high among pregnant and non-pregnant women in both states, but 
it was particularly alarming in the AP intervention district where over 50% of mothers of 
children 6-23 months were malnourished. However, malnutrition prevalence decreased 
significantly among first trimester pregnant women and mothers of children 6-23 months 
of age in AP. In UP, malnutrition decreased among third trimester pregnant women only. 
The increase in BMI of ~1 unit among the first (AP) and third (UP) trimester pregnant 
women in the intervention districts over a 2-year evaluation period is an impressive 
improvement. 
• The INHP-II was not effective in reducing the prevalence of anemia among pregnant 
women in either state. Overall prevalence was comparable to the NFHS-2 prevalence 
although the severity of anemia was greater in this study population. 
Infant and child feeding
• The INHP-II program was highly effective at improving newborn feeding practices, 
including increasing early initiation of breastfeeding and reducing the provision of 
prelacteal feeds. The intervention effects were particularly dramatic in UP state. More 
modest changes in these feeding practices occurred in the UP cohort. 
• Exclusive breastfeeding among infants 0-5 months increased at endline in UP, but 
declined in AP. However, inappropriate early introduction of complementary foods and 
liquids before 6 months of age decreased significantly in both states. The decline in 
exclusive breastfeeding in AP was explained by an increase in breastfeeding with water, 
but the reasons for this are unknown. In the UP cohort, exclusive breastfeeding at 3-5 
months of age was more prevalent in the intervention than comparison district, but early 
introduction of complementary foods was not affected by the intervention. 
• The proportion of mothers of children 6-23 months of age who reported timely 
introduction of complementary foods at 6-8 months increased in both states, but only the 
increase in AP was attributable to the INHP-II intervention. Increases in cereal intake 
from the 24-hour recall among children 6-8 months were observed also in both states, 
but the increase in the intervention relative to the comparison district was significant in 
UP state only. By contrast, the intervention did not influence timely introduction of 
complementary foods among UP cohort mothers when defined by maternal recall or 
reported current dietary intake. 
• In UP, the INHP-II program was highly effective at increasing consumption of cereals 
and legumes among children 6-23 months. In AP, cereal intake increased equally in both 
districts and legume intake remained low. Fruit and vegetable intake was not influenced 
by INHP-II complementary feeding messages in either state, but intake increased in the 
2/20/2008 154 
UP districts and decreased in the AP districts. The intervention was highly effective also 
at increasing legume consumption from 9 to 18 months among children in the UP cohort. 
However, cereal intake increased equally in both districts and fruit and vegetable intake 
increased more rapidly among cohort children in the comparison than the intervention 
district. 
• INHP-II was effective at increasing complementary feeding indicators of diet quantity, 
quality and feeding frequency among children 6-23 months of age in UP, but not AP 
state. Diet quantity, quality, and frequency did increase in AP and remained more 
prevalent than in UP, but improvements were comparable in the two AP districts. 
However, the quality, frequency, and particularly quantity of dietary intake among 
children 6-23 months remained quite inadequate at the end of the evaluation. The 
intervention was less effective in the UP cohort, increasing the feeding frequency, but 
not diet quantity or quality among children from 9 to 18 months of age. 
• Receipt and consumption of SN among children 6-23 months was more prevalent in the 
intervention than comparison district in both states. However, recent declines in 
coverage and utilization of SN in late 2005 in the AP intervention district are of concern 
for the ongoing success of the program. The intervention was effective at distributing SN 
to families of weaning age children in the UP cohort as well. 
• Mothers in the intervention district of both states were more likely than those in the 
comparison district to report that they usually fed their child. In UP only, the 
intervention was also effective at encouraging mothers to feed their young children from 
a separate, rather than a shared, plate. In the UP cohort, some variation in the 
intervention effects was observed. While intervention district mothers were more likely 
to usually feed their child than comparison district mothers, they were less likely to feed 
them from a separate plate. The explanation for the inconsistency in the intervention 
effect in UP is unknown. 
Micronutrient supplementation
• Vitamin A supplementation coverage increased significantly in both states. Receipt of at 
least one dose of vitamin A syrup among children 12-23 months increased by ~50% to 
79% coverage in AP and four-fold to 69% coverage in UP. 
• Pediatric IFA supplementation coverage increased modestly to 30% in AP and 
dramatically to 69% in UP. Gains in coverage occurred during the second year of the 
intervention. However, few children reported consumption of more than a handful of 
tablets. 
Child nutritional status and anemia
• The INHP-II intervention did not have an effect on child nutritional status in the study 
populations in AP and UP states. 
2/20/2008 155 
• In AP state, underweight and stunting among all children did not change significantly. 
However, the mean HAZ score declined and severe wasting increased significantly in the 
intervention district. Stunting was increased significantly also among boys, but not girls. 
• In UP state, underweight, stunting, and wasting did not change significantly. However, 
the underweight prevalence increased and mean HAZ score declined significantly among 
girls, but not boys in the intervention district. 
• In the UP cohort, all 3 indicators of nutritional status followed the same worsening trend 
over the first 18 months of life that were observed in the cross-sectional data. However, 
there was limited evidence of a positive effect of the intervention on nutritional status and 
growth in the cohort sample. Mean WHZ was increased and wasting prevalence reduced 
among intervention children, particularly among girls from 3 to 12 months of age. The 
same benefits of the intervention were observed for mean WAZ and underweight 
prevalence at 6 and 9 months only. No effects on HAZ and stunting were observed. 
• Longitudinal modeling of weight and height gains among cohort children identified small 
positive effects of the intervention on physical growth from 0 to 18 months of age. The 
borderline significant weight gain in intervention district children was larger and 
statistically significant for girls only. The height gain among intervention district children 
was statistically significant also, but was larger for girls also. When calculated over a 12-
month period, children in the intervention district were predicted to have an additional 
200 g weight gain and 0.6 cm gain in height relative to those in the comparison district. 
The comparable gains for girls only were an additional 300 g weight gain and 0.8 cm 
gain in height. 
• The INHP-II program was not effective at reducing anemia among children 12-23 months 
of age in either state, and overall prevalence remained >85% at endline. 
Recommendations for Further Analyses 
The INHP-II program was quite successful at improving maternal dietary practices and nutrition, 
infant and child feeding practices, and micronutrient supplementation in the intervention districts 
of both states, but particularly in UP. The pre-test, post-test, quasi-experimental design of this 
study allowed for evaluation of program impact on various indicators of these outcomes that are 
considered key to improved child nutritional status and health. However, study data from both 
the pre-test, post-test evaluation and the cohort also provide a valuable and unique resource to 
explore in detail the sociodemographic, environmental, health care utilization, and program 
factors that are associated with or predict key study outcomes. Listed below are some examples 
of additional analyses of the Nutrition Evaluation Research data that would be of high priority to 
better understand which factors and program inputs were most strongly associated with positive 
maternal and child nutrition outcomes. The following list is a starting point only. 
Infant and child feeding
• Sociodemographic, environmental, pregnancy, antenatal care utilization and program 
factors/predictors associated with early initiation of breastfeeding 
2/20/2008 156 
o Survey data: cross-sectional association of the factors with early breastfeeding 
initiation in the endline data in AP and UP. 
o Cohort data: Factors assessed at pregnancy enrollment and immediately 
postpartum that predicted early breastfeeding initiation in the 2 UP districts 
• Sociodemographic, environmental, pregnancy, health care utilization, delivery, infant 
feeding, and program factors/predictors associated with exclusive breastfeeding and early 
introduction of complementary foods and liquids. 
o Survey data: cross-sectional association of factors with exclusive breastfeeding 
and early introduction of complementary foods and liquids among children 0-5 
months of age in AP and UP 
o Cohort data: Factors assessed at pregnancy enrollment, and 1, 3, and 6 months 
postpartum that predicted a) exclusive breastfeeding until 6 months of age or b) 
early introduction of complementary foods and liquids in 2 UP districts. 
• Sociodemographic, environmental, health care utilization, infant, and program 
factors/predictors associated with introduction of solids to breastfeeding infants 6-11 
months of age 
• Sociodemographic, environmental, health care utilization, infant, and program 
factors/predictors associated with adequate quantity, frequency, and diversity of dietary 
intake among children 6-23 months of age 
• The association between diversity in the diets of mothers and their children (cross￾sectionally from survey data and longitudinally from cohort data). 
• Evaluation of variation in intervention impact on infant and child feeding practices by 
child sex and household socioeconomic status (survey data). 
Child nutritional status and anemia
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
factors/predictors of a) malnutrition, and b) good nutritional status of children at 6, 12, 
18, and 24 months of age 
o Survey data: cross-sectional association of factors with a) malnutrition, and b) 
good child nutritional status 
o Cohort data: Factors assessed at pregnancy enrollment, and at 1, 3, 6, 9, 12, 15, 
and 18 months of age that predicted a) malnutrition and b) good nutritional status 
as 6, 12, and 18 months of age. 
• Intervention effect on nutritional status and growth from 0-18 months of age in the 2 UP 
districts (cohort – in process) 
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
predictors of good physical growth between birth and 18 months of age in the 2 UP 
districts (cohort). 
2/20/2008 157 
• Sociodemographic, environmental, health care utilization, maternal, infant, and program 
factors associated with anemia among children 12-23 months of age 
• The association between receipt and consumption of pediatric IFA supplements and child 
anemia. 
Micronutrient supplementation
• Sociodemographic, environmental, health care utilization, maternal, infant and program 
factors/predictors of receipt and consumption of pediatric IFA supplements (survey and 
cohort data) 
Maternal Nutrition
• Sociodemographic, environmental, dietary, and program factors/predictors (advice given, 
receipt of SN, etc.) of women’s BMI and malnutrition 
o Survey data: cross-sectional association of factors with BMI and malnutrition 
o Cohort data: Factors assessed at pregnancy enrollment associated with third 
trimester BMI 
• Association of third trimester BMI and neonatal weight (cohort data) 
• Sociodemographic, environmental, dietary, health utilization, pregnancy, delivery, and 
program (contacts with service providers, attendance at NHD, advice given, etc.) factors 
associated with the receipt and consumption of SN 
2/20/2008 158 
REFERENCES 
Bollen, K. A., J.L Glanville, and G. Stecklov.. Socioeconomic Status and Class in Studies of 
Fertility and Health in Developing Countries. Annual Review of Sociology, 2001, 27, 153-85. 
Diggle, P, K. Liang, and S. Zeger. Analysis of Longitudinal Data. New York: Oxford University 
Press, 1994. 
Filmer, D, and L. Pritchett. “Estimating wealth effects without expenditure data – or tears: an 
application to educational enrollment in states of India.” Demography, 2001, 38, 115-32. 
National Family Health Survey 1998-99: India. International Institute for Population Sciences 
(IIPS) and ORC Macro, 2000. 
Wagstaff, A. and N. Watanabe,, "What Difference Does theChoice of SES Make in Health 
Inequality Measurement?." The World Bank, Nov 2002. 
Wooldridge, J. M. Econometric Analysis of Cross Section and Panel Data. MIT Press , 2002. 
2/20/2008 159 
APPENDIX I: CARE India’s INHP-II Basic Package of Nutrition Interventions 
Life Cycle Period Basic Package 
Pregnancy 3 contacts 
• Iron Folic Acid (IFA large) supplementation 
• Rest for 1-2 hours 
• Dietary advice 
• Consume one additional meal everyday 
• Consume all available foods in the house 
• Consume Supplementary Nutrition (SN) 
Birth- 28 days 3+ contacts at 0, 3 and 7 days 
• Initiate breastfeeding within 1 hour 
• Exclusive breastfeeding 
• Make sure baby is getting as much milk as possible, e.g. correct position and attachment 
• Feed from one breast until soft and then switch to the other in order for child to receive fore and hind 
milk
•
• For newborn who does not feed vigorously; provide expressed breast milk with clean cup and spoon as long 
as breastfeeding is not adequately vigorous; frequent feeding 
• Dietary advice for mother 
• Consume one additional meal everyday 
• Consume all available foods in the house 
• Consume SN 
1-5 months 3 contacts at 1.5, 2.5 and 3.5 months 
• Exclusive breastfeeding for 6 months 
• Dietary advice for mother 
• Consume one additional meal everyday 
• Consume all available foods in the house 
• Consume SN 
6-8 months 1+ contact 
• Continued breastfeeding 
• Appropriate Complementary Feeding 
• Initiate complementary foods: rice, dal, suji, banana etc. 
• Start with few teaspoons and gradually increase the quantity 
• Child should be fed 1/2 katori of soft mashed food 3 times a day, with 1/2 to 1 tsp of oil added 
• Feed the child yourself from a separate plate or bowl to assess how much the child has eaten; interact 
with the child 
• During illness breastfeed more frequently, provide fluids 
• After illness give more food at each sitting and feed more frequently until the child has regained weight 
9-11 months 1+ contact 
• Continued breastfeeding 
• Appropriate Complementary Feeding 
• Child should be fed 1/ 2 katori of soft mashed food (rice, dal, vegetables, green leafy vegetables, fruits. 
Milk products, egg , meat , fish if available and feasible) 4 times with 1 tsp of oil added 
• Feed the child yourself from a separate plate or bowl to assess how much the child has eaten; interact 
with the child 
• During illness breastfeed more frequently, provide fluids 
• After illness give more food at each sitting and feed more frequently until the child has regained weight 
• Vitamin A and measles immunization at 9 months
12-24 months 2+ contacts 
• Continued breastfeeding 
• Appropriate Complementary Feeding 
• Child should be fed 3/4 katori of soft mashed food (rice, dal, green leafy vegetables, other vegetables, 
fruits. Milk products, egg , meat , fish if available and feasible) 4-5 times a day with 1-2 tsp oil 
• Feed the child yourself from a separate plate or bowl to assess how much the child has eaten; interact 
with the child 
• During illness breastfeed more frequently, provide fluids 
• After illness give more food at each sitting and feed more frequently until the child has regained weight 
• Vitamin A at 18 months 
• Give iron supplements beginning 12 months {GOI policy, available for about 50% of children} 
2/20/2008 160 
APPENDIX II: ADDITIONAL TABLES FROM THE PRE-TEST, POST-TEST 
EVALUATION
A. Antenatal Care Services 
Andhra Pradesh
Table A.1. Mothers of children 0-5 months of age reporting 1+ service provider home visits during the last 
pregnancy in Andhra Pradesh 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 
pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 34.9 56.0 21.1 24.3 29.9 5.6 15.5*
ANM 30.3 41.6 11.3 22.1 23.2 1.1 10.2*
AWW 13.6 47.6 34 8.2 16.1 7.9 26.1*
CA 0 3.0 3 0.2 0.2 0 3*
Total # of 
mothers 
624 563 646 622 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table A.2. Mothers of children 0-5 months of age reporting 1-2, 3+ service provider home visits during the 
last pregnancy in Andhra Pradesh 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 
pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 
1-2 13.9 13.5 -0.4 9.9 11.1 1.2 -1.6
3+ 21.0 42.5 21.5 14.4 18.8 4.4 17.1*
ANM 
1-2 13.6 17.2 3.6 11.0 12.7 1.7 1.9
3+ 16.7 24.3 7.6 11.2 10.5 -0.7 8.3*
AWW 
1-2 5.9 18.3 12.4 4.6 8.4 3.8 8.6*
3+ 7.7 29.3 21.6 3.6 7.7 4.1 17.5*
CA 
1-2 0 1.1 1.1 0 0.2 0.2 0.9*
3+ 0.2 2.0 1.8 0 0 0 1.8*
Total # of 
mothers 
624 563 646 622 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table A.3. Mothers of children 0-5 months of age reporting 1+ times receiving advice from places other than 
home during the last pregnancy in Andhra Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ times receiving advice from places other than home 
during during the last pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 56.7 55.6 
2/20/2008 161 
AWC 50.4* 40.8* 
NHD 15.1* 8.5* 
Mahila 0.4 0.6 
Others 7.1* 20.1* 
Total # of 
mothers 
 563 622 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
Table A.4. Mothers of children 0-5 months of age reporting 1-2, 3+ service provider home visit during the last 
3 months of the most recent pregnancy in Andhra Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 3 
months of pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 
1-2 19.9 12.9 
3+ 22.2 9.8 
ANM 
1-2 23.5 13.0 
3+ 6.9 4.0 
AWW 
1-2 25.8 9.2 
3+ 11.6 2.9 
CA 
1-2 1.4 0.2 
3+ 1.1 0 
Total # of 
mothers 
 563 622 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table A.5. Mothers of children 0-5 months of age reporting 1+ service provider home visit during the last 3 
months of the most recent pregnancy in Andhra Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 3 
months of the most recent pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 42.1* 22.7* 
ANM 30.4* 17.0* 
AWW 37.3* 12.1* 
CA 2.5* 0.2* 
Total # of 
mothers 
 563 622 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
2/20/2008 162 
Table A.6. Mothers of children 0-5 months of age reporting 1+ times receiving advice from places other than 
home during the last 3 months of the most recent pregnancy in Andhra Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ times receiving advice from places other than home 
during the last 3 months of the most recent pregnancy (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 40.9 43.6 
AWC 36.4* 30.4* 
NHD 9.6* 5.5* 
Mahila n/a n/a 
Others 5.5* 18.5* 
Total # of 
mothers 
 563 622 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
2/20/2008 163 
Table A.7. Pregnancy advice on diet, rest, and IFA supplements given during home visits to mothers of 
children 0-5 months of age who received any advice during last pregnancy in Andhra Pradesh (endline only)
Advice received (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Pregnancy 
advice by 
service provider 
type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Auxilary Nurse 
Midwife 
(ANM) 
 
Diet 54.6* 37.5* 
Rest 49.1* 33.7* 
IFA 
supplements 44.2 51.9
Breastfeeding 
advice 10.9* 1.9*
Total # of 
mothers 0-5m 
who received 
advice from 
ANM 
 165 104 
Anganwadi 
worker (AWW) 
 
Diet 54.2 58.9 
Rest 59.7* 41.1* 
IFA 
supplements 40.8* 17.8*
Breastfeeding 
advice 
 9.0* 1.4* 
Total # of 
mothers 0-5m 
who received 
advice from 
AWW 
 201 73 
Change Agent 
 
Diet 62.5 n/a 
Rest 50.0 n/a 
IFA 
supplements 37.5 n/a 
Breastfeeding 
advice 
 12.5 n/a 
Total # of 
mothers 0-5m 
who received 
advice from CA 
 16 n/a 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
2/20/2008 164 
Uttar Pradesh
Table A.8. Mothers of children 0-5 months of age reporting 1+ service provider home visit during the last 
pregnancy in Uttar Pradesh 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 
pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 23.1 70.4 47.3 18.4 25.1 6.7 40.6*
ANM 17.4 47.8 30.4 11.7 18.3 6.6 23.8*
AWW 7.6 57.4 49.8 4.8 9.4 4.6 45.2*
CA 1.2 28.2 27 0 0.3 0.3 26.7*
Total # of 
mothers 
804 763 727 682 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table A.9. Mothers of children 0-5 months of age reporting 1-2, 3+ service provider home visits during the 
last pregnancy in Uttar Pradesh 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 
pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 
1-2 15.4 15.2 -0.2 13.8 17.6 3.8 -4.0
3+ 7.7 55.2 47.5 4.7 7.5 2.8 44.7*
ANM 
1-2 13.6 26.9 13.3 9.8 16.9 7.1 6.2*
3+ 3.9 21.0 17.1 1.9 1.5 -0.4 17.5*
AWW 
1-2 4.5 21.0 16.5 2.6 5.7 3.1 13.4*
3+ 3.1 36.4 33.3 2.2 3.7 1.5 31.8*
CA 
1-2 0.5 11.9 11.4 0 0.3 0.3 11.1*
3+ 0.8 16.3 15.5 0 0 0 15.5*
Total # of 
mothers 
804 763 727 682 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
2/20/2008 165 
Table A.10. Mothers of children 0-5 months of age reporting 1+ times receiving advice from places other than 
home during the last pregnancy in Uttar Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ times receiving advice from places other than home 
during during the last pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 55.3* 20.8* 
AWC 38.0* 7.9* 
NHD 32.4* 2.2* 
Mahila 3.3* 0* 
Others 12.6 12.3 
Total # of 
mothers 
 763 682 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
Table A.11. Mothers of children 0-5 months of age reporting 1-2, 3+ service provider home visits during the 
last 3 months of the most recent pregnancy in Uttar Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 3 
months of pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall* 
1-2 19.1 16.7 
3+ 40.0 3.5 
ANM* 
1-2 29.8 14.8 
3+ 8.0 0.9 
AWW* 
1-2 27.3 5.1 
3+ 19.5 1.0 
CA* 
1-2 12.2 0 
3+ 9.8 0 
Total # of 
mothers 
 763 682 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table A.12. Mothers of children 0-5 months of age reporting 1+ service provider home visit during the last 3 
months of the most recent pregnancy in Uttar Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ service provider home visit during the last 3 
months of the most recent pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 59.1* 20.2* 
ANM 37.8* 15.7* 
AWW 46.8* 6.2* 
CA 22.0* 0* 
Total # of 
mothers 
 763 682 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
2/20/2008 166 
Table A.13. Mothers of children 0-5 months of age reporting 1+ times receiving advice from places other than 
home during the last 3 months of the most recent pregnancy in Uttar Pradesh (endline only) 
Mothers of 0-5 month olds reporting 1+ times receiving advice from places other than home 
during the last 3 months of the most recent pregnancy (%) 
Intervention (Barabanki) Comparison (Unnao) 
Service 
provider type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Overall 37.2* 12.0* 
AWC 26.3* 5.6* 
NHD 23.3* 1.2* 
Mahila 1.2 0.6 
Others 4.7 6.0 
Total # of 
mothers 
 763 682 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
2/20/2008 167 
Table A.14. Pregnancy advice on diet, rest, and IFA supplements given during home visits to Mothers of 
children 0-5 months of age who received any advice during last pregnancy in Uttar Pradesh (endline only)
Advice received (%) 
Intervention (Barabanki) Comparison (Unnao) 
Pregnancy 
advice by 
service provider 
type 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Auxilary Nurse 
Midwife 
(ANM) 
 
Diet 56.4* 19.6* 
Rest 42.9* 8.8* 
IFA 
supplements 54.2 52.0
Breastfeeding 
advice 14.5* 1.0*
Total # of 
mothers 0-5m 
who received 
advice from 
ANM 
 282 102 
Anganwadi 
worker (AWW) 
 
Diet 62.2* 35.3* 
Rest 51.6* 23.5* 
IFA 
supplements 47.4* 31.4*
Breastfeeding 
advice 19.8* 2.0*
Total # of 
mothers 0-5m 
who received 
advice from 
AWW 
 384 51 
Change Agent 
 
Diet 66.1 n/a 
Rest 54.8 n/a 
IFA 
supplements 44.6 n/a 
Breastfeeding 
advice 28.5 
Total # of 
mothers 0-5m 
who received 
advice from CA 
 186 n/a 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05 
2/20/2008 168 
B. Rest during Pregnancy and the Postpartum Period
Table B.1. Afternoon rest (≥2 hours) during the past week among pregnant women and mothers of children 
0-5 months in Andhra Pradesh
Afternoon rest for ≥2 hours during the past week (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
79.9 
(582) 
61.1 
(572) 
-18.8 76.1 
(632) 
57.0 
(703) 
-19.1 0.3 
Mothers of 
children 0-5 
months 
66.5 
(624) 
48.1 
(563) 
-18.4 79.2 
(638) 
58.0 
(622) 
-21.2 2.8 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
Table B.2. Afternoon rest (≥2 hours) during the past week among pregnant women and mothers of children 
0-5 months in Uttar Pradesh
Afternoon rest for ≥2 hours during the past week (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Pregnant 
women 
26.9 
(745) 
38.8 
(724) 
11.9 27.2 
(699) 
24.4 
(676) 
-2.8 14.7* 
Mothers of 
children 0-5 
months 
22.7 
(802) 
24.6 
(763) 
1.9 13.8 
(724) 
14.8 
(682) 
1.0 0.9 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05 
C. Delivery, Newborn, and Postpartum Care 
Andhra Pradesh
Table C. 1. Place of delivery among women of infants 0-5 months of age in Andhra Pradesh 
Place of delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Sub-Center, 
PHC, Govt 
Hospital 11.9 15.0 3.1 36.9 38.6 1.7 1.4
Private Clinic, 
Hospital 59.8 64.6 4.8 17.3 28.1 10.8 -6
At Home 27.7 20.2 -7.5 42.2 33.1 -9.1 1.6
Other 0.2 0.2 0 2.2 0 -2.2 2.2*
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 169 
Table C.2. Birth attendant among women of infants 0-5 months of age in Andhra Pradesh 
Birth Attendant (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
No assistance 
during delivery 0.8 0.5 -0.3 0.6 0.2 -0.4 0.1
Untrained birth 
attendant (Dai), 
relative, or Dai 
training status 
unknown 11.9 6.5 -5.4 32.7 16.9 -15.8 10.4*
Trained 
Dai/TBA 4.4 9.4 5 6.7 11.4 4.7 0.3
Doctor/nurse 
/ANM 80.5 81.2 0.7 57.6 69.3 11.7 -11*
Other 0.3 2.4 2.1 2.3 2.3 0 2.1
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.3. Postpartum contacts with ANM among mothers of infants 0-5 months of age in Andhra Pradesh 
1+ postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
37.1 47.7 10.6 33.3 40.5 7.2 3.4
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.4. Postpartum contacts with AWW among mothers of infants 0-5 months of age in Andhra Pradesh 
1+ postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
30.5 56.1 25.6 29.8 28.9 -0.9 26.5*
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.5. Postpartum contacts with CA among mothers of infants 0-5 months of age in Andhra Pradesh 
1+ postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0.6 4.3 3.7 n/a n/a 
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 170 
Table C.6. Postpartum contacts on the day of delivery with ANM among mothers of infants 0-5 months of age 
who delivered at home in Andhra Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
3.5 4.5 1 1.9 1.0 -0.9 1.9
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 112 271 204 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.7. Postpartum contacts on the day of delivery with AWW among mothers of infants 0-5 months of 
age who delivered at home in Andhra Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1.2 2.7 1.5 0 2.0 2 -0.5
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 102 271 174 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.8. Postpartum contacts on the day of delivery with CA among mothers of infants 0-5 months of age 
who delivered at home in Andhra Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0 0 n/a n/a 
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 102 271 174 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.9. Postpartum contacts in the first week after delivery with ANM among mothers of infants 0-5 
months of age who delivered at home in Andhra Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
11.1 14.3 3.2 6.6 4.4 -2.2 5.4
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 112 271 204 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 171 
Table C.10. Postpartum contacts in the first week after delivery with AWW among mothers of infants 0-5 
months of age who delivered at home in Andhra Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
10.5 18.8 8.3 9.2 4.4 -4.8 13.1*
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 102 271 174 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.11. Postpartum contacts in the first week after delivery with CA among mothers of infants 0-5 
months of age who delivered at home in Andhra Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0 1.8 1.8 n/a n/a 
Total # of 
mothers 0-5 mo 
who delivered 
at home 
172 102 271 174 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.12. Postpartum contacts (0, 1-2, 3+) with ANM among mothers of infants 0-5 months of age in 
Andhra Pradesh 
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 22.1 27.2 5.1 22.7 25.9 3.2 1.9
3+ 15.0 20.5 5.5 10.5 14.6 4.1 1.4
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.13. Postpartum contacts (0, 1-2, 3+) with AWW among mothers of infants 0-5 months of age in 
Andhra Pradesh 
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 20.2 29.0 8.8 23.0 18.3 -4.7 13.5*
3+ 10.1 26.1 16 6.5 9.6 3.1 12.9*
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 172 
Table C.14. Postpartum contacts (0, 1-2, 3+) with CA among mothers of infants 0-5 months of age in Andhra 
Pradesh 
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 0.5 2.4 1.9 n/a n/a 
3+ 0.2 2.2 2 n/a n/a 
Total # of 
mothers 0-5m 
625 551 648 617 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.15. Number of home visits by ANM in the first week (including the day of delivery) among mothers 
of infants 0-5 months of age in Andhra Pradesh (endline survey only)
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 11.0* - 5.0* 
3+ - 1.4 - 1.0 
Total # of 
mothers 0-5m 
- 551 - 617 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
Table C.16. Number of home visits by AWW in the first week (including the day of delivery) among mothers 
of infants 0-5 months of age in Andhra Pradesh (endline survey only) 
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 11.7* - 4.0* 
3+ - 3.0 - 1.5 
Total # of 
mothers 0-5m 
- 497 - 504 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
Table C. 17. Number of home visits by CA in the first week (including the day of delivery) among mothers of 
infants 0-5 months of age in Andhra Pradesh (endline survey only) 
postpartum contacts (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 0.5 - 0 
3+ - 0.4 - 0 
Total # of 
mothers 0-5m 
- 497 - 504 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
2/20/2008 173 
Table C.18. Advice received at postpartum contacts with ANM among mothers of infants 0-5 months of age 
in Andhra Pradesh 
Advice received among all mothers who received any advice (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 19.4 60.7 41.3 6.5 22.1 15.6 25.4*
Any 
complementary 
feeding advice 7.3 17.1 9.8 4.4 4.4 0 9.8*
Advice to have 
child weighed 14.1 34.6 20.5 28.3 15.3 -13 33.5*
Any child 
health advice 76.4 81.3 4.9 79.0 90.8 11.8 -6.9
Any maternal 
nutrition advice 31.4 24.1 -7.3 28.3 8.4 -19.9 12.6*
Total # of 
mothers 0-5 mo 
who received 
any advice 
191 257 138 249 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.19. Advice received at postpartum contacts with AWW among mothers of infants 0-5 months of age 
in Andhra Pradesh 
Advice received among all mothers who received any advice (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Type of advice 
received 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 19.3 64.2 44.9 4.2 30.7 26.5 18.4*
Any 
complementary 
feeding advice 6.0 28.7 22.7 2.8 13.5 10.7 12*
Advice to have 
child weighed 30.1 45.4 15.3 50.7 38.7 -12 27.3*
Any child 
health advice 64.5 62.1 -2.4 66.2 54 -12.2 9.8
Any maternal 
nutrition advice 36.1 40.3 4.2 31.7 26.4 -5.3 9.5
Total # of 
mothers 0-5 mo 
who received 
any advice 
166 293 142 163 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 174 
Table C.20. Advice received at postpartum contacts with CA among mothers of infants 0-5 months of age in 
Andhra Pradesh 
Advice received among all who mothers who received any advice (%) 
Intervention (Karimnagar) Comparison (Rangareddy) 
Type of advice 
received 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 50.0 80.0 30 n/a n/a
Any 
complementary 
feeding advice 0 45.0 45 n/a n/a
Advice to have 
child weighed 0 30.0 30 n/a n/a
Any child 
health advice 50.0 80.0 30 n/a n/a
Any maternal 
nutrition advice 50.0 30.0 -20 n/a n/a
Total # of 
mothers 0-5 mo 
who received 
any advice 
2 21 n/a n/a 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Uttar Pradesh
Table C.21. Place of delivery among mothers of infants 0-5 months of age in Uttar Pradesh 
Place of delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Sub-Center, 
PHC, Govt 
Hospital 7.5 8.9 1.4 5.0 9.9 4.9 -3.5
Private Clinic, 
Hospital 7.8 16.4 8.6 6.2 8.7 2.5 6.1*
At Home 82.7 73.9 -8.8 85.8 80 -5.8 -3
Other 0.8 0.8 0 2.6 1.5 -0.9 0.9
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 175 
Table C.22. Birth attendant among mothers of infants 0-5 months of age in Uttar Pradesh 
Birth Attendant (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
No assistance 
during delivery 3.7 1.2 -2.5 5.3 5.3 0 -2.5
Untrained birth 
attendant (Dai), 
relative, or Dai 
training status 
unknown 53.4 50.7 -2.7 68.7 62.5 -6.2 3.3
Trained 
Dai/TBA 3.2 8.2 5 4.0 2.4 -1.6 6.6*
Doctor/nurse 
/ANM 22.8 30.2 7.4 16.3 21.3 5 2.4
Other 16.9 9.6 -6.7 5.7 8.6 2.9 -9.6*
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.23. Postpartum contacts with ANM among mothers of infants 0-5 months of age in Uttar Pradesh 
1+ postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
29.2 66.8 37.6 25.6 25.8 0.2 37.4*
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C. 24. Postpartum contacts with AWW among mothers of infants 0-5 months of age in Uttar Pradesh 
1+ postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
17.9 71.2 53.3 10 14.8 4.8 48.5*
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 176 
Table C.25. Postpartum contacts with CA among mothers of infants 0-5 months of age in Uttar Pradesh 
1+ postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
5.1 33.7 28.6 n/a n/a 
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.26. Postpartum contacts on the day of delivery with ANM among mothers of infants 0-5 months of 
age who delivered at home in Uttar Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1.8 5.0 3.2 0.8 1.5 0.7 2.5*
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.27. Postpartum contacts on the day of delivery with AWW among mothers of infants 0-5 months of 
age who delivered at home in Uttar Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0.9 4.8 3.9 1.0 1.1 0.1 3.8*
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.28. Postpartum contacts on the day of delivery with CA among mothers of infants 0-5 months of age 
who delivered at home in Uttar Pradesh 
postpartum contacts on the day of delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0.8 6.3 5.5 n/a n/a 
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 177 
Table C.29. Postpartum contacts in the first week after delivery with ANM among mothers of infants 0-5 
months of age who delivered at home in Uttar Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
4.2 16.8 8.6 3.9 1.7 -2.2 10.8*
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.30. Postpartum contacts in the first week after delivery with AWW among mothers of infants 0-5 
months of age who delivered at home in Uttar Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1.8 28.0 26.2 1.1 1.9 0.8 25.4*
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.31. Postpartum contacts in the first week after delivery with CA among mothers of infants 0-5 
months of age who delivered at home in Uttar Pradesh 
postpartum contacts on the first week after delivery (%) 
Intervention (Barabanki) Comparison (Unnao) 
1+ contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
0.6 12.4 11.8 n/a n/a 
Total # of 
mothers 0-5 mo 
who delivered 
at home 
665 556 621 529 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C. 32. Postpartum contacts (0, 1-2, 3+) with ANM among mothers of infants 0-5 months of age in Uttar 
Pradesh 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 21.3 37.7 16.4 20.4 21.5 1.1 15.3*
3+ 8.0 29.1 21.1 5.2 4.3 -0.9 22*
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 178 
Table C.33. Postpartum contacts (0, 1-2, 3+) with AWW among mothers of infants 0-5 months of age in Uttar 
Pradesh 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 12.4 31.9 19.5 6.7 11.4 4.7 14.8*
3+ 5.4 39.3 33.9 3.3 3.4 0.1 33.8*
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.34. Postpartum contacts (0, 1-2, 3+) with CA among mothers of infants 0-5 months of age in Uttar 
Pradesh 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 4.1 17.5 13.4 n/a n/a 
3+ 1.0 16.2 15.2 n/a n/a 
Total # of 
mothers 0-5m 
804 759 731 679 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.35. Number of home visits by ANM in the first week (including the day of delivery) among mothers 
of infants 0-5 months of age in Uttar Pradesh (endline survey only) 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 17.6* - 3.2* 
3+ - 2.9* - 0.2* 
Total # of 
mothers 0-5m 
- 759 - 679 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
Table C.36. Number of home visits by AWW in the first week (including the day of delivery) among mothers 
of infants 0-5 months of age in Uttar Pradesh (endline survey only) 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 24.9* - 2.4* 
3+ - 5.5* - 0* 
Total # of 
mothers 0-5m 
- 759 - 679 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
2/20/2008 179 
Table C.37. Number of home visits by CA in the first week (including the day of delivery) among mothers of 
infants 0-5 months of age in Uttar Pradesh (endline survey only) 
postpartum contacts (%) 
Intervention (Barabanki) Comparison (Unnao) 
Number of 
contacts 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
1-2 - 11.1 - 0 
3+ - 5.8 - 0 
Total # of 
mothers 0-5m 
- 759 - 679 
* Statistical test result comparing intervention and control districts at endline, significant at p<0.05
Table C.38. Advice received at postpartum contacts with ANM among mothers of infants 0-5 months of age 
in Uttar Pradesh 
Advice received among all mothers who received any advice (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 13.1 75.4 62.3 5.7 20.3 14.6 47.7*
Any 
complementary 
feeding advice 2.5 18.3 15.8 1.4 7 5.6 10.2*
Advice to have 
child weighed 1.9 13.0 11.1 0 2.1 2.1 9.0*
Any child 
health advice 49.4 64.6 15.2 41.4 86.7 45.3 -29.9*
Any maternal 
nutrition advice 1.9 8.8 6.9 1.4 2.8 1.4 5.5
Total # of 
mothers 0-5 mo 
who received 
any advice 
160 475 140 143 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
2/20/2008 180 
Table C.39. Advice received at postpartum contacts with AWW among mothers of infants 0-5 months of age 
in Uttar Pradesh 
Advice received among all mothers who received any advice (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 23.8 81.9 58.1 13.6 27 13.4 44.7*
Any 
complementary 
feeding advice 4.8 20.4 15.6 7.6 5.7 -1.9 17.5*
Advice to have 
child weighed 1.9 22.6 20.7 1.5 16.2 14.7 6
Any child 
health advice 35.2 52.3 17.1 31.8 43.2 11.4 5.7
Any maternal 
nutrition advice 9.5 16.7 7.2 4.6 18.9 14.3 -7.1
Total # of 
mothers 0-5 mo 
who received 
any advice 
105 514 66 74 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05
Table C.40. Advice received at postpartum contacts with CA among mothers of infants 0-5 months of age in 
Uttar Pradesh 
Advice received among all mothers who received any advice (%) 
Intervention (Barabanki) Comparison (Unnao) 
Baseline Endline Δ Baseline Endline Δ
Difference 
of Δ
Any 
breastfeeding 
advice 53.3 85.2 31.9 n/a n/a
Any 
complementary 
feeding advice 0 22.4 22.4 n/a n/a
Advice to have 
child weighed 6.7 15.2 8.5 n/a n/a
Any child 
health advice 53.3 48.5 -4.8 n/a n/a
Any maternal 
nutrition advice 0 17.3 17.3 n/a n/a
Total # of 
mothers 0-5 mo 
who received 
any advice 
15 238 
* Statistical test result comparing change from baseline to endline in intervention and control districts, significant at 
p<0.05