EVALUATION OF THE AIN-C 
PROGRAM IN HONDURAS
EVALUATION OF THE AIN-C 
PROGRAM IN HONDURAS 
October 2008 
This publication was produced for the United States Agency for International Development. It was 
prepared by Tom Schaetzel, Marcia Griffiths, Joy Miller Del Rosso, and Beth Plowman on behalf of 
BASICS. 
The authors’ views expressed in this publication do not necessarily reflect the views of the United 
States Agency for International Development or the United States Government. 
ii
Recommended Citation 
Schaetzel, Tom et al. 2008. Evaluation of the AIN-C Program in Honduras. Arlington, VA, USA: 
Basic Support for Institutionalizing Child Survival (BASICS) for the United States Agency for 
International Development (USAID). 
U.S. Agency for International Development 
Bureau for Global Health 
Office of Health, Infectious 
Diseases and Nutrition 
Ronald Reagan Building 
1300 Pennsylvania Ave., NW 
Washington, D.C. 20523 
Tel: (202) 712-0000 
Email: globalhealth@phnip.com 
www.usaid.gov/our_work/global_health 
BASICS 
4245 N. Fairfax Dr., Suite 850 
Arlington, VA 22203 
Tel: (703) 312-6800 
Fax: (703) 312-6900 
Email: basics@basics.org 
www.basics.org
Support for this publication was provided by the USAID Bureau for Global Health.
BASICS (Basic Support for Institutionalizing Child Survival) is a global project to assist 
developing countries in reducing infant and child mortality through the implementation of 
proven health interventions. BASICS is funded by the U.S. Agency for International 
Development (contract no. GHA-I-00-04-00002-00) and implemented by the Partnership for 
Child Health Care, Inc., comprised of the Academy for Educational Development, John 
Snow, Inc., and Management Sciences for Health. Subcontractors include The Manoff 
Group, Inc., the Program for Appropriate Technology in Health, and Save the Children 
Federation, Inc.
iii
Abstract 
The Honduras AIN-C (Atencion Integral a la Ninez en la Comunidad [Integrated Community 
Child Health Program]) is a volunteer-led, community-based growth monitoring and 
promotion program aimed at preventing under-nutrition in children under two years of age. 
AIN-C was initially launched in the mid 1990s and intensely developed and supported by 
USAID over the period 1998-2005. An impact evaluation covering this period was conducted 
in 2002. The original quasi-experimental, “intention-to-treat” evaluation design called for pre￾and post-comparisons between intervention and control communities. A multi-stage cluster 
sampling design was employed; data were collected on 1467 households at baseline and 
1343 households in the final evaluation. Due to issues including non-equivalent groups, 
extensive contamination of control communities, and reduced intensity of implementation due 
to unanticipated diversion of funding during the evaluation period, the pre-post, intervention￾control community comparison was replaced with an analysis based on individual, 
community-level participation in AIN-C. Data were analyzed from approximately 1300 
households (600 AIN-C; 700 “No-GMP”) and 1200 children (580 AIN-C; 611 “No-GMP”). 
The results demonstrate that AIN-C had a significant, positive impact on young child feeding 
and care practices and on nutritional status, especially among poorer households and 
children who participated more regularly in the program. Exclusive breastfeeding rates 
among infants younger than six months of age, the proportion of caregivers who fed children 
6–23 months appropriately, the percentage of children who were fully immunized by 13 
months, and the percentage of children through 23 months who received iron and vitamin A 
supplementation were all significantly higher in the AIN-C group than the group that did not 
participate in GMP. Controlling for socio-economic status and age, AIN-C participants had 
mean weight-for-height Z-scores .122 higher than non-participants. The impact of AIN-C was 
two to three times greater among poorer households. Intensity of participation also increased 
impact; each additional month of participation yielded a .042 increase in weight-for-age Z￾score.
iv
Table of Contents
ABSTRACT...............................................................................................................................................................III
TABLE OF CONTENTS .................................................................................................................................IV
LIST OF TABLES ......................................................................................................................................... V
LIST OF FIGURES .......................................................................................................................................VI
ACKNOWLEDGEMENTS ..............................................................................................................................VII
ACRONYMS..............................................................................................................................................VIII
EXECUTIVE SUMMARY................................................................................................................................ 1
INTRODUCTION .......................................................................................................................................... 7
METHODOLOGY ....................................................................................................................................... 12
DESCRIPTION OF THE SAMPLE .................................................................................................................. 20
CHARACTERISTICS OF HOUSEHOLDS ........................................................................................................ 22
CAREGIVER CHARACTERISTICS................................................................................................................. 27
PARTICIPATION IN AIN-C PROGRAM ACTIVITIES ........................................................................................ 30
CHILD GROWTH AND DEVELOPMENT KNOWLEDGE AND ACTIONS FOR IMPROVEMENT .................................. 37
INFANT AND YOUNG CHILD FEEDING PRACTICES ....................................................................................... 44
CARE SEEKING AND TREATMENT FOR COMMON CHILDHOOD ILLNESSES..................................................... 55
DELIVERY AND POSTPARTUM CARE .......................................................................................................... 62
VACCINATIONS AND MICRONUTRIENTS ...................................................................................................... 66
NUTRITIONAL STATUS AND AIN-C PARTICIPATION ..................................................................................... 69
13. CONCLUSIONS ................................................................................................................................... 78
v
List of Tables 
TABLE 1.1 TIMELINE OF EVENTS IN THE DEVELOPMENT AND IMPLEMENTATION OF AIN-C: 1990-2002 ........... 9 
TABLE 2.1 SUMMARY OF THE LEVELS AND PROCESS OF SELECTION OF THE SAMPLE .................................. 14 
TABLE 2.2 BILATERAL USAID PROGRAM SUPPORT IN HEALTH AREAS INCLUDED IN THE EVALUATION.......... 16 
TABLE 3.1 FINAL EVALUATION SAMPLE SIZE, BY HEALTH AREA.................................................................. 20 
TABLE 3.2 FINAL EVALUATION SAMPLE SIZE, BY SEX OF THE INDEX CHILD ................................................. 20 
TABLE 3.3: FINAL EVALUATION SAMPLE SIZE, BY AGE GROUP OF CHILD..................................................... 21 
TABLE 4.1 WATER AND SANITATION CHARACTERISTICS OF HOUSEHOLDS................................................... 22 
TABLE 4.2 CHARACTERISTICS OF HOUSING............................................................................................... 23 
TABLE 4.3 HOUSEHOLD POSSESSION OF AMENITIES.................................................................................. 24 
TABLE 4.3 CONTACT WITH OTHER SOCIAL ASSISTANCE PROGRAMS ........................................................... 25 
TABLE 6.1 AGE OF ENTRY INTO AIN-C...................................................................................................... 30 
TABLE 6.2 ACCURACY OF PLOTTING ON GROWTH CARDS ALL CHILDREN COMBINED .................................. 32 
TABLE 7.1 CAREGIVERS’ PERCEPTIONS OF THEIR OWN CHILD’S GROWTH, BY PROGRAM PARTICIPATION ..... 37 
TABLE 7.2 RECOGNITION OF AIN-C COUNSELING CARDS .......................................................................... 41 
TABLE 8.1 MEDIAN DURATION OF EXCLUSIVE AND ANY BREASTFEEDING BY PARTICIPATION........................ 45 
TABLE 8.2 ESTIMATED MONTHS BREASTFEEDING (KAPLAN-MEIER SURVIVAL ANALYSIS) ............................ 46 
TABLE 8.3 AMOUNT OF FOOD RECEIVED AMONG CHILDREN 9-11 MONTHS................................................. 48 
TABLE 8.4 AMOUNT OF FOOD RECEIVED AMONG CHILDREN 12-23 MONTHS............................................... 49 
TABLE 8.5 FEEDING PRACTICES DURING DIARRHEAL ILLNESS FOR CHILDREN 6 MONTHS OR OLDER............ 51 
TABLE 8.6 FEEDING PRACTICES DURING ACUTE RESPIRATORY ILLNESS CHILDREN 6 MONTHS OR OLDER.... 52 
TABLE 9.1 PREVALENCE OF DIARRHEA IN CHILDREN BY PROGRAM TYPE.................................................... 55 
TABLE 9.2 USE OF ORAL REHYDRATION THERAPY OR OTHER FLUIDS FOR DIARRHEAL DISEASE .................. 56 
TABLE 9.3 SOURCE OF CARE FOR DIARRHEA ............................................................................................ 56 
TABLE 9.4 LEADING TYPES OF ADVICE RECEIVED BY CARETAKERS OF CHILDREN WITH DIARRHEA .............. 57 
TABLE 9.5 CAREGIVERS’ PERCEPTIONS CONCERNING DANGER SIGNS OF SEVERE DIARRHEA..................... 58 
TABLE 9.7 CAREGIVERS’ PERCEPTIONS OF DANGER SIGNS IN CHILDREN WITH ARI .................................... 60 
TABLE 10.1 PLACE OF DELIVERY BY PROGRAM PARTICIPATION.................................................................. 62 
TABLE 10.2 PERSONNEL MAKING POSTPARTUM HOME VISITS ................................................................... 63 
TABLE 10.3 ACTIVITIES DURING MONITORA’S POSTPARTUM VISIT TO AIN-C PARTICIPANTS ........................ 64 
TABLE 12.1 MEAN AND MEDIAN VALUES FOR ANTHROPOMETRIC MEASURES OF NUTRITIONAL STATUS AMONG 
AIN-C AND “NO-GMP” CHILDREN.................................................................................................... 69 
TABLE 12.2: FACTORS ASSESSED IN THE CREATION OF AN ASSET SCORE AND THEIR CODING .................... 71 
TABLE 12.3 MEAN AND MEDIAN VALUES OF THE ASSET SCORE FACTOR USED FOR ESTIMATING ECONOMIC 
STATUS, BY PARTICIPATION STATUS ................................................................................................ 71 
TABLE 12.4 AGE DISTRIBUTION OF INFANTS/CHILDREN ACCORDING TO PARTICIPATION STATUS.................. 72 
TABLE 12.5 REGRESSION ANALYSIS OF NUTRITIONAL STATUS IN RELATION TO AIN-C PARTICIPATION,
CONTROLLED FOR INFANT/CHILD AGE.............................................................................................. 73 
TABLE 12.6 RESULTS OF MULTIPLE LINEAR REGRESSION ANALYSIS PREDICTING NUTRITIONAL STATUS ON 
THE BASIS OF PARTICIPATION INTENSITY, INFANT/CHILD AGE AND HOUSEHOLDS ASSET SCORE......... 74 
TABLE 12.7 LINEAR REGRESSION RESULTS FOR THE PREDICTION OF NUTRITIONAL STATUS BY PARTICIPATION 
INTENSITY AND INFANT/CHILD AGE, BY ASSET SCORE CATEGORIES .................................................. 75 
vi
List of Figures
FIGURE 1.1 OPERATIONAL FRAMEWORK: AIN-C PROGRAM ....................................................................... 10
FIGURE 4.1 TIME SPENT TO REACH HEALTH SERVICES ............................................................................. 24
FIGURE 5.1 AGE DISTRIBUTION OF MOTHERS............................................................................................ 27
FIGURE 5.2 MATERNAL EDUCATIONAL LEVELS .......................................................................................... 28
FIGURE 5.3 NUMBER OF LIVE BIRTHS REPORTED ...................................................................................... 28
FIGURE 6.1 ATTENDANCE AT GMP SESSION IN LAST THREE MONTHS FOR CHILDREN > 3 MONTHS OF AGE. 31
FIGURE 6.2 ACTIVITIES AT PREVIOUS GROWTH PROMOTION SESSION AS REPORTED BY CARETAKERS ........ 32
FIGURE 6.3 FREQUENCY OF HOME VISITS AS REPORTED BY CAREGIVERS.................................................. 33
FIGURE 6.4: REASONS FOR LAST HOME VISIT BY AN AIN-C MONITORA...................................................... 33
FIGURE 6.5 SPECIFIC ADVICE GIVEN BY AGE OF CHILD, DURING LAST HOME VISIT ..................................... 34
FIGURE 6.6 ADVICE GIVEN FOR SICK CHILD, BY AGE OF CHILD, DURING LAST HOME VISIT.......................... 35
FIGURE 7.1: SIGNS OF GOOD GROWTH NAMED BY CAREGIVERS, BY PROGRAM PARTICIPATION .................. 38
FIGURE 7.2 SIGNS OF GOOD GROWTH NAMED BY CAREGIVERS WHO GAVE A SINGLE RESPONSE, BY 
PROGRAM PARTICIPATION ............................................................................................................... 39
FIGURE 7.4 ADVICE PROVIDED FOR CHILDREN WITH INADEQUATE GAIN WEIGHT-AIN-C ............................. 41
FIGURE 8.1 INFANTS < THREE AND SIX MONTHS EXCLUSIVELY BREASTFED BY PARTICIPATION ................... 44
FIGURE 8.2 SURVIVAL FUNCTION FOR TIME............................................................................................... 46
FIGURE 8.3 SURVIVAL FUNCTION FOR TIME TO.......................................................................................... 46
FIGURE 8.4 APPROPRIATE FEEDING AMONG CHILDREN 6-8 MONTHS.......................................................... 47
FIGURE 8.5 APPROPRIATE FEEDING AMONG CHILDREN 9-11 MONTHS........................................................ 48
FIGURE 8.6 APPROPRIATE FEEDING AMONG CHILDREN 12-23 MONTHS...................................................... 49
FIGURE 8.7 APPROPRIATE FEEDING OF CHILDREN 6-23 AND 9-23 MONTHS ................................................ 50
FIGURE 8.8 BREASTFEEDING PRACTICE DURING DIARRHEAL ILLNESS......................................................... 51
FIGURE 8.9 BREASTFEEDING PRACTICES DURING ARI ILLNESS.................................................................. 53
FIGURE 9.1 CAREGIVER KNOWLEDGE OF SIGNS OF SEVERE DIARRHEA ILLNESS........................................... 57
FIGURE 9.2 PREVALENCE OF ACUTE RESPIRATORY INFECTION (ARI) IN CHILDREN...................................... 58
FIGURE 10.1: PLACE OF DELIVERY, HOME WITH ASSISTANCE, HOME WITHOUT ASSISTANCE, HEALTH FACILITY 
OR OTHER BY PROGRAM PARTICIPATION ........................................................................................... 62
FIGURE 10.2 POSTPARTUM HOME VISITS BY PROGRAM PARTICIPATION ..................................................... 63
FIGURE 11.1 IMMUNIZATION COVERAGE AMONG CHILDREN 13 THROUGH 23 MONTHS OF AGE BY PROGRAM 
PARTICIPATION ............................................................................................................................... 66
FIGURE 11.2 IRON SUPPLEMENTATION AMONG CHILDREN 4 THROUGH 23 MONTHS OF AGE BY PROGRAM 
PARTICIPATION ............................................................................................................................... 67
FIGURE 11.3 VITAMIN A SUPPLEMENTATION AMONG CHILDREN 6 THROUGH 23 MONTHS OF AGE BY PROGRAM 
PARTICIPATION ............................................................................................................................... 67
FIGURE 11.4 DEWORMING IN CHILDREN AIN-C COMPARED TO “NO-GMP" ................................................. 68
vii
Acknowledgements 
The planning, implementation, analysis, and reporting of an evaluation that tries to capture a 
rapidly expanding and evolving program is a complex undertaking requiring the collaboration 
and hard work of numerous individuals. This evaluation of the AIN-C program is no 
exception, as it depicts results of the initial four plus years of implementation of a dynamic 
program. 
USAID/BASICS is indebted to its partner, The Manoff Group, Inc., for providing invaluable 
support from the inception of AIN-C through the completion of this evaluation report, 
including championing the issues addressed here and advocating the use of this analysis to 
improve global programming in child health and nutrition. 
Although space does not permit a list of all of the professionals who worked in some capacity 
on this evaluation, a few individuals who played key roles in ensuring that this work came to 
conclusion are listed below, along with their role in the evaluation: 
Tom Schaetzel, Data analysis and interpretation, lead author 
Marcia Griffiths, Development of evaluation content, author 
Beth Plowman, Evaluation and data analysis planner, author 
Victoria Vivas de Alvarado, Evaluation planner 
Carlos Villalobos, Evaluation field work coordination, training and quality assurance 
Karen Van Roekel, Evaluation planner, data analysis 
Lydia d’Alois, Evaluation planner 
Jorge Matute, Evaluation sampling 
Miguel Calderon, Evaluation planning, field work logistics 
Evaluation Review Panel: Ed Frangillo, Raphael Perez-Escamilla, Jack Fielder, Judith 
McGuire 
The tremendous effort of the field workers who collected the information reported here and 
those who entered the data, as well as the families who gave their time, is gratefully 
acknowledged. 
More information about the Honduran AIN-C Program can be found at the following websites: 
USAID/BASICS—www.basics.org 
The Manoff Group, Inc.—www.manoffgroup.com 
World Bank/Nutriition in Latin America and the Carribbean—www.worldbank.org/lacnutrition 
viii
Acronyms
AIN Atención Integral a la Niñez (Integrated Care of the Child) 
AIN-C Atención Integral a la Niñez en la Comunidad (Integrated Care of the Child in 
the Community) 
ARI Acute Respiratory Infection 
BASICS Basic Support for Institutionalizing Child Survival Project 
BCG Bacillus Camille-Guerin (Tuberculosis vaccine) 
CDC Centers for Disease Control 
CESAMO Centro de Salud con Medico (Health Center with Doctor) 
CESAR Centro de Salud Rural (Rural Health Center) 
COMSAIN Comunicación en Salud Infantil (Communication in Child Health) 
DHS Demographic and Health Surveys 
DPT Diptheria, Pertussis, Tetanus 
FANTA Food and Nutrition Technical Assistance Project 
GOH Government of Honduras 
GMP Growth Monitoring and Promotion 
IDB Inter-American Development Bank 
IMCI Integrated Management of Childhood Illness 
INE Instituto Nacional de Estadísticas (National Institute of Statistics) 
MDGs Millennium Development Goals 
MOH Ministry of Health 
MMR Measles, Mumps, and Rubella 
NGO Non-Governmental Organization 
ORS Oral Rehydration Solution 
ORT Oral Rehydration Therapy 
PAHO Pan-American Health Organization 
PPS Probability Proportional to Size 
PRAF Programa de Asignación Familiar (Family Assistance Program) 
SES Socio-Economic Status 
TBA Traditional Birth Attendant 
TOT Training of Trainers 
UNICEF United Nations International Children’s Fund 
UPS Unidades Proveedores de Salud (health centers) 
USAID United Status Agency for International Development 
WHO World Health Organization 
1
Executive Summary 
Background 
History of AIN. The Atención Integral a la Niñez (AIN) program in Honduras began in the 
early 1990s when the Ministry of Health (MOH) revised health center1
 norms to require the 
detection of faltering growth2
 in children and to use that indicator for providing services. The 
U.S. Agency for International Development (USAID) has long assisted the Honduran 
government (GOH) in its efforts to increase access to and equity in the use of health 
services. The Atención Integral a la Niñez en la Comunidad (AIN-C) program is a major 
effort to extend the reach and coverage of public health services to poor communities. 
USAID’s support to the MOH to develop and implement AIN-C began in 1991, starting with 
the bilateral Health Sector I and II programs (1991–2000), continuing with the Nueva
Programación agreement between the MOH and USAID for 2001–2003, followed by the 
Sustainable Improvements Program (2003–2005). USAID also provided technical support to 
the MOH for AIN-C through the Basic Support for Institutionalizing Child Survival (BASICS) I 
and II projects, and more recently through the Food and Nutrition Technical Assistance 
(FANTA) project. 
Description of AIN-C. The goal of the AIN-C program is to prevent under-nutrition by 
maintaining newborn health or recuperating the infant with low birth weight by ensuring 
adequate monthly growth (weight gain) until 24 months of age. After that period, AIN-C 
targets children until they are five years old only when they are sick, to be sure that they 
receive timely and adequate attention. All children in the community are eligible at birth, and 
parents are encouraged to participate in the monthly growth monitoring and promotion 
(GMP) session, which includes weighing and counseling based on growth performance, until 
their child reaches 24 months of age. The counseling focuses on adequate nutrition, care, 
and health-seeking practices. When a child is seriously ill or has persistent or acute growth 
failure, the health worker refers the caregiver to the health clinic. Other activities flow from 
the monthly GMP session, making AIN-C a model of community-based growth promotion 
(CBGP): home visits for children who did not attend or who need special attention; reporting 
of child growth information to the community; and detection, assessment, and treatment of 
common childhood illnesses. 
Evaluation Design and Sample 
Original Design. The AIN-C program evaluation plan was developed in 1997 during the first 
phase of BASICS support to the AIN-C program. The original design was quasi-experimental 
and called for pre- and post-intervention comparisons between intervention and control 
communities with baseline, midterm, and final assessments. Data collection for the 2002 final 
evaluation survey employed the same set of survey parameters used in the 1998 baseline 
 
1
 These health facilities are known locally as CESAMOs (Centros de Salud con Médico) or CESARs
(Centros de Salud Rural), depending on whether there is a physician on staff. 
2
 Faltering growth is the failure to gain the expected minimum amount of weight according to 
population-based norms.
2
household survey. Multi-stage cluster sampling was used at baseline. For the follow-up 
evaluation impact survey in 2002, a new random sample of households with children younger 
than 2 years was taken from the same community clusters surveyed in the baseline. The 
actual baseline sample obtained was 1,467 households from 100 communities. The final 
impact evaluation sample obtained from the same communities included 1,343 households. 
Revised Analysis. Due to issues including non-equivalent groups, extensive “contamination” 
of control communities, and reduced intensity of implementation during the evaluation period 
due to an unanticipated diversion of funding to other areas, the pre/post, intervention/control 
community comparison was replaced with an analysis based on individual, community-level 
participation in AIN-C. Using the follow-up data, children who participated in AIN-C were 
compared with children who did not, regardless of the community in which they lived. 
Children who participated in GMP only at a Health Center were excluded so that the 
evaluation focused on the impact of CBGP. The total sample size available for the revised 
analytical approach was 1336 households. With this change, an “intention-to-treat” 
assessment was not possible, but rather this evaluation shows the impact of AIN-C services 
on individuals who received them. Also, no comparison of communities was possible and the 
socio-economic differences between AIN-C and control communities remained. 
Characteristics of Sample 
Household Characteristics. Significant differences were noted between households that 
participated in AIN-C and those who did not (or “No-GMP”). A greater percentage of 
households who enrolled in AIN-C had access to an improved sanitation facility. Generally, 
improved sanitation indicates greater wealth, but in this case, the more likely explanation is 
that AIN-C households were targeted for assistance with sanitation facilities by both 
government and NGOs, as they were in “poorer” communities. Point-of-use water purification 
was more common among AIN-C than “No-GMP” households, possibly a reflection of the 
specific AIN-C project recommendation that families chlorinate their water if any child has 
diarrhea. Additional differences between AIN-C and “No-GMP” households indicate that on 
average, AIN-C households were not as wealthy as the “No-GMP” households. AIN-C 
households lived further from health centers. Significantly more (43.5 percent) AIN-C 
households had dirt floors than did non-participating households (29.9 percent) and fewer 
AIN-C households had electricity or owned expensive durable goods (television, refrigerator, 
telephone, or motor vehicle). Radio ownership was equally common all households. 
Caregiver Characteristics. AIN-C and “No-GMP” caregivers attained different levels of 
education. The percentage of mothers with any formal education was not significantly 
different between AIN-C and “No-GMP” households (p=0.057) however, among those 
mothers who had some schooling, the “No-GMP” group had more: a significantly smaller 
percentage of AIN-C mothers had completed primary or had any secondary education 
compared to the “No-GMP” group. In other respects—size of household, presence of father 
or male in the family—the households were not different. Mothers of AIN-C participants had 
similar reproductive health histories, although parity was significantly higher among AIN-C 
mothers. 
3
AIN-C Implementation 
Age of Entry and Participation. The GMP program was well established in the AIN-C 
communities. Almost 100 percent of the participants in the AIN-C program knew the activities 
supported by the program and the person in the community responsible; 90 percent knew 
this person as the monitora. Indicators related to entry into and participation in the program 
suggested that the quality of implementation was high. Although not quite achieving the 
program goal of registration of all children in the first month of life, more than half were 
registered within two months and 75 percent within three months. Attendance among AIN-C 
participants was consistent: almost 90 percent of AIN-C children had been weighed two or 
more times in the three months prior to the survey. This weighing occurred through monthly 
sessions as well as home visits. Almost 90 percent of growth cards reviewed were found to 
have an accurate interpretation of the growth trend. 
Monthly GMP and Home Visits. Seventy-five percent of caregivers reported receiving 
counseling on breastfeeding and feeding practices during GMP monthly sessions. The 
project called for counseling of all caregivers regardless of the growth trend of the child, so 
this result did not fully meet project expectations. Home visits were another major element of 
the AIN-C program; about 30 percent of caregivers reported receiving a home visit. Many of 
these visits were to follow-up on sick children, and referral to the health center was three 
times more frequent in home visits than in the monthly GMP sessions. Very few caregivers 
(based on the very small sample of mothers who had sick children at the time of the survey) 
recalled receiving messages related to continued and increased frequency of feeding during 
illness, indicating some area for improvement in nutrition counseling for sick children. 
Understanding Adequate Weight-Gain. The caregivers who participated in the AIN-C 
program were significantly more likely than the “No-GMP” group to understand the signals 
(especially weight gain) of healthy growth, and a small proportion—although still significantly 
more than the “No-GMP” group—understood that inadequate weight gain was a sign of poor 
growth. About half of AIN-C participants, compared to a third of the “No-GMP” caregivers, 
responded that gaining weight was a sign of good growth. For a sign of poor growth, 17.1 
percent of AIN-C participants, compared to 12.7 percent of “No-GMP” caregivers, named 
inadequate weight gain as a sign of poor growth. Even among AIN-C participants, very few 
articulated the relationship between poor growth and inadequate weight gain. Given that one 
of the key messages of AIN-C was the relationship between adequate weight gain and 
health, this suggests that the communication of messages either at the level of trainer to 
monitora, or monitora to caregiver, needs additional strengthening. 
Quality of Counseling. More than 80 percent of AIN-C participants whose children were 
faltering in growth received counseling. The specific advice received as reported by these 
caregivers shows mixed results in terms of the appropriateness and specificity of the 
messages. Ten percent of caregivers of children less than six months recalled being 
counseled to give their child more food than accustomed, despite the recommendation of 
exclusive breastfeeding for the first six months. Very few respondents reported messages 
other than “continue breastfeeding” and “give more food than accustomed.” More specific 
messages related to the quality of foods for children over six months of age, such as “give 
thick soups,” “give mashed foods” and “pay attention to the amount eaten” were mentioned 
only 18.1, 13.7, and 6.6 percent, respectively. The “special” counseling cards used in the 
AIN-C program were widely, although not universally, recognized by AIN-C participants. 
While more than 70 percent of AIN-C participants had seen the cards, 45 percent of the “No-
4
GMP” group also recognized the cards. Among the “No-GMP” participants, 91 percent had 
seen the cards in the health center. 
Breastfeeding, Complementary Feeding, and Care during Illness 
Improved Breastfeeding Practices. AIN-C mothers practiced exclusive breastfeeding longer, 
as shown both in terms of the percentage of mothers who were practicing exclusive 
breastfeeding at six months (55.8 percent of AIN-C mothers, compared to 40 percent of “No￾GMP” mothers) and in terms of the median duration of exclusive breastfeeding. Among AIN￾C mothers, the mean duration of breastfeeding was 1.5 months longer than among “No￾GMP” mothers. Furthermore, the rate at which mothers ceased exclusive breastfeeding was 
much slower in the early months and more rapid as six months neared; while true for both 
groups, a greater proportion of AIN-C mothers ceased exclusive breastfeeding closer to six 
months than the “No-GMP” mothers. 
Improved Complementary Feeding Practices. Overall, the AIN-C program had a positive 
impact on the frequency of complementary feeding for children 6–8 months and 9–11 
months; in AIN-C, 98.7 and 94 percent, respectively, met the recommendation compared to 
88.9 and 84 percent, respectively, in the “No-GMP” group. However, even among the AIN-C 
group, where more than 90 percent of children 9–11 months met the recommendation for 
frequency of feeding, only 15.6 percent met the recommendation for the amount of food 
needed. Although almost double the proportion of children in the “No-GMP” group, this 
reflects very few children being fed adequately. Among children 12–23 months of age, the 
AIN-C group showed some improvement in the frequency of providing food compared to the 
“No-GMP” group: 16.1 percent compared to 5.8 percent, respectively. However, this shows a 
dramatic decline in the proportion of children who are meeting the recommendation for 
frequency of feeding compared to the 9–11 month age group. The percentage of children 
who received the appropriate quantity of food dropped in the AIN-C group; there was no 
difference in meeting this recommendation between the AIN-C and “No-GMP” children for 
this age group. The results show that the AIN-C program did not adequately address 
practices related to the amount of food for children 6–23 months. 
Care during Illness. AIN-C mothers were more likely to increase the quantity and frequency 
of feeding and less likely to decrease or cease breastfeeding during diarrheal illness and an 
episode of acute respiratory infection (ARI) than mothers in the “No-GMP” group. About half 
of both the AIN-C and “No-GMP” mothers either increased or maintained the quantity and 
frequency of food provided to their children during diarrheal disease or ARI. The results for 
care-seeking related to ARI show that AIN-C and “No-GMP” participants had comparable low 
rates of care-seeking for ARI; 55.4 percent and 58.8 percent, respectively. The difference 
between the two groups is seen in the length of time it took to seek care among the group 
who ultimately sought care from a trained health provider and in the knowledge of danger 
signs. AIN-C participants took the child for care 2.6 days after the beginning of the illness, 
compared to 3.0 days for the “No-GMP” group. AIN-C caregivers were significantly more 
knowledgeable on danger signs associated with diarrhea and ARI. 
Use of Health Services 
Births and Postpartum Visits. Fewer AIN-C mothers used facilities for delivery of infants, 
likely a reflection of the targeting of the AIN-C program toward communities with greater 
socio-economic needs. At the time of this evaluation, AIN-C participants received 
significantly more postpartum visits than “No-GMP” mothers, although the percentage of AIN-
5
C mothers was relatively small (not quite 40 percent). The main postpartum activities, for all 
who made visits, including the monitoras, were examination of the mother and baby. 
Enrollment in AIN-C did not occur through these visits. 
Vaccination and Micronutrient Supplementation. Significantly more children in AIN-C had 
records of their immunizations than non-enrollees. Children enrolled in AIN-C had higher 
rates of complete immunization coverage at 13 months than non-enrollees. Children enrolled 
in AIN-C were far more likely to receive iron (65.6 percent [AIN-C]; 29.5 percent [“No-GMP”]) 
and vitamin A (94.3 percent [AIN-C]; 87.5 percent [“No-GMP”]) supplementation than were 
children not enrolled in any form of growth promotion. 
Impact of AIN-C on Nutritional Status 
Bivariate Analysis Suggests Positive Impact on Nutrition. Descriptive results (mean and 
median z-scores) show that AIN-C participants did not differ from the “No-GMP” group in 
terms of weight-for-height (wasting), but AIN-C children in the 0–6 month age group did have 
lower height-for-age and weight-for-age z-scores than their “No-GMP” counterparts, as did 
the combined 0–23 month age group. This analysis did not account for differences in socio￾economic status and age, both of which are correlated with nutritional status. The major 
difference in nutritional status between the two groups is in height-for-age under six months. 
Children 0–5 months of age have statistically lower z-scores in the AIN-C group compared to 
the “No-GMP” group, while there are no differences in nutrition status in the 6–11 and 12–23 
month age groups. This suggests that AIN-C participation protects against height-for-age 
faltering. AIN-C participating infants may begin life shorter-for-age than the “No-GMP” group, 
but they “catch-up” to the “No-GMP” children by the age of 6–11 months, indicating more 
rapid height gain among AIN-C participants. This conclusion is not definitive, since a 
pre/post-intervention comparison of height-for-age status of these two groups was not 
possible. 
Controlling for Age and Socio-Economic Status Shows Positive Impact of AIN-C. There are 
highly significant differences between AIN-C and “No-GMP” groups in terms of asset score, a 
reasonable proxy for economic status. In addition, the samples of the two groups differ in 
distribution of ages of children: AIN-C participation is more common for younger 
infants/children than for older ones. Analyses were restricted to infants/children from either 
group whose household asset score was equal to or lower than the median value for both 
groups combined—the poorer half of all households. Comparisons between groups were 
controlled for infant/child age by including age in a multiple linear regression predicting 
nutritional status. 
Infant/child age was highly and negatively associated with all measures of nutritional status: 
with each additional month of age, weight-for-height, height-for-age, and weight-for-age 
decreased 0.057, 0.073, and 0.072 Z-score, respectively. No negative association between 
AIN-C participation and height-for-age and weight-for-age is seen in the controlled, restricted 
analysis. A positive and marginally significant (p<0.10) effect of AIN-C participation on 
weight-for-height was observed in the controlled, restricted analysis (participation was 
negatively associated with weight-for-height in uncontrolled comparisons). The mean weight￾for-height Z score is 0.122 higher among AIN-C participants, holding age constant and 
excluding wealthier households. 
6
Participation Increases Impact on Nutritional Status. The intensity of participation was 
positively associated with nutritional status for all measures: strongest for weight-for-age 
(p=0.004), less strong but still highly significant for weight-for-height (p=0.011), and 
marginally significant for height-for-age (p=0.083). Holding infant/child age and assets scores 
constant, for every 1 percent increase in participation intensity, weight-for-age increased 
0.005 Z-score, and weight-for-height increased 0.004 Z-score. Age and participation intensity 
explain a much greater proportion of the variability in nutritional status among poorer 
households than among wealthier households. The size of the positive effect of participation 
intensity is 2–3 times greater among poorer households than it is among wealthier 
households. The association between participation intensity and nutritional status is highly 
significant for weight-for-height and weight-for-age among poorer households, while no 
association exists between participation intensity and nutritional status among wealthier 
households. 
7
1. Introduction 
1.1 Situational Analysis of Honduras 
Honduras is a Central American republic with an estimated population of 7.2 million and an 
annual growth rate of 2.4 percent.3
 The per capita gross national income of $1030 in 2004 
makes Honduras the third poorest country in Latin America, after Haiti and Bolivia. Fifty-one 
percent of the population is below the poverty line and seventy-five percent of rural 
households are extremely poor. Although Honduras is modernizing, almost one-half of the 
population resides in rural areas. About eighteen percent of Honduran households have no 
connection to a water supply system and almost one-third lack an adequate sanitary facility.4
 
The infant mortality rate in Honduras declined from 34 to 30 per 1,000 live births between 
2001 and 2005 as a result of improvements in post-neonatal mortality. There has been no 
change in the level of the neonatal contribution to infant mortality.5
 Infant and child deaths 
are concentrated in the rural areas, particularly among mothers with no or little formal 
education. Diarrhea with dehydration and acute lower respiratory infection are two leading 
causes of child death. The prevalence of both these illnesses has increased since 1996, 
especially that of acute respiratory infection (ARI). Contributing to the under-five mortality 
rate is under-nutrition in young children. Nationally, chronic under-nutrition, measured by 
height-for-age, is 34.5 percent (based on 2001 data)—about double the average for Latin 
America. Overall, under-nutrition as measured by weight-for-age was 12.5 percent in the 
same period. These national statistics mask much higher levels of chronic and under￾nutrition in some regions of the country. Chronic under-nutrition was over 50 percent and 
overall under-nutrition was 20 percent or more in Regions Two and Five6
 in the western part 
of the country.7
Public health services are provided by the Secretaría de Salud (Ministry of Health, or MOH). 
The MOH is responsible for implementation of major health strategies at national and local 
levels. Central headquarters carries out health programs, oversees regulatory issues, and 
provides technical support to its own regional offices and the private subsector. Major on￾going health programs by the MOH include the HIV/AIDS Prevention program; the Sexually 
Transmitted Infections Control Program; the Tuberculosis Control Program; the Cervical 
Cancer Control Program; the National Oral Health Program; and the Expanded Program on 
Immunization. The Comprehensive Family Health Care Department of the MOH administers 
women’s health programs (prenatal care, deliveries, postpartum care, and family planning 
services) and programs for children under the age of five and for adolescents. 
 
3
 Instituto Nacional de Estadisticas (INE). Proyecciones de Población 2001-2005. Tegucigalpa:INE; 
2004. 
4
 Honduras, Instituto Nacional de Estadistica. Censo de Población y Vivienda 2001. Tegucigalpa: INE; 
2003. 
5
 Honduras, Instituto Nacional de Estatdistica. Proyecciones de población 2001-2015, Tegucigalpa; 
INE; 2003. 
6
 In May 2004 the MOH was reorganized along functional lines. 18 Departmental Health Regions and 
2 Metropolitan Health Regions were created to coincide with the country’s political-administrative sub￾divisions. 
7
 World Health Organization, who.int/nutgrowthdb/database/countries/who_standards/hnd.pdf 
8
The U.S. Agency for International Development (USAID) has been a major supporter of the 
efforts of the MOH, and continues to provide assistance in basic child health prevention and 
treatment services in the most vulnerable areas of the country. USAID has long assisted the 
Honduran government (GOH) in its efforts to increase access to and equity in the use of 
health services. The Atención Integral a la Niñez en La Comunidad (AIN-C)program is a 
major effort to extend the reach and coverage of public health services to poor communities. 
USAID’s support to the MOH to develop and implement AIN-C began in 1991, starting with 
the bilateral Health Sector I and II programs (1991–2000), continuing with the Nueva
Programación agreement between the MOH and USAID for 2001–2003, followed by the 
Sustainable Improvements Program (2003–2005). USAID also provided technical support to 
the MOH for AIN-C through the Basic Support for Institutionalizing Child Survival (BASICS) I 
and II projects, and more recently through the Food and Nutrition Technical Assistance 
(FANTA) project. 
1.2 History and Development of the AIN and AIN-C Programs 
The AIN program in Honduras began in the early 1990s when the MOH revised health 
center8
 norms to require the detection of faltering growth9
 in children, and to use that 
indicator for providing services. By 1994, the majority of health centers had adopted the 
growth promotion approach to well child visits. A review of the program in 1994 led to the 
realization that public health impact could not be achieved by limiting services to the facility 
level. This led to the development, testing and implementation of AIN-C— the community 
AIN approach. In December 2000, the MOH issued a decree establishing both the facility 
and community components of AIN as the National Child Health and Nutrition Program. 
Guidance issued with the decree specified that AIN-C was first to focus on prevention and 
early detection and referral of sick children under two, and then to introduce the illness 
management model for all children under five. Beginning in 2001, a number of NGOs and a 
government entity, PRAF (Programa de Asignación Familiar—an income transfer program of 
the President’s office funded by the Inter-American Development Bank) supported the 
implementation of AIN at the community level. 
USAID support to the MOH for AIN-C implementation continued over the evaluation period 
1998–2005, although the health areas receiving assistance shifted over this time period. 
Under USAID’s Health Sector II support, nine health areas were supported from 1998–2001. 
In 2001, when USAID renegotiated its assistance, four of these nine health areas were 
dropped, while five new areas were added. From 2001–2005, USAID supported 10 health 
areas. All tallied, by 2002–2003 there were 24 different health areas that had received some 
kind of support for AIN-C implementation from a variety of financial sources. By the time data 
were collected for this evaluation, two of the six areas that participated in the baseline survey 
and were part of the impact evaluation design had been without extra-budgetary support for 
the AIN-C program for almost one year.10 See Table 1.1 for a timeline of the history and 
 
8
 These health facilities are known locally as CESAMOs (Centros de Salud con Médico) or CESARs
(Centros de Salud Rural), depending on whether there is a physician on staff. 
9
 Faltering growth is the failure to gain the expected minimum amount of weight according to 
population-based norms. 
10 In 2004, the MOH received resources from both the Inter-American Development Bank (IDB) and 
the World Bank to expand AIN to the most disadvantaged communities nationwide. The AIN program 
also was included in the MOH National Nutrition Strategy to help meet the Millennium Development 
Goals (MDGs). Today, in 2008, AIN-C is present in all health areas and continues to expand into new 
communities, while an effort is made to give periodic support to communities in areas that do not have 
extra budgetary support. 
9
development of the AIN-C Program, as well as important events that occurred during the 
course of the initial implementation of AIN-C. 
Table 1.1 Timeline of Events in the Development and Implementation of AIN-C: 1990–
2002 
Year Event or Activity 
1990 Secretariat of Health replaced nutrition status tracking with growth faltering as the indicator of 
child health at the clinic level—beginning of AIN. 
1992-
1993 
Pilot testing of the community component of AIN in a few communities in Copan, Lempira, and 
La Paz. 
1994 Review of clinic-based strategy and recognition of need to reach the community. 
1995 BASICS I study of initial experience with AIN-C finds strong commitment among health workers 
and high participation rates among mothers, but inconclusive results. 
1996 National Workshop convened by MOH identified need for: systematizing program operations, 
incorporating inter-country lessons learned, developing tools for volunteer workers and 
supervisors, creating a training system, and linking AIN with national strategy for equity in 
health care delivery (ACCESO11).
1996–
1997 
BASICS I helped produce: guide for MOH personnel, manual for volunteer community workers, 
training curriculum, TOT (training of trainers) guide, series of 20 counseling cards (láminas). 
1998 Re-launch of the revised community program began in nine USAID-supported health areas in 
most populated and poorest regions. Baseline survey conducted prior to implementation. 
1998 Introduction of WHO/UNICEF Integrated Management of Childhood Illness (IMCI) protocols at 
facility level to assess, classify, and treat sick children under 5. 
1998 Hurricane Mitch hit many communities that had just begun or were scheduled to begin AIN-C; 
implementation stopped for three months. When program continued, reports indicated that 
communities with AIN-C in hurricane-affected areas “recuperated” the growth of children more 
quickly than non-AIN communities. 
1999 AIN-C initiated in the hardest hit health areas with the extra-budgetary support from Mitch 
reconstruction. Ten additional health areas were fully trained in AIN-C and an initial set of 
communities began with USAID support in 1999. 
1999–
2000 
Integration of IMCI protocols and training within existing AIN protocols at the facility level. 
Extension of IMCI to the community under AIN-C; training for community workers and 
supplying them with antibiotics, oral rehydration solution (ORS), and micronutrients for all 
children under five. 
1999–
2000 
Mitch resources used to develop communication campaign (COMSAIN) to support health￾promoting behaviors identified by AIN-C with a strong emphasis on hygiene and diarrheal 
disease control. Program launched with print materials and radio. 
2000 Mid-term Evaluation in 60 of 100 original communities showed improvements in knowledge 
and behavior related to childcare and feeding practices. 
2001 World Bank-supported pilot project and several NGOs adopt AIN-C; interagency AIN-C 
implementation group formed. 
2002 COMSAIN expanded nationally. 
 
11 A project to improve health care coverage in disadvantaged districts
10
1.3 Description of the AIN Program 
The goal of the AIN-C program is to prevent under-nutrition by maintaining newborn health or 
recuperating the infant with low birth weight by ensuring adequate monthly growth (weight 
gain) until 24 months of age. After that, children are covered under the program until they are 
five years old only when they are sick, to be sure that they receive timely and adequate 
attention. All children in the community are eligible at birth, and parents are encouraged to 
participate in the monthly growth monitoring and promotion (GMP) session, which includes 
weighing and counseling based on growth performance, until their child reaches 24 months 
of age. The counseling focuses on adequate nutrition, care, and health-seeking practices. 
When a child is seriously ill or has persistent or acute growth failure, the health worker refers 
the caregiver to the health clinic. The nurse auxiliary from the health post often attends the 
GMP sessions, where she updates immunizations, distributes micronutrient supplements, 
provides medicines to children, and discusses family planning with new mothers. Other 
activities flow from the monthly GMP session, making AIN-C a model of community-based 
growth promotion (CBGP): home visits for children who did not attend or who need special 
attention; reporting of child growth information to the community; and detection, assessment 
and treatment of common childhood illnesses. 
The AIN-C operational framework is shown in Figure 1.1 below. In addition to what happens 
in the community, growth information is shared with municipality officials, and they support 
community activities that foster healthy growth in children. 
Figure 1.1 Operational Framework: AIN-C Program 
Health Center Nurse / NGO
Mobilizes 
community
Community identifies and 
supports a team of 
volunteers; encourages all 
families to have <2s 
weighed each month and 
all sick <5s taken to 
volunteers for care
Each month volunteers: 1) weigh children (0-
2 yrs); 2) compare weight to expected weight 
to detect faltering; 3) counsel mothers; 4) 
treat and/or refer sick children; 5) make home 
visits to children who fail to get weighed at 
the session, fail to gain weight or are sick; 6) 
treat on demand children 0-5 for illness; 7) 
support health system programs; 8) report 
program information to the health center.
Community takes collective action to address problems 
affecting child growth (unsafe water, child care and food insecurity)
Provides health services in the 
community (e.g., immunization, 
micronutrients, 
family planning)
Monitors program 
progress 
(participation, 
immunization 
coverage, children 
failing to grow)
Volunteers 
inform 
community 
about progress 
(participation, 
child growth)
Trains, supervises and 
provides material 
support for volunteers
Health Sector support Municipal/local government support
 Source: Manoff Group, 2004 
11
1.3.1 AIN-C Training and Supervision 
A week-long training for health area staff and nurses initiates the program in a health area to 
ensure that those who will be supervising the program are able to support it. Five days of 
training are also provided to the community volunteer and community-level auxiliary health 
staff. Monthly meetings also serve as in-service trainings. Supervision is designed to be 
supportive and regular: for the first six months of implementation, a supervisor attends the 
community weighing sessions. A key practice of the program is the use of a team of 
volunteers in the community rather than relying on one person. This aspect of program 
design ensures that the work is not too burdensome for one person and adds to overall 
capacity by drawing on the different strengths of the various team members. 
1.3.2 AIN-C Program Monitoring 
A simple information system is an important management tool of AIN-C. At the end of each 
month, growth information from all of the children in a community is compiled into bar graphs 
that use five simple indicators: number of children under two in the community, number 
weighed that month, number gaining adequate weight, number with inadequate weight gain, 
and number gaining inadequate weight for two or more months.12 These indicators help 
volunteers to target home visits, focus supervision, and mobilize the community. These data 
are also shared within the health system and sent to municipal authorities. 
By March/April 1998, the initial group of communities had concluded the preparatory phase 
and begun the work of the program, seeing children under two monthly to detect and resolve 
their health problems. The evaluation described in this report looks at the impact of the first 
communities to implement the AIN-C program in areas of the country where USAID 
collaborated with the GOH on implementation. 
 
12 An adaptation of the SKDN system used in Indonesia. SKDN is the abbreviation for the columns in 
the bar chart; S=number of children, K=number of children enrolled in the program, D=number of 
children weighed, N=number of children who gained adequate weight. 
12
Program 
Exposure 
Program 
Participation 
Improved 
Knowledge 
& Attitudes 
Improved 
Practices 
Improved 
Nutrition 
& Health 
2. Methodology 
2.1 Conceptual Framework for the Impact Evaluation 
Figure 2.1 describes the conceptual framework for the impact evaluation of the AIN-C 
program in Honduras. The design was an “intention-to-treat” analysis of the effect of 
available program services on child nutrition and health, mediated through caregivers’ 
participation in the program, caregivers’ knowledge and attitudes about child health and 
nutrition, and caregivers’ actions promoting and protecting child nutrition and health. 
Figure 2.1 Conceptual Framework for the AIN-C Evaluation 
2.2 Original Evaluation Design 
The AIN-C program evaluation plan was developed in 1997, during the first phase of BASICS 
support to the AIN-C program. The original design was quasi-experimental and called for 
pre- and post-intervention comparisons between intervention and control communities with 
baseline, midterm, and final assessments. Data collection for the 2002 final evaluation 
survey employed the same set of parameters used in the 1998 baseline household survey. 
The design was based on the following assumptions: 
1. Reasonably uniform CBGP services were provided in all areas targeted for the 
introduction of AIN-C, and GMP services were absent in areas not targeted for AIN-C. 
2. Caregivers participated when and where services were available, and caregivers did not 
participate when and where services were unavailable. 
3. Basics health services (e.g., immunizations) were received by those who participated in 
the program, as well as individual counseling. 
4. Individual counseling improved knowledge and attitudes among caretakers. 
5. Improved knowledge and attitudes resulting from individual counseling led to improved 
caregiver practices. 
6. Improved nutritional and health status resulted from improved caregiver practices. 
The results presented in the sections below demonstrate the extent to which each of these 
assumptions proved to be true. 
Internationally-recognized indicators related to improved feeding practices for children less 
than two years of age, as well as improved care-seeking and treatment for diarrhea and 
acute respiratory illness (ARI) are among the priority outcome and impact indicators that are 
measured in this evaluation. The following is a list of the key indicators measured in this 
evaluation: 
13

 Proportion of children 0–23 months of age with diarrhea who received oral 
rehydration solution (ORS); 

 Proportion of children 0–23 months of age with diarrhea who received increased 
fluids; 

 Proportion of children 0–23 months of age with diarrhea who continued feeding; 

 Proportion of children 0–23 months of age with cough and difficult and/or rapid 
breathing who were taken early to a trained provider; 

 Proportion of children fully immunized by 13–23 months; 

 Proportion of children receiving vitamin A and iron supplementation; 

 Proportion of infants less than six months exclusively breastfed; 

 Proportion of children 6–23 months who are appropriately fed for their age; 

 Height-for-age z scores; 

 Weight-for-age z scores; 

 Weight-for-height z scores. 
2.2.1 Sampling Design 
Multi-stage cluster sampling was employed at baseline as follows (see Table 2.1): 
1. Random selection of health areas in three urbanization strata (urban, mixed, and rural);13
2. Probability proportional to size (PPS) random sampling of Unidades Proveedoras de 
Salud (UPS, or health centers); 
3. PPS random sampling of communities from a MOH list of those selected for the 
introduction of AIN-C in the first year. Communities with fewer than 40 households and, 
at the direction of the MOH, those too distant from Health Centers for appropriate 
monitoring, were excluded from the list; 
4. Random sampling of households based on a community map list showing households 
with children younger than two years of age. One child was sampled per household. 
Individual randomization was not possible, as is the case with most evaluation studies of 
social service programs. For the follow-up evaluation impact survey in 2002, a new random 
sample of households with children younger than two years was taken from the same 
community clusters surveyed in the baseline. 
 
13 The purpose for creating urbanization strata was to ensure that the range of AIN-C program 
contexts were captured, but the categories are relative rather than based on a standard definition. 
That is, they characterize the degree of urbanization in the health areas relative to each other only. 
The majority of the communities in five of the six selected health areas were in fact predominantly rural 
(Metro San Pedro Sula being the exception). 
14
Table 2.1 Summary of the Levels and Process of Selection of the Sample 
Level Selected Criteria Type of sampling 
1 Health Areas 3 urbanization strata 
USAID support 
Random within stratum 
2 UPS None PPS* random 
AIN-C 
Communities
AIN-C current or planned** 
40+ households 
3 Random 
Control 
Communities 
Similar in size, government Socio-Economic Status 
(SES) ranking and UPS proximity 
No planned AIN-C 
Matching 
4 Household At least 1 child < 2 years Random 
* Probability proportional to size; ** Generally, based on government SES ranking 
2.2.2 Sample Size 
To estimate sample size, the minimum programmatically meaningful effect size for nutritional 
indicators (e.g., stunting) was set at 10 percent. The one-tailed probability of a Type I error 
and the power of the test were fixed at 5 percent (α = 0.05) and 80 percent, respectively. 
Because the variance of the key nutritional indicators was unknown, an estimated variance of 
0.25 for a dichotomous variable (assumed prevalence of 50 percent) was used, with an 
estimated design effect of 2 resulting from cluster sampling. Non-response/refusal was 
estimated at 20 percent. Based on these parameters, a minimum sample of 1,350 
households was necessary, and a target sample of 100 communities was adopted (50 AIN￾C, 50 control; 15 households/village yielding a total of 1,500 households). The actual 
baseline sample obtained was 1,467 households from 100 communities. The final impact 
evaluation sample obtained from the same communities was 1,343 households. 
Eight of the original 100 baseline communities were omitted from the follow-up survey. 
Reasons for removing these communities were that some “AIN-C” designated communities 
did not implement the program and thus were removed along with their matched controls. 
Damage from Hurricane Mitch in 1998 completely destroyed one control community, 
necessitating its elimination along with its AIN-C pair. 
2.3 Issues with the Original Evaluation Design and Analysis 
For several reasons, the original quasi-experimental, pre/post-intervention comparisons 
between intervention and control communities were not valid. 
2.3.1 Non-Equivalent Groups
Analysis of the baseline data indicated that matching failed: intervention and control 
communities differed significantly according to the distance and size matching criteria. In 
addition, intervention and control communities differed according to a number of socio￾economic factors because MOH targeting criteria for AIN-C gave priority to the most 
15
disadvantaged communities.14 Nutritional status did not differ between the AIN-C and control 
communities at baseline; however the differences in distance/size matching criteria and in 
economic status may have influenced endline nutritional status. If such bias did occur, its 
magnitude and direction are unknown. 
2.3.2 “Contamination” of Control Communities 
Over the course of the evaluation period, spillover of AIN-C activities increasingly 
“contaminated” the control areas: AIN-C and matched control communities were within the 
same Health Center catchment area; health Centers overseeing AIN-C participating 
communities improved GMP activities at the Health Center; services offered by the Health 
Center staff lacked the community-based aspects associated with AIN-C (e.g., collective 
response to non-behavioral causes of malnutrition), but were otherwise similar; and nurses 
were trained as AIN-C supervisors and had easy access to the counseling cards and other 
AIN-C tools. As a case in point, the final evaluation found that 48 percent of respondents in 
control communities recognized the laminas, and 88 percent of those who recognized them 
had seen them at the CESAR/CESAMO (health center). 
The MOH also decided to introduce AIN-C into some communities that were designated at 
baseline as control communities. In these cases, the GMP intervention received by children 
in the control area was identical to that being evaluated in the intervention areas. At the 
follow-up survey, 8.4 percent of children in the control communities were enrolled in AIN-C. 
The original evaluation design failed to anticipate these “contaminating” influences; however, 
these issues are not uncommon in evaluation research. Furthermore, AIN-C (and facility￾based growth promotion) services are presumed to be beneficial—even to save lives—and 
restricting the availability of these services for the sake of analytical purity would have been 
unethical. The spillover of GMP and the introduction of AIN-C into the control group may 
have been undesirable for the AIN-C evaluation, but it was desirable for child health and 
nutrition. 
2.3.3 Reduced Intensity of Intervention during the Evaluation Period 
Although USAID provided continuous support to the MOH for AIN-C throughout the 
evaluation period, four of the original six selected Health Areas lost AIN-C assistance when 
funds were shifted to support activities in other areas in 2001 (see Table 2.2). By the time of 
the final evaluation survey, these four areas had been without extra-budgetary support for 
almost one year. 
 
14 The Government of Honduras implemented AIN-C under the ACCESO program, which ranked 
communities according to a variety of socio-economic factors. AIN-C communities tended to rank in 
the lowest two of ACCESO’s five socio-economic categories. 
16
Table 2.2 Bilateral USAID Program Support in Health Areas Included in the Evaluation 
Health Areas included in AIN-C evaluation Years of USAID Support for AIN-C 
Metro San Pedro Sula 1998–Sept. 2001 
Choloma/Lima 1998–Sept. 2001 
Puerto Cortes 1998–Sept. 2001 
Santa Cruz de Yojoa 1998–Sept. 2001 
Siguatepeque 1998–present 
La Paz 1998–present 
This withdrawal of funds influenced the intensity of supervision and the involvement of health 
center personnel in AIN-C. Perhaps as a result, at the follow-up survey in 2002, 17 percent of 
eligible children in the AIN-C communities were not enrolled in a growth promotion program 
of any kind (community-based or at the Health Center). The extent to which this non￾participation rate resulted from changes in funding (or simply from mothers’ choices not to 
participate) is unknown. 
Non-participation or drop-out, for whatever reason, is an important element of intention-to￾treat analysis, which attempts to understand the overall effect of available treatment (AIN-C) 
on the community, not just the effect on the individuals who receive it. Had funding remained 
constant, a strong case could be made that the non-participation rate is characteristic of the 
AIN-C project in Honduras. However, because of the withdrawal of funding, equally plausible 
is that the high non-participation rate existed because services were not characteristic of 
AIN-C. 
2.4 Revised Data Analysis 
Due to these issues—non-equivalent groups, extensive contamination of control 
communities, and reduced intensity of implementation during the evaluation period—the 
pre/post, intervention/control community comparison was replaced with an analysis based on 
individual, community-level participation in AIN-C. Using the follow-up/final evaluation data, 
children who participated in CBGP were compared with children who did not, regardless of 
the community in which they lived. Children who participated in growth promotion only at a 
Health Center were excluded, so that the evaluation focused on the impact of community￾based growth promotion. 
Analysis by participation was deemed the best option based on the issues in the sample; 
however, this type of analysis has several important ramifications that must be considered in 
the interpretation of the results. 
1. Analysis by participation does not provide the desired “intention-to-treat” assessment of 
the AIN-C program’s impact. Rather, it provides an assessment of impact of AIN-C 
services on individuals who receive them. 
17
2. Since “treatment” (AIN-C) was allocated to communities, not individuals, proper analysis 
would require comparisons between communities, not individuals—but this analysis is not 
possible. 
3. While analysis by participation may circumvent the problems caused by contamination of 
the control group and changes in the intervention (e.g., changing levels of external 
support for Health Area), the problems due to failed matching and the socio-economic 
differences between AIN-C and the control group remain. Most AIN-C participants 
resided in AIN-C communities and most non-participants resided in the original control 
communities. The socio-economic differences that exist between these communities are 
more likely to have reduced the observed impact of the intervention, rather than 
enhanced it. 
2.5 Design and Content of the Questionnaires 
The household and individual child questionnaires15 used in the follow-up impact evaluation 
survey were based closely on the questionnaires used for the baseline in 1998. They were 
modified slightly for the following reasons: 
• The rewording of some questions was needed to allow better distinction between the 
activities of AIN monitoras and those of traditional birth attendants (parteras) and 
CESAR/CESAMO health center staff; 
• Some questions were adjusted to allow for pre-coding of responses; 
• Questions were added to obtain more information on the programmatic aspects of AIN; 
• The order of the questions in the questionnaire was revised to create a more logical flow 
including better skip patterns; 
• A section on participation in community activities, including growth promotion programs, 
was moved from the individual questionnaire to the household questionnaire to ensure 
that responses relating to the AIN monitoras were clearly identified from the beginning of 
the interview. 
The household questionnaire contained items that addressed general information on the 
child’s environment including aspects of housing, socioeconomic status, age and gender 
composition of the family, and participation in community social assistance programs. 
The child questionnaire focused on the “index child” of each household: a child under two 
years of age living in the household at the time of the survey. This questionnaire contained 
questions on the following topics: 

 Characteristics of the caregiver 

 Delivery and postpartum care 

 Vaccinations and micronutrient 
supplementation 

 Growth and development 

 Diarrheal disease 

 ARI 

 Breastfeeding 

 Feeding practices 

 Height and weight measurements for 
the child 
 
15 Copies of the household and individual questionnaires are available at www.basics.org 
18
2.6 Training, Field Work, and Logistics 
A senior Honduran program manager with extensive experience in field research led the final 
evaluation survey operations. Four study teams collected the data; each team included one 
supervisor, two interviewers, a dietary interviewer, a person responsible for anthropometry, 
and one driver. The majority of the personnel in these field teams were experienced in 
conducting surveys on maternal and child health. 
Supervisors received training in the content of the evaluation and approach to completing the 
questionnaires. A manual was prepared and distributed at the training. A separate training 
session was conducted on anthropometry. 
Training on the content of the questionnaire and interview methodology took place from June 
3–7, 2002. The classroom training included: 

 Discussion of the technical content of the interviews; 

 Review of each section of the questionnaire, with explanations of the concepts 
involved and the manner in which to ask the questions and record responses; 

 Discussions on the AIN model and the job of the monitoras to assess and record 
weights for children; 

 Discussions on the content of counseling provided to mothers; and 

 Role plays of interviewing techniques. 
Interviewers and supervisors field-tested and validated the questionnaire to test its 
functionality and to complement the classroom training. 
 
Fieldwork began on June 12, 2002 and continued until July 22, 2002. Supervisors were 
supplied with logistical information to facilitate travel and overnight stays for the fieldwork 
teams, with listings of children under the age of two in the study communities, and with maps 
of the study communities. The maps were prepared in the weeks prior to the initiation of 
survey field work by two cartographers, and provided the interview teams with color-coding 
for all households in the community having a child under two years of age. 
Upon completion of each interview, the interviewers reviewed the questionnaires before 
leaving the household. Each child’s weight and recumbent length was measured using 
standard anthropometric techniques and recorded on the individual’s general questionnaire. 
At the end of the day’s work, the survey teams met to exchange experiences, review the 
questionnaires with their supervisors, and make any necessary corrections before leaving 
the community. The supervisors and the study coordinator rotated among the interviewer 
staff to observe interviews and ensure high standards of data quality. 
2.7 Data Handling and Analysis 
Data entry was conducted using the EpiInfo Program (Version 6) developed by the U.S. 
Centers for Disease Control (CDC). Data editing, entry and cleaning were handled by a team 
comprised of a Data Coordinator, three Data Entry Clerks, and a Data Entry Supervisor. 
Oversight and technical guidance were provided by a BASICS II Technical Officer. 
All questionnaires were double-entered for validation purposes. Data entry began concurrent 
to the survey fieldwork and was completed in mid-August 2002. General guidelines on data 
cleaning were prepared and used to guide the programmer responsible for the cleaning process. 
19
2.8 Statistical Analysis 
Statistical analysis was conducted using SPSS for Windows Version 12.0. Bivariate 
relationships between variables of interest and participation status were tested using 
Pearson Chi-Square tests and analysis of variance methods. Specific p values for test 
statistics appear as footnotes to the tables displaying the results or in the discussion of the 
results in the text. All data presented in this report are unweighted. 
For some program outcome and impact indicators, further analyses were carried out using 
regression in an attempt to control for socio-economic differences between participants and 
non-participants. Nominal or linear regression models were employed as appropriate for 
controlled analysis of the relationship between participation and diarrhea treatment, health 
care-seeking behavior, exclusive breastfeeding, young child feeding practices and nutritional 
status. 
20
3. Description of the Sample
3.1 Total Sample for Revised Analysis 
The total sample size obtained for the final evaluation using the revised analytical approach 
to overcome the issues in the original evaluation design was 1336 households. This is 
somewhat less than the projected sample size of 1500 households described in the survey 
design, but it is adequate to show statistically significant change. The final evaluation sample 
is comprised of 603 AIN-C and 733 no-GMP households. Table 3.1 shows the distribution of 
the sample by health area. 
Table 3.1 Final Evaluation Sample Size, by Health Area 
Health Area AIN-C No GMP Total 
% of 
 total 
Number of 
Households 
% of 
total 
Number of 
Households 
% of 
total 
Number of 
Households 
Metro San Pedro Sula 10.8 65 11.2 82 11.0 147 
Choloma/Lima 15.3 92 14.7 108 15.0 200 
Puerto Cortés 17.7 107 15.8 116 16.7 223 
Siguatepeque 27.4 165 25.6 188 26.4 353 
Santa Cruz de Yojoa 14.9 90 13.9 102 14.4 192 
La Paz 13.9 84 18.7 137 16.5 221 
Total number of 
households 603 733 1336 
3.2 Sample by Gender and Age 
The total number of children in the impact evaluation sample for which there was a gender 
designation was 1211. Overall, there were slightly more males than females in the sample, 
52.7 percent versus 47.3 percent, respectively. This proportion of males to females was 
similar in the AIN-C and the no-GMP groups, as shown in Table 3.2 below. 
Table 3.2 Final Evaluation Sample Size, by Sex of the Index Child 
AIN-C No GMP Total 
Sex of the child % of 
total 
Number of 
Children 
% of 
total 
Number of 
Children 
% of 
total 
Number of 
Children 
Male 52.3 309 53.1 329 52.7 638 
Female 47.7 282 46.9 291 47.3 573 
Total number of 
children 591 620 1211 
21
The final evaluation sample was distributed across age groups as shown in Table 3.3. In 
each household, the index child selected was the youngest child of the children under two 
years of age living in the household. 
Table 3.3: Final Evaluation Sample Size, by Age Group of Child 
AIN-C No GMP Total 
Age of the child in months % of 
total 
Number 
of 
children 
% of 
total 
Number 
of 
children 
% of 
total 
Number 
of 
children 
 0 to 5 20.3 120 21.7 135 21.0 255 
 6 to 11 30.1 178 26.8 167 28.4 345 
12 to 17 26.6 157 22.2 138 24.3 295 
18 to 23 23.0 136 29.3 182 26.2 318 
Total number of 
children 591 622 1213 
22
4. Characteristics of Households 
The surveys collected data on the basic measures of housing that describe the source and 
type of water used, the type of sanitary facilities available to the household, the number of 
rooms and bedrooms in the house, whether one room was dedicated to use as a kitchen, the 
type of fuel used in cooking, the presence of certain appliances, and the type of flooring. The 
time and cost of reaching the nearest health center were also queried as an indication of 
household access to health services. 
4.1 Water and Sanitation 
As seen in Table 4.1, by far the most common source of water for both AIN-C (77.3 percent) 
and “No-GMP” (78.5 percent) households was a tap located inside the house or on the 
property. There were no significant differences between the two groups in water sources. 
Significantly more AIN-C (41.8 percent) households purified their water compared to “No￾GMP” households (32.1 percent). Overall, almost two-thirds of households consumed water 
directly from the source with no further treatment. AIN-C households (58.2 percent) were 
significantly more likely to have an improved toilet (indoor, sealed latrine) than “No-GMP” 
households (50.5 percent). 
Table 4.1 Water and Sanitation Characteristics of Households 
AIN-C NO-GMP TOTAL 
% Number of 
Households
% Number of 
Households
% Number of 
Households
Principal source of water 
Tap inside house or on property 77.3 78.5 77.9 
Tap off property < 100m 7.6 9.0 8.3 
Tap off property > 100m 4.2 3.5 3.9 
Natural source (river, lake, etc.) 5.9 4.3 5.1 
Well (with and without pump) 4.1 4.5 4.3 
Other 0.8 
N/A 
0.2 
N/A 
0.5 
N/A 
Water purification practices ‡
Purify water by electro-purification, 
chlorination or boiling 
41.8 32.1 36.8 
Consumed straight from source 58.2 
582 
67.9 
614 
63.2 
1169 
Type of sanitation * 
Improved (indoor toilet, sealed 
latrine) 
58.2 50.5 54.2 
Non-improved (open pit, open 
air/none) 
40.8 48.6 44.8 
Other 1.0 
591 
1.0 
622 
1.0 
1213 
* (p<0.05); ‡ (p<0.001) 
23
4.2 Housing and Amenities 
Most of the households surveyed in the final evaluation had three rooms or fewer. AIN-C 
households (71.4 percent) were more likely to use a separate room as a kitchen compared to 
“No-GMP” households (66.1 percent) and more likely to use firewood as cooking fuel (82.6 
percent) compared to the “No-GMP” households (68.2 percent). The principal material used 
for flooring was significantly different in the AIN-C compared to the “No-GMP” households: 
43.5 percent compared to 29.9 percent, respectively, had earth floors and 47.9 percent 
compared to 58.0 percent, respectively, had cement floors. 
Table 4.2 Characteristics of Housing 
AIN-C NO-GMP TOTAL 
% Number of 
Households
% Number of 
Households 
% Number of 
Households
Number of Rooms in the Household 
One room 20.6 20.7 20.7
Two rooms 31.8 29.1 30.4
Three rooms 20.1 20.3 20.2
Four rooms 13.7 15.8 14.8
Five or more rooms 13.7
591 
14.2
622 
13.7
1213 
Number of Rooms used as a Bedroom║
One 65.7 61.7 63.6
Two 21.7 27.3 24.6
Three or more rooms 12.7
591 
10.6
622 
11.6
1213 
Separate Room used for Kitchen*
Yes 71.4 66.1 68.7
No 28.6 591 33.9 622 31.3 1213 
Type of Cooking Fuel‡
Firewood 82.6 68.2 75.2
Liquid gas/kerosene 3.6 8.7 6.2
Propane 9.5 15.0 12.3
Electricity 4.4
591 
8.2
622 
6.3
1213 
Principal Material in Flooring‡
Earth 43.5 29.9 36.5
Wood 0.7 1.0 0.8
Cement 47.9 58.0 53.1
Clay Tile 0.2 1.1 0.7
Ceramic Tile 7.8
591 
10.0
622 
8.9
1213 
║ (p=0.052); * (p< 0.05); ‡ (p<0.001) 
Significant differences were also found between the two groups of households in terms of 
their household possessions. Almost two-thirds (65.6 percent) of the “No-GMP” households 
had electricity, compared to 44.7 percent of AIN-C households. AIN-C households were also 
less likely than “No-GMP” households to own televisions, refrigerators, telephones, and 
motor vehicles: 42.6 and 54.5 percent for televisions, 17.6 and 25.4 percent for refrigerators, 
3.0 and 6.1 percent for telephones, and 4.7 and 7.1 percent for motor vehicles, respectively. 
24
Table 4.3 Household Possession of Amenities 
AIN-C NO-GMP TOTAL 
% Number of 
Household
s 
% Number of 
Household
s 
% Number of 
Household
s 
Utilities and Possessions 
 Electricity ‡ 44.7 65.6 55.4
 Radio 82.1 83.6 82.9
 Television ‡ 42.6 54.5 48.7
 Refrigerator ‡ 17.6 25.4 21.6
 Telephone * 3.0 6.1 4.6
 Motor vehicle ¶
4.7
591 
7.1
622 
5.9
1213 
* (p< 0.05); ‡ (p<0.001); ¶ (p=0.085) 
4.3 Access to Health Services in Terms of Time 
As already noted, access to the health center was found to be significantly different between 
the two “original” groups: the AIN-C and control communities. Therefore it is not surprising to 
see in this analysis that the AIN-C participants are more likely to reside further from the 
health center than the “No-GMP” group (statistically significant p<0.05); 65 percent of AIN-C 
participants compared to 57 percent of “No-GMP” households live half an hour or more away 
for the health center. (See Figure 4.1) 
Figure 4.1 Time to Reach Health Services 
0
10
20
30
40
50
60
70
Percent of households
Less than half hour Half hour or more
AIN-C
No-GMP
25
4.4 Social Assistance Programs 
In addition to the AIN-C program, several other social assistance programs were operating at 
the time of the evaluation survey. AIN-C participants had more contact with all of these 
programs than the “No-GMP” group, as shown in Table 4.4 below. 
Table 4.4 Contact with other Social Assistance Programs 
AIN-C NO-GMP TOTAL 
% Number 
of House￾holds 
% Number of 
House￾holds 
% Number 
of House￾holds 
Social Programs Active in the Community 
AIN-C 88.3 15.3 50.9
Maternal Child “Bono” 30.1 17.4 23.6
Plan International/Honduras 21.2 16.7 18.9
School Lunch 17.4 10.5 13.8
Aldea Global 15.4 4.0 9.6
World Vision 3.9 0.5 2.1
CARE 3.6 0.8 2.1
Other 6.6
591 
4.5
622 
5.4
1213 
4.5 Discussion of Findings 
Overall, these results on the characteristics of AIN-C and “No-GMP” households indicate 
some significant differences in sanitation facilities, water purification practices, housing 
features, and the availability of household amenities. 
The World Health Organization (WHO) and UNICEF define sanitary facilities by the type of 
technology used. Facilities may either be improved or unimproved. An improved sanitation 
facility includes a household connection to a public sewer system, a connection to a septic 
system, a pour-flush latrine, a covered pit latrine and a ventilated pit latrine. Unimproved 
sanitation facilities include any public or shared latrine, open pit latrine, a bucket latrine and 
any open-air location. A significantly greater percentage of households who enrolled in AIN-C 
had access to an improved sanitation facility. Improved sanitation may indicate greater 
wealth, but in this case, the fact that AIN-C households have better access to improved 
sanitation is assumed to reflect that both government and NGOs have targeted these 
“poorer” communities for assistance with sanitation facilities. These improvements should 
reduce the risk of fecal contamination in AIN-C compared to “No-GMP” households. 
Access to clean water also relates to fecal contamination. The majority of households in both 
groups had access to tap water either in their house or on their property. Nevertheless, point￾of-use water purification was more common among AIN-C than “No-GMP” households, most 
likely reflecting the specific AIN-C recommendation for families to chlorinate their water if any 
child has diarrhea. 
Exposure to indoor smoke increases the risk of respiratory infection. Significantly more 
households participating in AIN-C had a separate room in the house for a kitchen than “No￾GMP” households. However, significantly more AIN-C participant households used firewood 
for cooking fuel than “No-GMP” households, which may explain this difference. While only 
17.5 percent of AIN-C households used kerosene, propane, or electricity for cooking, almost 
26
a third (31.9 percent) of non-participating households used these more expensive fuel 
alternatives. Among households that did not have a separate room for the kitchen, a greater 
percentage of AIN-C households relied on firewood for cooking (71.6 percent vs. 55.5 
percent; p< 0.001), indicating that a greater percentage of children in the AIN-C group were 
exposed to indoor smoke. 
Significantly more AIN-C households (two-fifths) had dirt floors than did non-participating 
households, and fewer AIN-C households had electricity or owned expensive durable goods 
(television, refrigerator, telephone, or motor vehicle). Radio ownership was as common 
among AIN-C households as it was among “No-GMP” households. Overall, the differences 
between the two groups for important indicators of wealth suggest that the households 
participating in the AIN-C program were, on average, not as wealthy as the “No-GMP” 
households. 
Household Characteristics: Summary of Key Findings 

 Households with children participating in AIN-C had similar access to safe water, but 
they were more likely to purify their water. 

 A greater percentage of AIN-C households had improved sanitation, perhaps due to 
recognition by NGOs and the Government of Honduras that these are “poorer” 
communities. 

 Fewer AIN-C households had more than one room for the family to sleep in, however the 
total number of rooms in the living structures were not significantly different in the AIN-C 
compared to the “No-GMP” houses. 

 More AIN-C houses had earthen floors and used firewood for cooking; both of these are 
an indication of lower economic status. 

 AIN-C households lived further from health centers. 

 Fewer AIN-C households had electricity or owned a television, refrigerator, telephone, or 
motor vehicle. Radio ownership was the same in both groups. 

 AIN-C households had more contact with other social assistance programs. 
27
5. Caregiver Characteristics
In this section, basic characteristics of the caretakers and mothers are analyzed, including 
age, educational level, employment, number of deaths among children under four, parity, the 
nature of the relationship between the child and the caretaker (e.g., mother, grandmother, 
other relative), and the presence of a husband or male companion in the household. 
5.1 Relationship of Caregivers to Children 
The vast majority of children in these communities were cared for by their mothers (AIN-C: 
95.9 percent; “No-GMP:” 93.6 percent). Most of the other children were cared for by their 
grandmothers (AIN-C: 3 percent; “No-GMP:” 4 percent), a few by other relatives and a very 
few were cared for by a non-relative. 
5.2 Age and Educational Background of the Mother 
Figure 5.1 below shows the age distribution of mothers among AIN-C and “No-GMP” 
participants. The mean age of mothers in both groups is similar; 25.96 percent for AIN-C 
participants and 25.92 for “No-GMP.” 
Figure 5.1 Age Distribution of Mothers 
22.7
3.2
18.5
13.5
8.3 5.9
0.9
17.5
35.2
18.7
12.410.7
5
0.5 0
10
20
30
40
Percent of 
Mothers
AIN-C No-GMP
Age in Years
<20 20-24 25-29 30-34 35-39 40-44 >45
Educational data collected on these mothers show that 13.6 percent and 10.0 percent of 
mothers in AIN-C and “No-GMP” groups, respectively, have not completed any formal 
education. Mothers of children who participated in AIN-C were much less likely to have 
completed primary school compared to mothers of children in the “No-GMP” group. Very few 
in either group went on to complete secondary school; however a significantly greater 
proportion of the “No-GMP” mothers did so. See Figure 5.2 for this educational data in detail. 
28
Figure 5.2 Maternal Educational Levels 
0
10
20
30
40
50
60
Percent of Mothers
AIN-C No-GMP
None
Some Primary
Complete Primary
Some Secondary
Complete Secondary
University
 
5.3 Live Births and Mortality among Children under Four Years of Age
About 24 percent of both AIN-C and “No-GMP” mothers reported having only one live birth. 
This is consistent with the age distribution of the population; almost one-fifth of AIN-C and 
“No-GMP” mothers surveyed were under 20 years of age. An additional 20.8 percent in AIN￾C and 22.3 percent of “No-GMP” mothers reported two live births. The remaining 55.2 
percent in the AIN-C group and 53.6 percent in the “No-GMP” group had three or more live 
births. The two groups are significantly different (p < .05) if one compares the mean number 
of live births; the mean for AIN-C was 3.7 compared, to 3.4 for the “No-GMP” group. 
Figure 5.3 Number of Live Births Reported 
24
20.8
15.7
9.8 9
20.6
24.2
22.3
14.1
15.3
7.4
16.7
0
5
10
15
20
25
Percent of Mothers
AIN-C No-GMP
One Two Three Four Five Six or more
Of the mothers represented by these groups, 16.4 percent of AIN-C participants reported 
having at least one case of a child born alive who subsequently died before age four. For the 
29
“No-GMP” participants, the percentage was slightly lower at 15.1 percent, but this was not 
statistically significant. 
 
5.4 Mother’s Employment and Male Presence in Households 
In response to questions concerning whether a mother works for pay, both groups reported 
that about 18 percent worked for compensation. Also similar was the proportion of mothers 
who reported that they worked outside of the home; about half of AIN-C mothers (54.3 
percent) and “No-GMP” mothers (51.9 percent) were employed outside of their home. For 
both groups, 82 percent of the families had a husband or a companion living at home. 
5.5 Discussion of Findings 
The percentage of mothers with any formal education was not significantly different between 
AIN-C and “No-GMP” households (p=0.057). Among those mothers who had some 
schooling, the “No-GMP group had more: a significantly smaller percentage of AIN-C 
mothers had completed primary or had any secondary education compared to the “No-GMP” 
group. Since children’s nutritional status generally is positively associated with maternal 
education, this lower level of formal school for AIN-C mothers called for attention to health 
and nutrition practices. 
The slightly higher parity among AIN-C mothers (3.70) compared to “No-GMP” mothers 
(3.40) is not likely to represent a significantly higher child care burden. No other 
characteristics—having a husband/companion living at home, maternal employment and 
place of employment, maternal age, principal caregiver, number children under four years old 
who had died—were significantly different in the AIN-C and “No-GMP” groups. 
Summary of Key Findings: Caregiver Characteristics 

 Mothers of children enrolled in AIN-C had completed fewer years of formal education.

 Mothers of AIN-C participants had similar reproductive health histories, although parity 
was significantly higher among AIN-C mothers.
30
6. Participation in AIN-C Program Activities
This section provides background information on AIN-C program participants. It includes a 
description of some of the important characteristics of “participation,” age of entry into the 
program, and consistency of participation, as well as the activities that take place during the 
GMP sessions, accuracy of plotting weights on growth cards by monitoras and activities in 
follow-up home visits. 
6.1 Knowledge of AIN-C 
Respondents were asked about their knowledge and participation in growth promotion 
programs. Nearly all respondents (99.8 percent) from AIN-C-enrolled households knew a 
person in their community who was responsible for weighing children every month. Ninety 
percent of these respondents cited “monitora” as the person who takes these weights. 
6.2 Age of Entry into Program 
During the period under study, AIN-C norms required newborn registration in AIN-C no later 
than three months after delivery. Registration of infants during the first month was an 
objective of the program. Based on age at first attendance recorded on growth promotion 
cards presented by respondents, only about 25 percent of children who participated in AIN-C 
were enrolled within one month of birth. However, approximately 75 percent were enrolled 
within three months and half of the participating infants/children were enrolled within 2 
months of delivery. 
Table 6.1 Age of Entry into AIN-C 
Age in Months Percent 
< 1 24.2 
1 33.3 
2 17.9 
3 6.3 
4 4.9 
5 3.4 
6-8 5.6 
9–11 1.7 
12–23 2.7 
Mean Age of Entry 1 month 
31
6.3 Level of Participation 
The norm for the AIN-C program is for children to attend monthly until they reach two years 
of age. The program objective is for a child to participate 80 percent of the time, i.e. attend 
monthly growth promotion sessions 10 out of 12 times a year. According to the growth cards 
presented by respondents, 67 percent of AIN-C participants attended all three monthly 
sessions during the three months prior to the survey (Table 6.2). Eighty-seven percent had 
attended two or more times. 
Figure 6.1 Growth Promotion Attendance in Last Three Months; Children > 3 Months 
6.4 Activities during Growth Promotion Sessions 
Information on activities taking place at the previous growth promotion session was obtained 
both from caregivers’ spontaneous reports and with prompting. The main activity at the 
growth promotion session is weighing the children. Ninety-nine percent of caregivers 
reported that their child had been weighed at the previous session (only 83.9 percent did so 
without prompting [Figure 6.1]). A slightly lower but high percentage reported being informed 
of their child’s weight (93.6 percent), and of the adequacy of the child’s weight gain (85.4 
percent). More then 75 percent reported that the monitora explained child feeding issues to 
them, and nearly 70 percent (67.7 percent) reported discussing breastfeeding. The 
percentage of caregivers reporting breastfeeding discussions was scarcely different among 
infants younger than 6 months and infants/children 6–23 months (69.0 percent and 67.5 
percent, respectively; data not shown). 
2.7
10.3
19.7
67
0.4 0
10
20
30
40
50
60
70
None One Two Three 
or more
Doesn't
know/remember
Number of Months Attended
Percent of AIN-C 
Participants
32
Figure 6.2 Activities at Previous Growth Promotion Session as Reported by Caretakers 
 
6.5 Accuracy of Growth Card Plotting by Monitoras
The accuracy of plotting child weights from month to month on the growth chart (the 
detection of growth faltering) by monitoras was assessed for children participating in AIN-C 
who had at least two weights marked on their growth cards. Ninety-three percent of children 
had at least two weights recorded on their cards with no greater than a two month interval 
between two weights. The interpretation of the growth trend for these children was 
determined to be accurate in 87.9 percent of these cases. 
Table 6.2 Accuracy of Plotting on Growth Cards All Children Combined 
Category of Child Percent Total Number 
of Children 
Two or more weights marked on card 93.3 511 
Two or more weights marked on card within 3 month period 92.0 477 
Two or more weights marked on card within 3 months period with 
accurate interpretation of growth trend 
87.9 477 
14.2
37.9
25.2
30.5
83.9
53.5
39.9
60.2
63.1
15.2
0 10 20 30 40 50 60 70 80 90 100
Discussed breastfeeding
Explained child feeding
Informed adequacy of
weight gain
Informed child's weight
Weighed child
Percent
Mentioned without prompting Mentioned when prompted
33
6.6 Home Visits by Personnel from GMP Programs 
Home visits are made by monitoras for a variety of reasons, including to enroll newborns and 
to follow-up on children who have missed growth promotion sessions or who have failed to 
gain adequate weight. Almost 30 percent of AIN-C participants reported having received at 
least one home visit from a monitora (Figure 6.3). Among those families receiving visits, 35.6 
percent had received one visit and another 32.2 percent reported receiving multiple visits. 
Figure 6.3 Frequency of Home Visits as Reported by Caregivers 
35.6%
32.2%
26.6%
5.6%
Once
Sometimes
Almost every month
Several times a month
Among the caregivers who received home visits, the main reasons cited for the home visit 
were that the child was sick, had missed a weighing session, or had been found to have 
inadequate growth at the growth promotion session (Figure 6.4). 
 
Figure 6.4: Reasons* for Last Home Visit by an AIN-C Monitora 
2.5
3.2
13.6
16.5
25.6
0 5 10 15 20 25 30
Other: w eigh child
Other: newborn visit
Inadequate growth
Missed w eighing
session
Child was sick
Percent
 *Not mutually exclusive 
34
The most common activity reported among caregivers (almost 72 percent) as occurring 
during a home visit was counseling/advice provided to the caregiver from the monitora. Other 
activities included weighing the child (32.2 percent), reviewing the growth card (10.7 
percent), and referring to the health center (8.5 percent).
Figure 6.5 shows caregivers’ recall of the counseling topics or advice discussed during home 
visits by the age of the child. Among caregivers of infants younger than six months, 
counseling for increased amount or frequency of breastfeeding was the most common topic 
for counseling. However, more than 10 percent of these caregivers recalled receiving advice 
concerning feeding of solid or semi-solid food. 
Figure 6.5 Specific Advice Given* by Age of Child during Last Home Visit 
*Not mutually exclusive 
Only 45 caregivers cited infant/child illness as the reason for their visit from the monitora, and 
only 5 of these caregivers had infants younger than 6 months. When the reason for the last 
home visit was child illness, the percentage of caregivers receiving advice was approximately 
the same as for all visits, but referrals to the health center were nearly three times as 
common. None of the caregivers of children younger than six months who received a sick 
child visit from the monitora recalled advice concerning feeding of solid or semi-solid food. 
Only one of the five recalled receiving advice related to increased frequency or quantity of 
breastfeeding16 (Figure 6.7). Very few of the mothers of sick children aged 6–11 months or 
12–23 months received advice to increase frequency and quantity of breastfeeding, but 53 
 
16 Response to any of Figure 6.6 categories for “breastfeed more” or “breastfeed more often”. 
2.9
5.7
11.4
31.4
31.4
5.7
2.9
6.1
18.4
12.2
8.2
12.2
12.2
2.0
15.2
22.8
19.6
10.9
12.0
22.8
5.4
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Give vitamins
Feed more often 
in smaller
quantities
Feed softer/thicker 
food
Breastfeed more 
often
Breastfeed more
Give more food
Feed more often
Percent
12 - 23 months (n=92)
 6 - 11 months (n=49) 
 < 6 months (n=35)
35
percent of mothers with 6 to 11-month-olds and 36 percent of mothers with 12 to 23-month￾olds received some advice concerning increased feeding of solid or semi-solid food.17
Figure 6.6 Advice Given* for Sick Child, by Age of Child, during Last Home Visit 
0.0
0.0
0.0
20.0
20.0
0.0
0.0
0.0
33.3
26.7
13.3
20.0
20.0
0.0
20.0
24.0
24.0
4.0
4.0
8.0
4.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Give vitamins
Feed more often in smaller
quantities
Feed softer/thicker food
Breastfeed more often
Breastfeed more
Give more food
Feed more often
Percent
12 - 23 months (n=25)
6 - 11 months (n=15)
< 6 months (n=5)
*Not mutually exclusive 
6.7 Discussion of Findings 
The GMP program was well established in the AIN-C communities. Almost 100 percent of 
participants in the AIN-C program knew the activities supported by the program and the 
person in the community responsible; 90 percent knew this person as the monitora. 
Community-based growth promotion is based on the understanding that children under the 
age of two grow rapidly, and those who don’t grow adequately are at increased risk for 
malnutrition, which is associated with increased morbidity and mortality. To effectively 
prevent malnutrition in a community, all children under the age of two should have the 
opportunity to participate on a monthly basis. Indicators related to entry into and participation 
in the program suggest that the quality of AIN-C implementation was high. Although not quite 
achieving the program goal of registration of all children in the first month of life, more than 
half were registered within two months and 75 percent within three months. Monthly 
participation of all children under two is also critical to effective program functioning. 
Attendance among AIN-C participants was consistent: almost 90 percent of AIN-C children 
had been weighed two or more times in the three months prior to the survey. This weighing 
occurred through monthly sessions as well as home visits. Almost 90 percent of growth cards 
reviewed were found to have an accurate interpretation of the growth trend. 
 
17 Response to any of Figure 6.6 categories for “Feed more often”, “Give more food”, or “Feed more 
often in smaller quantities”. 
36
In addition to the weighing activity, which included informing caregivers of the child’s weight 
and the adequacy of weight gain, counseling on breastfeeding and feeding practices were 
major activities cited by a large majority (75 percent) of AIN-C caregivers during GMP 
sessions. This is less than the 100 percent expected to have received counseling, since AIN￾C is designed to provide support to caregivers regardless of the growth trend of the child. 
The content of the counseling varied according to the age of the child and whether the child 
was healthy or sick. In most cases the focus of counseling for caregivers with infants less 
than six months of age was on breastfeeding, although ten percent of caregivers recalled 
messages related to food as well as breastfeeding, which is not consistent with the 
recommendation for exclusive breastfeeding for the first six months. 
Home visits are another major element of the AIN-C program; about 30 percent of caregivers 
reported receiving a home visit. Many of these visits were to follow-up on sick children, so 
the content of the advice included referral to the health center three times more often than 
during the monthly growth promotion session. Based on this small sample of mothers (45) 
whose children were sick at the time of the survey, monitoras may need additional support to 
provide appropriate advice on nutritional care of the sick child. Very few caregivers recalled 
receiving messages related to continued and increased frequency of feeding during illness. 
Summary of Key Findings: Participation in AIN-C 

 Three-quarters of children who participated in AIN-C were enrolled by the end of the 
second month of life. 

 Attendance at AIN-C growth promotion sessions was consistent and monitoras followed 
up through home visits with caregivers who miss a session. 

 Most caregivers indicated child weighing as the primary activity taking place at growth 
promotion sessions, with more than 90 percent reporting that monitoras informed them of 
their child’s weight and 85 percent reporting that monitoras informed them about the 
adequacy of their child’s weight gain. 

 The capacity of the monitoras to interpret growth trends was high; 87.9 percent of cards 
checked were determined to be accurate. 

 In addition to child weighing, mothers report that monitoras routinely discussed child 
feeding and breastfeeding with caregivers at growth promotion sessions. 

 Counseling during sick child visits appears to be inadequate (based on a very small 
sample of caregivers who had sick children at the time of the survey); very few recalled 
advice on continued feeding during illness. 
37
7. Child Growth and Development Knowledge and 
Actions for Improvement
This section presents contrasting information related to caregiver perceptions and knowledge 
of child growth among AIN-C and “No-GMP” participants, as well as their knowledge of the 
growth promotion program, including advice that they have been given in counseling. 
7.1 Caregiver Perceptions of Child Growth 
Caregiver’s perception of their child’s growth differed by participation status (Table 7.1). The 
percentage of caregivers who felt their child was growing well was essentially identical in 
both the AIN-C and the “No-GMP” groups. However, among caregivers who said that their 
child was not growing well, a higher percentage of those in the “No-GMP” group said they did 
not know their child’s growth status, while a higher percentage of those in the AIN-C group 
said that they knew that their child was not growing well. 
Table 7.1 Caregivers’ Perceptions of their own Child’s Growth 
Perception of Child’s Growth* AIN-C NO-GMP TOTAL 
% Number 
of 
Children 
% Number 
of 
Children 
% Number 
of 
Children 
Growing Well 81.2 82.3 81.8
Not Growing Well 15.2 11.6 13.4
Do not Know 3.6
591 
6.1
622 
4.9
1213 
* χ
2
 = 7.143; p=0.028 
7.2 Caregiver Knowledge of Signs of Good Growth 
Caregivers were asked what they feel indicates that a child is growing well (Figure 7.1). 
About half of AIN-C participants responded “gains weight,” compared to 29 percent of the 
“No-GMP” groups. This difference is statistically significant (χ
2
 = 42.523; p<0.001). About half 
of “No-GMP” participant caregivers mentioned that a child who is “healthy/not sick” is a child 
who is growing well, compared to only 43.8 percent of AIN-C respondents; this too was a 
statistically significant result (χ
2
 = 5.922; p<0.05). No other differences in the responses to 
this question by the two groups were statistically significant, however more than 30 percent 
of AIN-C participants said that growing normally and eating well indicated good growth. 
38
Figure 7.1: Signs of Good Growth Named by Caregivers* 
*More than one response possible 
Although “gains weight” is the primary sign of good growth promoted by the AIN-C project, 
the other responses given by caregivers also can be understood, in some sense, as accurate 
signs of good growth. 
An examination of caregivers who gave only a single response (rather than several 
responses as above) may provide insight into the extent to which AIN-C participants consider 
gaining weight to be the primary sign of good growth. Among the respondents who cited only 
one sign of good growth (29.4 percent of 581 AIN-C participants and 35.9 percent of 591 
non-participants provided a single response; χ
2
 = 5.522; p<0.05), a highly significantly 
greater percentage of AIN-C participants cited “gains weight” as the sign of good growth 
compared to “No-GMP” participants; 38.6 compared to 11.8 percent, respectively. However, 
less than half of AIN-C participants who named only one sign of good growth cited “gains 
weight.” In addition, a highly significantly smaller percentage of AIN-C participants (28.7 
percent) cited “healthy, not sick” compared to “No-GMP” participants (48.1 percent) as the 
sign of good growth. The percentage of AIN-C participants who responded that a child who is 
growing normally is a sign of good growth (18.1 percent) also was significantly smaller than 
the percentage of the “No-GMP” group who gave this response (26.9 percent). 
50.8
34.4
39.5
32.5
19.1
43.8
39.1
36.4
50.9
21.5
0 10 20 30 40 50 60
Healthy/not sick
Eats well
Is growing normally
Gains weight
Sleeps well
Percent
AIN-C
No GMP
39
Figure 7.2 Signs of Good Growth Named by Caregivers who gave a Single Response, 
by Program Participation 
1.9
11.8
26.9
11.3
48.1
4.1
38.6
18.1
10.5
28.7
0 10 20 30 40 50 60
Sleeps Well
Gains Weight
Is Growing
Normally
Eats Well
Healthy/not sick
Percent of Participants
AIN-C
No-GMP
7.3 Caregiver Knowledge of Signs of Poor Growth 
Caregivers also were asked to name signs of poor child growth (Figure 7.3). In both groups 
the most common responses were “thin, underweight”, “undernourished,” and “does not eat.” 
Fewer than 20 percent of AIN-C participants named “not gaining weight adequately” as a 
sign of poor growth. This percentage, although low, was significantly higher than the 12.7 
percent of “No-GMP” participants who cited “not gaining weight” as a sign of poor growth (χ
2
= 4.619; p<0.05). The proportion of AIN-C participants who responded that a child who was 
“thin, underweight” or lethargic” was a sign of poor growth responses was significantly higher 
than among non-participants (χ
2
 = 6.387, p<0.05 and χ
2
 = 6.472, p<0.05, respectively). The 
groups did not differ in the proportion of caregivers who responded that poor growth was 
signaled by a child who is “undernourished,” “sick,” and “does not eat.” (See Figure 7.3.)
40
Figure 7.3: Signs of Poor Growth Named by Caregivers, by Program Participation 
Among caregivers who gave only one response for signs of poor growth, “thin, underweight” 
and “sick” were the most common responses in both groups. No differences were seen 
between groups for any response category, and fewer than 10 percent of AIN-C participants 
who named only a single sign of poor growth named “not gaining weight adequately” as that 
sign. 
7.4 Counseling for Improved Child Growth 
AIN-C program norms call for all mothers to receive counseling, facilitated by the program’s 
special counseling cards. Approximately 75 percent of AIN-C participant caregivers recalled 
having seen the AIN-C counseling cards. Most recalled seeing these during a contact with an 
AIN-C monitora, but some also reported seeing them at the public health facility (Table 5.2). 
Twenty-five percent of the AIN-C participants interviewed reported that they had not seen the 
counseling cards; this despite the fact that the program expectation is for all mothers to be 
counseled at every weighing session. 
Forty-five percent of the “No-GMP” group also reported having seen the cards; four percent 
of them reported that they had seen the counseling cards from an AIN-C monitora despite 
the fact that their children were not enrolled in the program. 
44.7
43.7
21.2
40.7
19.5
12.7
46.9
44.2
20.6
47.9
25.5
17.1
0 10 20 30 40 50 60
Does not eat
Sick
"Undernourished"
Thin, underweight
Lethargic
Is not gaining weight adequately
Percent
AIN-C
No GMP
41
Table 7.2 Recognition of AIN-C Counseling Cards 
 
AIN-C No GMP 
% of 
Total 
Number of 
Households 
% of 
Total 
Number of 
Households 
Have seen counseling cards 
 Yes 73.3 44.9 
 No 26.7 
591 
55.1 
622 
Location where cards were seen 
CESAR / CESAMO 35.8 91.0 
 With AIN-C Monitora 76.7 4.3 
 Elsewhere 5.3 
510 
6.1 
283 
Interviewers noted whether the child had an episode of growth faltering (inadequate monthly 
weight gain) indicated on their growth card and, if so, asked the respondents whether they 
received advice related to improving the child’s weight gain and the specific content of that 
counseling. Eighty-one percent of AIN-C participants reported receiving advice. The most 
frequent advice received was “give more food than accustomed.” Sixty-four percent of 
caregivers of infants and children six months or older and ten percent of caregivers with 
children under six months reported receiving this advice. Fifty percent of caregivers of infants 
younger than six months with faltering growth on the card reported “continue breastfeeding” 
as the advice they received, although the sample was only ten caregivers. “Continue 
breastfeeding” also was the second most common advice reported for caregivers whose 
children with faltering growth were six months or older. 
Figure 7.4 Advice Provided for Children with Inadequate Gain Weight: AIN-C 
0.0
0.0
0.0
10.0
0.0
0.0
10.0
0.0
50.0
10.0
1.3
2.6
6.6
7.5
12.8
13.7
15.0
18.1
22.5
66.1
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
Give food directly in child's
mouth
Do not withhold beans
Pay attention to amount eaten
Referred to health care
Feed more frequently 
Give mashed foods
Continue breastfeeding, giving
other foods
Give thick soups
Continue breastfeeding
Give more food than
accustomed
Percent
6+ months (n=227)
< 6 months (n=10)
42
7.5 Discussion of Findings 
The AIN-C program is built around the concept that adequate weight gain is a good proxy for 
health, and that by regular monitoring of weight gain in children (0–24 months of age) and 
tailored counseling based on that growth information, malnutrition can be prevented. The 
results of this evaluation show that the caregivers who participated in the AIN-C program 
were significantly more likely than the “No-GMP” group to understand the signals (especially 
weight gain) of healthy growth, and a small proportion, although still significantly more than 
the “No-GMP” group, understood that inadequate weight gain was a sign of poor growth. 
About half of AIN-C participants compared to a third of the “No-GMP” caregivers responded 
that gaining weight was a sign of good growth. For a sign of poor growth, 17.1 percent of 
AIN-C participants compared to 12.7 percent of “No-GMP” caregivers named inadequate 
weight gain as a sign of poor growth; even among AIN-C participants, very few articulated 
the relationship between poor growth and inadequate weight gain. 
One of the key messages of AIN-C was “El Niño o Nina que crece adecuadamente es sano; 
el Niño o Nina que no crece adecuadamente esta enfermo” [A child who grows adequately is 
healthy, while a child who grows inadequately is sick]. The significant differences in the two 
groups of caregivers understanding of the relationship between weight gain and growth 
suggest that this message was communicated to some extent through the program. 
However, the fact that only 50 percent of AIN-C participants could articulate this relationship 
is a signal that the project failed to effectively convey this concept. The reason for this result 
is unclear since the survey didn’t provide any data to identify whether this was a problem with 
the training of the monitoras or a communication issue between the monitoras and the 
caregivers. 
One of the most important aspects of the AIN-C program was the counseling provided to the 
caregiver by the monitora, based on the growth information. A set of 20 laminated counseling 
cards was a key job aid for the monitoras developed through the AIN program and intended 
to facilitate the counseling process. These cards covered topics including establishing 
breastfeeding for children 0–7 days old, managing adequate and inadequate growth with or 
without breastfeeding for children 0–2 and 3–5 months, and managing adequate or 
inadequate growth with proper feeding for children 6–8, 9–11, 12–17, or 18–23 months (see 
BASICS.org website for sample cards). At the time of this evaluation, 75 percent of the AIN￾C participants had seen the counseling cards; this despite the expectation that all caregivers 
would be counseled regardless of the growth status of their children. Notable is that almost 
half (45 percent) of the “No-GMP” caregivers had also seen the counseling cards. This result 
bolsters the earlier conclusion that the AIN-C program had “spilled over” into the “control” 
areas where the majority of the “No-GMP” group resided (see Chapter two). This suggests 
that the “No-GMP” group is likely to have some familiarity with the feeding and care 
messages contained in the counseling cards. 
Knowledge of the growth status of children under two years of age on a regular basis is the 
foundation for the counseling messages provided through the AIN-C program. Furthermore, 
when a child is encountered with inadequate growth, the monitora is expected to provide 
specific, tailored advice to the caregiver of that child to help improve weight gain. The results 
of this evaluation show that more than 80 percent of AIN-C participants whose children were 
faltering in growth received counseling. The specific advice received as reported by these 
caregivers shows mixed results in terms of the appropriateness and specificity of the 
messages provided. Ten percent of caregivers of children younger than six months recalled 
being counseled to give their child more food than accustomed. Very few respondents 
reported messages other than “continue breastfeeding” and “give more food than 
43
accustomed;” more specific messages related to the quality of foods for children over six 
months of age such as give thick soups, give mashed foods, and pay attention to the amount 
eaten were mentioned only 18.1, 13.7, and 6.6 percent of caregivers, respectively. 
Summary of Key Findings: Child Growth Concepts and Actions for Improvement 

 The concept of weight gain as a signal of “good” growth was understood by more than 50 
percent of AIN-C participants; this result was significantly different from “No-GMP” 
participants who only cited weight gain one-third of the time. 

 Half of the AIN-C participants did not mention adequate weight gain as a sign of health, 
despite the fact that this was a key message of the AIN-C program. 

 The concept of the “inadequate weight gain” was not well understood; only 17 percent of 
AIN-C participants identified this as a sign of “poor” growth. 

 The “special” counseling cards used in the AIN-C program were widely, although not 
universally, recognized by AIN-C participants. 

 While nearly 75 percent of AIN-C participants had seen the cards, 45 percent of the “No￾GMP” group also recognized the cards. 

 Among the “No-GMP” participants who recognized the cards, 91 percent had seen them 
in the health center. 

 Among caregivers who had children with faltering growth, more than 80 percent received 
counseling. 

 “Give the child more food than accustomed” was the most common message recalled by 
caregivers of children with inadequate growth over six months of age. 

 Among caregivers of children younger than six months with inadequate growth, 
“continue breastfeeding” was the most common message recalled (50 percent), while 10 
percent recalled “give the child more food than accustomed.” 
44
8. Infant and Young Child Feeding Practices
Mothers were asked a series of questions about breastfeeding and the introduction of other 
liquids, complementary foods, and the frequency of daily feeding. This information was 
analyzed for differences in optimal feeding practices between the AIN-C and the “No-GMP” 
participants according to the age of the child. 
8.1 Exclusive and Any Breastfeeding Practices 
Girls and women who were breastfeeding during the time of the survey were asked the 
number of times that they gave their child any other liquids or food in the 24 hours prior to the 
survey interview. Infants whose mothers reported that they received no food or other liquids 
were categorized as exclusively breastfed. The percentage of infants younger than three 
months who were exclusively breastfeeding was 68.3 percent in the AIN-C group compared 
to 56.7 percent in the “No-GMP” group, but this difference was not statistically significant. For 
infants under six months of age, the difference in proportion exclusively breastfed between 
the AIN-C and the “No-GMP” groups was statistically significant; 55.8 percent compared to 
40.0 percent respectively (see Figure 8.1). 
Figure 8.1 Infants < 3 and < 6 Months Exclusively Breastfed by Participation 
0
10
20
30
40
50
60
70
80
Infants < 3months Infants < 6 months
Percent
AIN-C
No-GMP
 
The median duration of exclusive breastfeeding and of any breastfeeding were calculated 
following the method used for the Demographic and Health Surveys (DHS). Three-group 
moving averages were calculated for numerators and denominators, with ages grouped into 
months. Starting from the youngest age, the value of the median was determined by linear 
interpolation of the percentage of breastfeeding children (or exclusively breastfeeding) in the 
first age group where the percentage was less than 50 percent and the percentage of 
breastfeeding children in the next youngest age group. The median durations of exclusive 
breastfeeding and of any breastfeeding were both 1.5 months longer among AIN-C 
participants than in the “No-GMP” group, although the duration of any breastfeeding was 
high in both groups. 
45
Table 8.1 Median Duration of Exclusive and Any Breastfeeding by Participation 
AIN-C No GMP 
Age in 
Months 
Number of 
Children 
Age in 
Months 
Number of 
Children 
Median duration (months) of exclusive 
breastfeeding* 
4 120 2.4 135 
Median duration (months) of any 
breastfeeding* 
20.9 535 19.5 543 
* Source for methodology: Rutstein, SO and G Rojas. “Guide to DHS Statistics”. Calverton, MD: ORC 
MACRO, 2003. 
Although the median duration of breastfeeding (and of exclusive breastfeeding) is one of the 
most commonly used breastfeeding indicators, it has specific limitations. First, for ease of 
calculation, the percentage breastfeeding (or exclusive breastfeeding) is assumed to 
decrease monotonically from birth. This assumption is unlikely to be true, especially for 
exclusive breastfeeding: the rate at which infants in the population cease exclusive 
breastfeeding is slower during the first few months of life than it is as six months approaches. 
Furthermore, if breastfeeding (or exclusive breastfeeding) does not decline monotonically, 
the shape of the “decay curve” may be different among different population groups, 
confusing direct comparisons of “median duration.” 
For this reason, breastfeeding outcomes in the AIN-C and “No-GMP” groups also were 
compared using Kaplan-Meier survival analysis (Table 8.2), considering cessation of 
breastfeeding (and exclusive breastfeeding) as “events,” and continued exclusive 
breastfeeding and breastfeeding, respectively, as “survival.” This procedure estimates 
survival time for the group by taking into account both the rate at which “death” occurs 
(cessation of exclusive breastfeeding or cessation of breastfeeding) and the age at which 
surviving (still exclusively breastfeeding or still breastfeeding) individuals in the group are lost 
to follow-up (their age at the survey). Survival curves (Figure 8.2 and 8.3) allow visual 
comparison of the rates at which breastfeeding decreases in the groups, and the variability of 
these rates over time. Procedures associated with survival analysis allow statistical 
comparisons of these rates to be made. 
46
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Age (months)
0.0
0.2
0.4
0.6
0.8
1.0
Cum Survival
Participation Status
Community￾based Growth 
Promotion
No Growth 
Promotion
Community￾based Growth 
Promotion￾censored
No Growth 
Promotion￾censored
Survival Functions
0.0 5.0 10.0 15.0 20.0 25.0
Age (months)
0.0
0.2
0.4
0.6
0.8
1.0
Cum Survival
Participation Status
Community￾based Growth 
Promotion
No Growth 
Promotion
Community￾based Growth 
Promotion￾censored
No Growth 
Promotion￾censored
Survival Functions
Table 8.2 Estimated Months Breastfeeding (Kaplan-Meier Survival Analysis) 
Exclusive breastfeeding † Any breastfeeding††
AIN-C No-GMP AIN-C No-GMP 
Age in 
Months 
Number 
of 
Children
Age in 
Months 
Number 
of 
Children 
Age in 
Months 
Number 
of 
Children 
Age in 
Months 
Number 
of 
Children 
75% 
continue 
4.1 2.8 18.8 18.0 
50% cease 5.2 4.4 21.6 21.4 
25% 
continue 
5.8 5.2 22.9 22.9 
Mean 4.8 
120 
4.0 
135 
20.3 
535 
19.4 
543 
Note: Maximum age of child is 23 months 
†
 Log-rank statistic = 15.94 (p=0.0001); Breslow’s Test = 12.90 (p=0.0001); analysis limited to infants < 
6 months; 
†† Log-rank statistic = 4.22 (p=0.0399); Breslow’s Test = 9.43 (p=0.0021) 
Figure 8.2 Survival Function for Time to Figure 8.3 Survival Function for Time to 
Cessation of Exclusive Breastfeeding Cessation of any Breastfeeding 
As can be seen in Figure 8.2, the rate at which infants in the “No-GMP” group cease 
exclusive breastfeeding is relatively constant up to 6 months. Among AIN-C participants, the 
rate of cessation is slower in the early months and more rapid as 6 months nears. Thus, 
before 3 months of age, 25 percent of “No-GMP” infants cease exclusive breastfeeding (75 
percent continue), but at 4 months of age less than 25 percent of the AIN-C group infants 
cease exclusive breastfeeding (Table 8.2). This difference between the groups in the rate of 
“survival”, or continued exclusive breastfeeding, is highly significant. 
47
Similarly, even though the median age of breastfeeding is approximately two years in both 
groups, the rate at which breastfeeding ceases is slower in the AIN-C group. Figure 8.3 
illustrates that, at any given age, a larger proportion of infants in the AIN-C group continue to 
breastfeed, and the difference between the groups is greatest at early ages (the much higher 
significance of the Breslow’s test, which gives more weight to early events, in comparison 
with the log-rank test, emphasizes this difference). 
8.2 Infant and Young-Child Feeding 6–23 Months 
Adequacy of feeding was determined by age group, as food requirements change rapidly 
during the first two years of life. For each age group, appropriate feeding was defined as 
continued breastfeeding, plus the frequency and quantity of feeding recommended for that 
age group in the AIN-C counseling materials as follows: 
• Infants 6–8 months of age: 2 or more meals daily 
• Infants and children 9–11 months of age: 3 or more daily meals, with 4 or more 
tablespoons of food at each meal 
• Children 12–23 months of age: 5 or more daily meals, with 5 or more tablespoons of food 
at each meal 
8.2.1 Breastfeeding and Frequency of Feeding in Children Under 12 Months 
AIN-C and “No-GMP” groups differ significantly in their breastfeeding and frequency of 
feeding practices among children 6–11 months of age. For children 6–8 months old, 80 
percent of AIN-C children were appropriately fed, compared to 60.3 percent of “No-GMP” 
children. The differences were apparent and significant in both the proportion of children who 
were breastfed and the frequency of feeding solid/semi-solid foods (see Figure 8.4). 
Figure 8.4 Feeding According to Recommendations among Children 6–8 Months 
0
20
40
60
80
100
120
Breastfeeding Frequency of Feeding
per Recommendation
Breastfeeding and
Frequency Combined
Percent
AIN-C
No-GMP
Likewise for children 9–11 months of age, the AIN-C group compared to the “No-GMP” group 
had higher rates of breastfeeding (85.5 percent compared to 81 percent), better frequency of 
feeding rates (94 percent compared to 84 percent), and overall, a higher proportion of 
children who met the recommendation for both breastfeeding and frequency of feeding for 
this age group (81 percent compared to 68.3 percent). These differences were also 
statistically significant (see Figure 8.5). 
48
Figure 8.5 Feeding According to Recommendations among Children 9–11 Months 
0
10
20
30
40
50
60
70
80
90
100
Breastfeeding Frequency of
Feeding per
Recommendation
Breastfeeding and
Frequency
Combined
Appropriately Fed
BF, Frequency and
Amount
Percent
AIN-C
No-GMP
An important element of appropriate feeding is the amount of food. Caregivers who 
described their food intake as normal in the day previous to the survey were asked questions 
regarding the amount of food the child received. Among all children, including AIN-C 
participants, the proportion of children who received the appropriate amount of food was 
extremely low, however the proportion of children 9–11 months who received the 
recommended amount of food was almost double in the AIN-C group; 15.6 percent (AIN-C) 
compared to 7.9 percent (“No-GMP”). 
Table 8.3 Amount of Food Received among Children 9–11 Months 
AIN-C NO-GMP TOTAL 
Amount of Feeding* # of 
Tbs 
Number 
of 
Children 
# of 
Tbs 
Number 
of 
Children 
# of 
Tbs 
Number 
of 
Children 
Mean Tablespoons food per meal 2.8 2.7 2.7
Median Tablespoons food per meal 2.0 64 2.0 63 2.0 127 
4+ Tablespoons food per meal (%) 15.6 7.9 11.8
* Children whose caregiver described the previous day’s food intake as normal 
8.2.2 Breastfeeding and Frequency of Feeding among Children 12–23 Months 
AIN-C and “No-GMP” groups also differed significantly in feeding practices among children 
12–23 months of age; AIN-C participants had significantly better practices related to 
breastfeeding and the frequency of feeding. Although the difference in the frequency of 
feeding was highly significant (p <0.001), the proportion of children who received the 
recommended number of feedings was very low for both groups: 16.1 percent for AIN-C 
participants and 5.8 percent for the “No-GMP” group (see Figure 8.6). 
49
Figure 8.6 Appropriate Feeding among Children 12–23 Months 
0
10
20
30
40
50
60
70
Breastfeeding Frequency of Feeding
per Recommendation
Breastfeeding and
Frequency Combined
Appropriately Fed BF,
Frequency and
Amount
AIN-C
No-GMP
As was the case for younger children, the proportion of children 12–23 months in both 
groups AIN-C participants (5.8 percent) and the “No-GMP” group (6.3 percent) who provided 
the recommended amount of food was extremely low; there was no statistical difference 
between the groups in this indicator (Table 8.4). 
Table 8.4 Amount of Food Received among Children 12–23 Months 
AIN-C NO-GMP TOTAL 
Amount of Feeding* # of 
Tbs 
Number 
of 
Children 
# of 
Tbs 
Number 
of 
Children 
# of 
Tbs 
Number 
of 
Children 
Mean Tablespoons food per meal 2.6 2.7 2.7
Median Tablespoons food per meal 2.0 208 2.0 224 2.0 432 
5+ Tablespoons food per meal (%) 5.8 6.3 6.0
8.2.3 Appropriate Feeding of Children 6–23 Months 
Comparing the results of the AIN-C and the “No-GMP” groups across the period of 6–23 and 
9–23 months, the AIN-C participants have a greater proportion of children who met the 
recommendations for breastfeeding and the frequency of feeding. For children 6–23 months, 
38 percent of AIN-C children met both recommendations (breastfed and frequency of 
feeding) compared to 24.9 percent of the “No-GMP” group. This difference is highly 
statistically significant (p< 0.001). Likewise for children 9–23 months, more AIN-C children 
were fed according to the recommendations (breastfeeding and the frequency of feeding) 
compared to “No-GMP” children: 28.4 percent and 17.6 percent, respectively. This difference 
is statistically significant (p<0.01). When the amount of food per feeding is included, which 
reflects a complete assessment of appropriate feeding practices, the proportion of children 
who are fed appropriately drops to very low levels (2.8 percent for AIN-C participants and 0.8 
percent of “No-GMP”) however, this difference remains statistically significant (p<0.05). 
50
Figure 8.7 Appropriate Feeding of Children 6–23 and 9–23 months 
0
5
10
15
20
25
30
35
40
Met Recommendations BF,
Frequency, Children 6–23
Months
Met Recommendations BF,
Frequency, Children 9–23
Months
Appropriately Fed BF,
Frequency, Amount of Food,
Children 9–23 Months
Percent
AIN-C
No-GMP
8.3 Feeding Practices during Childhood Illnesses 
Respondents who reported that their children were sick in the two weeks prior to the survey 
with either diarrhea or acute respiratory infection were asked about the quantity and 
frequency of breastfeeding and complementary feeding during the child’s illness. Analysis of 
these feeding practices was restricted to those infants and children aged six months or older 
who had been introduced to solid foods by the time of the illness. Results are presented 
separately for the illnesses of diarrhea and acute respiratory infection. 
51
Percent 
8.3.1 Feeding during Diarrheal Illness 
A greater percentage of AIN-C participants increased both the quantity and frequency of 
feeding during diarrheal illness (Table 8.5). The percentage that maintained, decreased, or 
stopped feeding, or maintained or decreased feeding frequency, was not statistically different 
between the two groups. In both groups, practices were split 50–50: roughly half of 
caregivers increased or maintained the quantity of food, and half decreased it or stopped 
feeding.
Table 8.5 Feeding Practices during Diarrheal Illness for Children 6 Months or Older 
AIN-C NO-GMP TOTAL 
% Number 
of 
Children 
% Number 
of 
Children 
% Number 
of 
Children 
Quantity of food per feeding†
Increased†† 10.6 3.9 7.4
Maintained 41.1 48.0 44.4
Decreased 38.6 38.0 38.3
Stopped feeding 9.8
246 
10.0
229 
9.9
475 
Frequency of feeding† ** 
Increased†† 12.7 4.9 8.9
Maintained 54.8 59.7 57.1
Decreased 32.6
221 
35.4
206 
34.0
427 
* Refers to the number of children with diarrhea who were eating solid foods at the time of the episode. 
** Only mothers who did not stop feeding. 
†
 p < 0.05; †† p < 0.01 
If the child was still breastfeeding, the respondent was asked about the frequency of 
breastfeeding during the diarrhea episode. Women whose children were enrolled in AIN-C 
were more likely to have increased breastfeeding (45 percent) during the recent diarrhea 
episode compared to women whose children were not enrolled in CBGP (34.6 percent). This 
difference was statistically significant (p<0.05). Also, a highly significantly (p<0.01) greater 
percentage of “No-GMP” participants decreased or ceased breastfeeding (10.9 percent) 
compared to AIN-C participants (3.3 percent). 
Figure 8.8 Breastfeeding Practice during Diarrheal Illness 
0
10
20
30
40
50
60
Increased Breastfeeding Maintained Breastfeeding Decreased or Ceased
Breastfeeding
AIN-C
No-GMP
52
8.3.2 Feeding During Acute Respiratory Infection 
Respondents whose children were six months of age or older and had an episode of acute 
respiratory infection (ARI) in the two weeks prior to the survey were asked about the quantity 
of food fed during this illness and the frequency of those feedings. The proportion of mothers 
who increased the quantity and frequency of feeding during ARI was greater in the AIN-C 
group (4.9 percent and 8.5 percent, respectively) than in the “No-GMP” group (0.7 percent 
for both). This difference was statistically significant (see Table 8.6). However, it is important 
to note that the percentage of AIN-C participants who maintained feeding was lower than the 
percentage of “No-GMP” mothers who did so; the combined percentage who increased or 
maintained feeding was higher in the “No-GMP” group, although this difference was not 
statistically significant. 
Also, as was the case for feeding during diarrhea, roughly half of the caregivers in both 
groups increased or maintained the quantity of food they fed during the ARI episode. For the 
frequency of feeding, roughly 60 percent of caregivers in both groups increased or 
maintained the frequency of feeding during the ARI episode. 
Table 8.6 Feeding Practices during ARI: Children 6 Months or Older 
AIN-C NO-GMP TOTAL 
% Number 
of 
Children 
% Number 
of 
Children 
% Number 
of 
Children 
Quantity of food per feeding†
Increased†† 4.9 0.7 2.9
Maintained 42.6 51.0 46.7
Increased or Maintained 47.5 51.7 49.6
Decreased 46.9 41.8 44.4
Stopped feeding 5.6
162 
6.5
153 
6.0
315 
Frequency of feeding†† **
Increased††† 8.5 0.7 4.7
Maintained 53.6 57.7 55.6
Increased or Maintained 62.1 58.4 60.3
Decreased 37.9
153 
41.5
142 
39.7
295 
** Only mothers who did not stop feeding;† p < 0.10; †† p < 0.05; ††† p < 0.01 
Among caregivers who were breastfeeding, those participating in AIN-C were highly significantly 
(p<0.01) likely to increase breastfeeding during the ARI episode, and significantly (p<0.05) less likely 
to decrease or cease breastfeeding (See figure 8.9). 
53
Figure 8.9 Breastfeeding Practices during ARI Illness 
0
10
20
30
40
50
60
70
Increased Maintained Decreased or Ceased
Percent
AIN-C
No-GMP
 
8.4 Discussion of Findings 
The conceptual framework for the impact of the AIN-C program on malnutrition shows the 
pathway to improved nutrition and health being mediated by improved practices. Infant and 
young child breastfeeding and complementary feeding practices among healthy children and 
also among children who are ill have a major impact on nutritional status. The results of this 
evaluation show that the AIN-C program significantly improved practices related to 
breastfeeding and frequency and quantity of meals for children 6–23 months, as well as 
feeding practices during illness when compared to practices in the “No-GMP” group. The 
results also show a major gap in appropriate feeding practices specifically related to the 
amount of food provided to children 6–23 months that was not adequately addressed in the 
AIN-C program. 
Exclusive breastfeeding is recommended for all children from birth to six months of age. 
Several different analyses of breastfeeding practices conducted for this evaluation support 
the conclusion that the AIN-C program had a positive impact on breastfeeding practices. 
Mothers practiced exclusive breastfeeding longer, as shown both in terms of the percentage 
of mothers who were practicing exclusive breastfeeding at six months (55.8 percent of AIN-C 
mothers compared to 40 percent of “No-GMP” mothers) and in terms of the median duration 
of exclusive breastfeeding. Among AIN-C mothers, the median duration was 1.5 months 
longer than among “No-GMP” mothers. Furthermore, the rate at which mothers cease 
exclusive breastfeeding is much slower in the early months and more rapid as six months 
nears; while true for both groups, a greater proportion of AIN-C mothers ceased exclusive 
breastfeeding closer to six months that the “No-GMP” mothers. 
The AIN-C program also effectively improved some practices related to complementary 
feeding of children 6–23 months compared to the “No-GMP” group, but the results show that 
much more attention needs to be paid to the amount of food that children are provided. 
Overall, the AIN-C program had a positive impact on the frequency of complementary 
feeding for children 6–8 months and 9–11 months: 98.7 and 94 percent, respectively, in AIN￾C met the recommendation, compared to 88.9 and 84 percent, respectively, in the “No-GMP” 
group. However, even among the AIN-C group, where more than 90 percent of children 9–11 
months met the recommendation for frequency of feeding, only 15.6 percent met the 
recommendation for the amount of food needed. Although almost double the proportion of 
children in the “No-GMP” group, this reflects very few children being fed adequately. Among 
children 12–23 months of age, the AIN-C group showed some improvement in the frequency 
of providing food compared to the “No-GMP” group; 16.1 percent compared to 5.8 percent, 
54
respectively, however this shows a dramatic decline in the proportion of children who are 
meeting the recommendation for frequency of feeding compared to the 9–11 month age 
group. The percentage of children who received the appropriate quantity of food dropped in 
the AIN-C group; there was no difference in meeting this recommendation between the AIN￾C and “No-GMP” children for this age group. 
The AIN-C program also provided counseling to caregivers regarding feeding when their 
child was ill. The evaluation shows significant differences in breastfeeding and provision of 
food to children with diarrhea or ARI. AIN-C participants are much more likely to increase 
breastfeeding and increase the frequency of feeding when their children have diarrhea or 
ARI. However, when looking at those who at least maintained the quantity of food provided 
and the frequency of feeding, the results show that about half of both the AIN-C and “No￾GMP” caregivers either maintained or increased the quantity and the frequency of feeding 
their children during diarrhea and ARI. Overall, the impact of AIN-C on feeding practices 
during illness at the time of this evaluation is uncertain. 
Summary of Key Findings: Infant and Young Child Feeding Practices 

 Exclusive breastfeeding rates among infants younger than six months of age were 
significantly higher in AIN-C children compared to “No-GMP” children. 

 Median duration of exclusive and any breastfeeding was higher in the AIN-C group. 

 The rates at which exclusive and any breastfeeding were discontinued were significantly 
different in the AIN-C and “No-GMP” groups: In the AIN-C group, mothers ceased 
exclusive breastfeeding much more slowly—75 percent of mothers were continuing to 
breastfeed exclusively their 4.1 month old infants. In the “No-GMP” group, 75 percent 
were continuing exclusive breastfeeding only to 2.8 months of age. 

 The proportion of children under 12 months who were fed appropriately, according to 
recommendations both in terms of breastfeeding and frequency of feeding 
complementary foods was significantly higher for AIN-C than “No-GMP.” 

 The proportion of mothers who met the recommendations for the amount of food to 
provide children between 6 and 23 months was extremely low for both groups. However, 
the percentage that did meet the recommendation was significantly greater in the AIN-C 
group than in the “No-GMP” group for children 9–11 months. There was no significant 
difference in meeting the recommendation for 12–23 month old children. 

 AIN-C mothers were more likely to increase the quantity and frequency of feeding during 
diarrheal illness and an episode of ARI, and less likely to decrease or cease 
breastfeeding than mothers in the “No-GMP” group. 

 About half of both the AIN-C and “No-GMP” mothers either increased or maintained the 
quantity and frequency of food provided to their children during a diarrhea disease or 
ARI. 
55
9. Care Seeking and Treatment for Common 
Childhood Illnesses 
Caregivers were asked whether the child had an episode of diarrhea or an acute respiratory 
infection (ARI) in the two weeks prior to the survey interview. For those children who had 
been ill, caregivers were asked about care seeking, home care, and feeding practices related 
to that episode of illness. Caregivers were also asked about their perceptions of danger signs 
for these illnesses. This chapter will present data collected on diarrheal and ARI. Feeding 
during illness episodes was covered in the previous chapter. 
9.1 Prevalence of Diarrhea 
Children enrolled in the AIN-C group were more likely to have had an episode of diarrhea 
(49.1 percent) compared to those in the non-participant group (46.4 percent). The survey 
also determined if the diarrhea varied on two measures of severity: the presence of blood in 
the stool and the duration of the episode. The prevalence of blood in the stool was similar in 
the two groups. Persistent diarrhea was more common among “No-GMP” than AIN-C 
children, however the difference was not statistically significant. 
Table 9.1 Prevalence of Diarrhea in Children by Program Type 
AIN-C No GMP 
Percent Number of 
Children Percent Number of 
Children 
Any Diarrhea† 49.1 43.1 
Bloody Diarrhea (percent among those who 
had diarrhea) 
4.5 5.2 
Persistent Diarrhea – 14+ days (percent 
among those who had diarrhea) 
5.2 
591 
6.0 
622 
 
†
 (p < 0.05) 
9.2 Use of Oral Rehydration Therapy 
Caregivers whose children had diarrhea were asked about the use of rehydrating fluids to 
treat it. Three types of rehydration fluids are reported in Table 9.2: breastfeeding; Oral 
Rehydration Solution (ORS); and a combination of ORS and any home fluid. Children 
enrolled in AIN-C were significantly more likely to receive each of these fluids in comparison 
with “No-GMP” children. 
Half of children (51 percent) enrolled in AIN-C were given increased fluids, including 
breastfeeding, during the episode of diarrhea, compared with 38.4 percent of non-enrolled 
children. Caregivers were asked if they gave the child ORS (known locally by its product 
name, Litrosol). One of every two children (49.7 percent) enrolled in AIN-C were given ORS 
compared to only 22.2 percent of children in the “No-GMP” group. Finally, caregivers were 
56
asked about a range of home fluids promoted by the MOH for use during diarrhea (teas from 
medicinal plants, coconut water, rice water, and natural juices). Sixty-two percent of children 
enrolled in AIN-C received either ORS or a home fluid, compared to 50.8 percent of the “No￾GMP” children. 
Table 9.2 Use of ORS or other Fluids for Diarrheal Disease 
AIN-C No GMP 
Type of fluid provided Percent Number of 
Children Percent Number of 
Children 
Increased fluids including breastfeeding† 51.0 148 38.4 103 
ORS (Litrosol) †† 49.7 144 22.0 59 
ORS or any home fluid†† 62.1 180 37.7 101 
 
†
 (p < 0.01); †† (p < 0.001) 
For those children who received ORS, 42.0 percent of children in AIN-C programs started 
their ORS on the first day of their diarrhea episode compared to 33.3 percent of “No-GMP” 
children. This difference was not statistically significant. 
9.3 Care-Seeking for Diarrhea 
Caregivers were asked whether they sought care for their child with diarrhea and if so, where 
they sought care and the type of advice they received. Caregivers of children participating in 
AIN-C were significantly more likely to seek care from an AIN monitora, other trained 
volunteer, or health professional than were “No-GMP” children. This difference between 
groups is attributable to the presence of the monitora. The proportion of care-seeking from a 
CESAR/CESAMO was similar among children enrolled in AIN-C (18.3 percent) and those in 
the “No-GMP” group (16.0 percent). 
Table 9.3 Source of Care for Diarrhea (responses are not mutually exclusive) 
AIN-C No GMP 
Percent Number of 
Children Percent Number of 
Children 
Did not seek care 55.2 66.8 
Sought care from any source††† 44.8 290 33.2 268 
Sought care from monitora, other trained volunteer, 
or health provider*†† 
82.3 69.7 
Sought care from monitora†††† 39.5 2.2 
Sought care from CESAR/CESAMO† 41.1 48.3 
Sought care from private doctor or clinic 14.0 
129 
16.8 
89 
* Monitora, CESAR/CESAMO, doctor or private clinic, hospital, or health worker other than monitora. 
† (p.10; χ
 2 test); †† (p.05; χ
 2 test); ††† (p.01; χ
2
 test); †††† (p.001; χ
 2 test) 
57
The leading types of advice provided by the CESAR/CESAMO nurses and the monitoras for 
diarrhea are presented in Table 9.4. Because care-seeking for diarrhea is relatively 
infrequent (care for common diarrhea can be effectively provided in the home), the figures 
presented in Table 9.4 are based on a small number of cases. Both the CESAR/CESAMO
staff and the monitora consistently recommend Litrosol for the child’s diarrhea. For both 
sources of care, the second leading type of advice is to give medication to a child with 
diarrhea. Staff at the CESAR/CESAMO rarely advise caregivers to provide increased fluids, 
continue breastfeeding, or continue feeding. 
Table 9.4 Leading Types of Advice Received by Caretakers of Children with Diarrhea 
AIN-C No-GMP 
By AIN-C Monitora By Health Providers By Health providers 
Advice Given 
% Number of 
caretakers 
consulting 
this provider 
% Number of 
caretakers 
consulting 
this provider 
% Number of 
caretakers 
consulting 
this provider 
Give Litrosol 80.4 74.5 62.8 
Give medication 39.2 52.9* 69.8* 
Continue giving food 15.7 11.8 4.7 
Continue breastfeeding 13.7 11.8 2.3 
Give boiled/treated water 9.8 23.5 26.3 
Give more liquids 13.7 
51 
3.9 
53 
2.3 
43 
*Gave the medication (not advice) 
9.4 Caregivers’ Knowledge Concerning Danger Signs in Children with Diarrhea 
All caregivers, regardless of whether their child had a recent episode of diarrhea, were asked 
about signs of severe diarrhea. Table 9.5 below presents responses that are related to 
dehydration caused by diarrhea (as opposed to general danger signs) that were 
spontaneously cited by the caregivers. Four of the five signs (sunken eyes, irritability/crying, 
excessive thirst, and wrinkled, dry skin [skin pinch]) were significantly more likely to be 
spontaneously cited by caregivers with AIN-C-enrolled children compared to caregivers in 
the “No-GMP” group (Figure 9.1). The AIN-C group (28.2 percent) also mentioned lack of 
appetite and not eating as signs of severe diarrhea, compared to 21.2 percent among the 
“No-GMP” group. The “No-GMP” group mentioned persistent diarrhea (lasting longer than 14 
days) as a sign of severe diarrhea more frequently than the AIN-C group, 13.8 and 10.5 
percent, respectively. 
Figure 9.1 Caregiver Knowledge, Signs of Severe Diarrhea 
0
10
20
30
40
50
60
70
Sunken eyes Irritability/crying Excessive thirst Wrinkled/dry skin
AIN-C
No-GMP
58
Table 9.5 Caregivers’ Perceptions Concerning Danger Signs of Severe Diarrhea 
AIN-C No GMP 
% who 
know 
sign 
Number of 
Caregivers % who 
know sign 
Number of 
Caregivers 
Individual danger signs cited spontaneously: 
Sunken eyes†††† 57.6 40.8 
Irritability/crying††† 29.9 22.7 
Excessive thirst†† 9.8 5.8 
Lethargy/unconsciousness 19.9 20.1 
Wrinkled and dry skin (skin pinch) †††† 25.5 
588 
12.1 
622 
Caregivers mentioning any 2 or more of the 
above signs of dehydration 
46.9 591 29.1 622 
Caregivers mentioning blood in the stool as 
sign of severe diarrhea 
1.2 7 1.0 6 
Caregivers mentioning persistent diarrhea (> 
14 days) as a sign of severe diarrhea††
10.5 62 13.8 86 
Caregivers mentioned lack of appetite or not 
eating as a sign of severe diarrhea†††
28.2 166 21.2 132 
†† (p.05; x2
 test); ††† (p.01; x2
 test); †††† (p.001; x2
 test) 
9.5 Prevalence of Acute Respiratory Infections 
Caregivers were asked if their children had experienced cough or difficulty breathing in the 
two weeks prior to the survey interview. Those women who responded affirmatively were 
then asked about a set of additional symptoms. For the purposes of this analysis, caregivers 
who reported that their child experienced rapid breathing in addition to cough or difficulty 
breathing were considered to have an ARI. 
The prevalence of ARI was similar in the AIN-C and the “No-GMP” groups (Figure 9.2): 
among AIN-C children, the prevalence was 22 percent compared to 21.1 percent in the “No￾GMP” group. 
Figure 9.2 Prevalence of Acute Respiratory Infection (ARI) in children 
10
15
20
25
30
ARI episode in past two weeks
AIN-C
No-GMP
59
9.6 Care-Seeking for Acute Respiratory Infection (ARI) 
 
Of those who sought care from an appropriate provider for an ARI episode, the results were 
analyzed by type of provider: (1) AIN monitoras trained in the disease management module, 
(2) CESAR/CESAMO staff, and (3) medical personnel outside the CESAR/CESAMO
including doctors, clinics, and hospitals. 
Forty-four percent of children sampled with ARI were taken to a trained provider for 
treatment. The percentage seeking care from a trained source was similar among children 
enrolled in AIN-C (44.6 percent) and “No-GMP” children (41.2 percent). Monitoras were 
mentioned as a source of care only for those enrolled in CBGP. Twelve percent of children 
participating in AIN-C who had ARI were taken to a monitora. CESAR/CESAMO staff was 
most frequently mentioned as a source of care for both AIN-C participants (25 percent) and 
“No-GMP” participants (25.2%). Other medical providers were consulted less frequently by 
AIN-C participants (13.1 %) compared to “No-GMP” participants (16.8 %). There were no 
significant differences by type of program participation in care-seeking for ARI at a 
CESAR/CESAMO or from other medical providers. 
Table 9.6 Care-Seeking for ARI 
AIN-C No-GMP 
Source of Care 
% Number 
of 
caretakers 
% Number of 
caretakers 
Did not seek care from any trained source 55.4 58.8
Sought care from a pneumonia volunteer, 
monitora or professional health provider18
44.6 41.2
Sought care from an AIN-C monitora 12.3 0.0
Sought care from CESAR/CESAMO 25.4 25.2
Sought care from private doctor/clinic 13.1
131 
16.8
130 
The mean number of days after the beginning of the ARI episode before children were taken 
for care was 2.6 days for AIN-C compared to 3.0 days for “No-GMP” children. 
9.7 Caregivers’ Knowledge Concerning Danger Signs in Children with ARI 
All women surveyed (regardless of whether their child had a recent episode of ARI) were 
asked about signs of severe respiratory illness. Results from this set of questions appear in 
Table 9.7. The responses focus specifically on signs of ARI and exclude general danger 
signs associated with childhood illnesses. Compared to the caregivers of “No-GMP” children, 
caregivers of children enrolled in AIN-C were more likely to spontaneously cite the three 
signs most commonly associated with ARI in Honduras: rapid or labored breathing; chest in￾drawing when breathing, and stridor (making strange sounds when breathing). 
 
18 Two cases of caregivers who sought care from a pneumonia volunteer were also included in this 
group as another trained, community-based source of care. 
60
Table 9.7 Caregivers’ Perceptions of Danger Signs in Children with ARI 
AIN-C NO-GMP TOTAL 
Danger signs mentioned 
spontaneously 
% Number 
of 
Women 
% Number 
of 
Women 
% Number 
of 
Women 
Rapid breathing (labored breathing)†
 69.5 408 63.8 397 66.6 805 
Chest in-drawing when breathing††† 27.6 162 12.4 77 19.8 239 
Stridor (strange noises when 
breathing) ††
28.8 169 21.6 134 25.1 303 
† (p.05; χ
2
 test); †† (p.01; χ
 2 test); ††† (p.001; χ
 2 test) 
9.8 Discussion of Findings 
The difference in the prevalence of illness between AIN-C and “No-GMP” children was 
minimal, although AIN-C children did experience more diarrhea than “No-GMP” children—
likely a reflection of socio-economic status. Through AIN-C, some efforts have been made to 
improve hygiene practices, and efforts to increase breastfeeding have already been noted 
(see chapter eight), but overall the program did not intervene to reduce the prevalence of 
diarrhea or ARI. AIN-C efforts concentrated on reducing the impact of an episode of illness 
on the growth and development of a child. The results of this evaluation show that AIN-C 
participants did receive better care during illness compared to “No-GMP” children. 
The better home care of children with diarrhea in the AIN-C group compared to the “No￾GMP” group reflects the fact that appropriate treatment at home—increasing fluid, especially 
breastfeeding, during a diarrheal episode occurred among 62.1 percent of AIN-C caregivers 
compared to 37.7 percent of “No-GMP” caregivers—was promoted in the initial AIN-C 
program materials. The complete disease management module, which addressed issues 
related to seeking care outside of the home, was introduced later in the program. Notably, 
however, the MOH also has conducted repeat campaigns about home fluids and the use of 
ORS for dehydration with diarrhea on a national basis. The significant difference between the 
groups appears to reflect the added value of the personal contact offered by the AIN-C 
program. 
Although caregivers need not seek assistance for every episode of diarrhea, knowing when 
support from a trained health care provider is needed can save many lives. A significantly 
higher proportion of AIN-C caregivers than “No-GMP” caregivers knew most of the danger 
signs of severe diarrhea. The exception was that neither group knew that blood in the stool 
nor prolonged diarrhea needed immediate attention. When care was sought, the advice 
provided most often by health providers for both groups was “give Litrosol.” Only 15.7 
percent of the monitoras provided the advice to continue feeding and more than one-third 
suggested giving medication. 
The results for care-seeking related to ARI show that AIN-C participants had comparable low 
rates of care-seeking for ARI to the “No-GMP” group: 55.4 percent and 58.8 percent, 
respectively. The groups differ, however, in the length of time they waited before seeking 
care from a trained health provider and in the knowledge of danger signs. AIN-C participants 
took the child for care 2.6 days after the beginning of the illness compared to 3.0 days for the 
“No-GMP” group. Also notable is that only 12.3 percent of caregivers sought support from the 
monitora when their child had an ARI episode. One explanation for the lack of appropriate 
responses to ARI episodes among the AIN-C participants is that the illness management 
61
module was introduced only a few months before this evaluation survey was conducted. 
Furthermore, four of the health areas surveyed had had the extra-budgetary support from 
USAID and the supervisory support from the BASICS project suspended. Also, perhaps 
because of the severity of ARI, when the child has difficulty breathing families tend to go 
directly to the health center, particularly since a process review of the AIN-C program 
showed that the monitoras do not always have the antibiotics needed to treat ARI. 
Summary of Key Findings: Care-Seeking and Treatment for Common Childhood 
Illnesses 

 Diarrhea is more common among AIN-C participants than among non-participants. 
Economic status likely influences this prevalence. 

 AIN-C-enrolled children are significantly more likely to receive rehydrating fluids for their 
diarrhea, whether it is increased fluids including breastmilk; Litrosol (ORS), or a 
combination of ORS and home fluids. 

 Among AIN-C children with diarrhea, the AIN monitora is sought for care/advice with 
equal frequency as the CESAR/CESAMO. 

 Whether care is sought from the CESAR/CESAMO or the AIN-C monitora, few mothers 
remember advice on continued feeding during diarrhea, and a significant portion are told 
to give some type of medication. 

 Prevalence of ARI is similar among children in AIN-C and “No-GMP” groups. 

 AIN-C caregivers are significantly more knowledgeable on danger signs associated with 
diarrhea and ARI. 

 AIN-C children with ARI are not more likely to be taken to a trained provider for care than 
“No-GMP” children. However when care is sought, AIN-C caregivers take their children 
for care significantly sooner than do caregivers in the “No-GMP” group. 
62
10. Delivery and Postpartum Care 
10.1 Delivery 
Data collected on the location of delivery showed that 47 percent of the children in the AIN-C 
households and 35.7 percent of children in the “No-GMP” households were delivered by 
midwives at home. About 45 percent of AIN-C children were delivered in a hospital or private 
clinic compared to almost 59 percent of the “No-GMP” group. 
Table 10.1 Place of Delivery by Program Participation 
AIN-C No-GMP 
Place of Delivery 
% Number of 
children 
% Number of 
children 
Home without specialized assistance 5.4 3.2 
Home with a midwife 47.2 35.7 
Home with doctor or nurse .2 .5 
At CESAR or CESAMO .7 .8 
At a hospital or private clinic 45.7 58.5 
Other .8 
591 
1.3 
622 
Figure 10.1: Place of Delivery by program participation: Home with assistance, Home 
without assistance, Health Facility, or other 
AIN-C
At home without
assistance
At home with
assistance
In a health facility
Other
No-GMP
63
10.2 Postpartum Home Visits 
When asked whether anyone from the community had visited the mother after delivery to 
check on her health and that of her child, AIN-C respondents were significantly more likely to 
have received a visit (37.7 percent) than “No-GMP” mothers (25.6 percent). This difference 
was statistically significant (p<=.001). 
Figure 10.2 Postpartum Home Visits by Program Participation 
0
10
20
30
40
50
60
70
80
Received home visit No visit
Percent
AIN-C
No-GMP
When asked who had made the postpartum visit, 30.9 percent of respondents in the AIN-C 
program who had received a visit reported that the monitora was the only person who had 
visited, 54 percent had received a visit from a traditional birth attendant (TBA), and only 5.4 
percent received a visit from the CESAR/CESAMO nurse. An additional 3.6 percent reported 
receiving a visit from both the monitora and a TBA. Few respondents reported receiving a 
visit from private doctors or nurses or other individuals. 
Although TBAs were the most common postpartum visitor, they were the only significant 
visitor for homes not enrolled in AIN-C. Among “No-GMP” women, 85.5 percent reported that 
they received a visit from a TBA, and 6.3 percent reported receiving a visit from a 
CESAR/CESAMO staff person. Again, few respondents reported receiving a visit from 
private doctors or nurses or other individuals. 
Table 10.2 Personnel Making Postpartum Home Visits 
AIN-C No-GMP 
Type of Personnel 
% Number 
of 
mothers 
% Number of 
mothers 
AIN monitora 30.9 0.6
Birth attendant (partera) 54.3 85.5
Staff from CESAR or CESAMO 5.4 6.3
Private doctor/nurse 0.9 3.1
Other (including combinations of personnel 
listed above) 
8.5
223 
4.4
159 
64
Respondents who reported receiving a postpartum visit from a monitora or a TBA were 
asked what was done during the visit.19 For those AIN-C participants who were visited by the 
monitora, 64.3 percent reported that the monitora gave the child a physical examination and 
25.7 percent reported that the monitora examined the mother. Just over half the caregivers 
reported that the monitora enrolled the child in AIN-C or invited the mother to participate in 
the program. More than eight percent received a referral to the health center from the 
monitora during the postpartum visit. 
Table 10.3 Activities during Monitora’s Postpartum Visit to AIN-C Participants 
Activity* Percent Total Number 
of Women 
Made physical exam of the mother 25.7 18 
Made physical exam of the child 64.3 45 
Registered the child in the AIN-C program 41.4 29 
Gave the mother a child health card 11.4 8 
Gave the mother a vaccination card for the child 1.4 1 
Asked mother to bring the child to monthly AIN-C weighing sessions 25.7 18 
Gave referral to higher level of care 8.6 6 
* These responses are not mutually exclusive. 
Of the respondents that were visited by a TBA, the main activities were examining the 
mother and the baby: 74 percent of AIN-C and “No-GMP” participants reported a maternal 
physical exam during the postpartum visit by the TBA, and 88.1 percent of AIN-C and 94.9 
percent of “No-GMP” mothers reported that the TBA conducted a physical examination of the 
child. The TBAs did not actively encourage participation in the AIN-C program. 
10.3 Discussion of Findings 
These results show fewer AIN-C mothers use facilities for delivery of infants. This is likely a 
reflection of the targeting of the AIN-C program toward communities with greater socio￾economic needs. At the time of this evaluation, a relatively small amount (not quite 40 
percent) of AIN-C participants received post-delivery visits, although the percentage was 
even smaller for “No-GMP” mothers (26 percent). Given the higher proportion of home births 
among AIN-C mothers, these postpartum follow-up visits are very important. Visits by the 
monitoras accounted for the higher percentage of AIN-C women receiving postpartum visits. 
This may be in part due to the fact that these monitoras are TBAs or have had midwifery 
training. The tasks carried out at the postpartum visit are those expected of TBAs—
examination of the mother and the child. Only half of the mothers who received post-delivery 
visits registered the child in the AIN-C program or were encouraged to bring the child to 
weighing sessions. This is a missed opportunity to begin promotion of appropriate newborn 
care and to discuss danger signs for the mother and newborn and for referral of problems. 
Another missed opportunity is reaching out to the TBAs (who are not monitoras) to 
encourage them to promote the enrollment of newborns in the AIN-C program and to make 
referrals to the monitora.
20
 
 
19 Data were unavailable for 8 respondents visited by a monitora and 5 respondents visited by a TBA. 
20 A newborn component for the AIN-C program was developed and introduced in 2003-2004. 
65
Summary of Key Findings: Delivery and Postpartum Care 

 AIN-C women are more likely to deliver their babies at home than “No-GMP” women, 
who are more likely to deliver at a hospital or private clinic. 

 AIN-C households were visited significantly more often in the postpartum period than 
non-participant households. 

 A TBA is the predominant person visiting homes; the higher visitation rate for AIN-C 
households is due primarily to postpartum visits made by monitoras. 

 The main postpartum activities for all who received postpartum visits, including from the 
monitoras, were examination of the mother and baby. Enrollment in AIN-C occurred at 
fewer than half of the postpartum visits to participating mothers. 
66
11. Vaccinations and Micronutrients
11.1 Vaccinations 
Respondents were asked if they had a vaccination card for their child. If the respondent 
presented a card, the interviewer recorded information from the card on the antigens and 
doses received and the month of each immunization. If the caregiver had no vaccination 
card, she was asked a series of questions about the types of vaccines that her child received 
and the number of doses. Respondents with children enrolled in AIN-C were more likely to 
have their child’s vaccination card (93.1 percent) than were caregivers of “No-GMP” children 
(85.5 percent). The difference in card ownership was significant (p=.002). Information from 
both of these sources (card and mother’s recall) was used to calculate the proportion of 
children 13 through 23 months who were fully immunized. “Fully immunized” refers to those 
children who have received BCG vaccine, three doses of diphtheria, pertussis and tetanus 
vaccine (DPT), three doses of polio vaccine, and a measles vaccination. 
It should be noted that the 13 through 23 month age construct is a slight departure from 
international norms for the calculation of vaccination coverage rates, which typically measure 
coverage from 12 months of age. This modification was made to recognize that the current 
immunization schedule in Honduras calls for application of measles at, but not before, 12 
months of age. Thus, all children 13 through 23 months have had one month in which to 
receive their immunization with measles to complete the childhood vaccination schedule. 
Results presented in Table 11.1 show no significant differences in the percent of children 
fully immunized at 23 months of age: 76.5 percent of children participating in AIN-C were 
fully immunized, compared to 71.3 percent of “No-GMP” children. The difference between 
groups was not statistically significant (p=.168). However, AIN-C children were much more 
likely to have received all necessary immunizations by 13 months in accordance with the 
GOH policy: 51.5 percent of AIN-C participants were fully immunized by 13 months, as 
opposed to 42.8 percent of “No-GMP” children. 
Figure 11.1 Immunization Coverage through 23 Months by Program Participation 
76.5
51 5
71.3
42.8
40
45
50
55
60
65
70
75
80
Fully Immunized age 13-23
months
Fully immunized by 13 months
among children 13-23 months
Percent
AIN-C
No-GMP
 
67
11.2 Vitamin A and Iron Supplementation 
Interviewers were instructed to record information from the vaccination card on the status of 
iron and vitamin A supplementation. If iron supplementation was not noted on the card or if 
no card was presented, then the interviewer asked the caregiver whether the child had 
received iron and, if so, how many times. Data presented in Figure 11.2 reflect the 
percentage of children who received iron, based either on the information extracted from the 
health card or the caregiver’s recall. As iron is to be given to children four months of age and 
older, these questions represent the results for children 4 months or older. Children enrolled 
in AIN-C were far more likely to receive iron supplementation (65.6 percent) than were 
children in the “No-GMP” group (29.5 percent). The difference is highly statistically significant 
(p<.001). 
Figure 11.2 Iron Supplementation among Children 4 through 23 Months of Age by 
Program Participation 
65.6
29.5
0
10
20
30
40
50
60
70
Received Iron Supplementation
Percent
AIN-C
No-GMP
Data on vitamin A supplementation were handled in a manner similar to that of iron, with 
coverage estimates based on both health card data, when available, and caregiver’s recall. 
Among children enrolled in AIN-C, 94 percent received vitamin A compared to 87.5 percent 
of “No-GMP” children. This difference between AIN-C-enrolled and “No-GMP” children was 
also highly significant; p<.001 (Figure 11.3). 
Figure 11.3 Vitamin A Supplementation among Children 6 through 23 Months of Age, 
by Program Participation 
94.3
87.5
80
85
90
95
100
Received Vitamin A Supplementation
Percent
AIN-C
No-GMP
 
68
11.3 Deworming 
All respondents were asked if their child was dewormed (un deparasitante). There were no 
significant differences in the rate of deworming between children in AIN-C and those in the 
“No-GMP” group; 59.4 percent compared to 63 percent, respectively. 
Figure 11.4 Deworming in Children, by Program Participation 
30
40
50
60
70
Percent
Received Deworming
AIN-C
No-GMP
11.4 Discussion of Findings 
The GOH has had an ambitious national effort to ensure that children receive vitamin A 
supplementation; note the high coverage (87.5 percent) in the “No-GMP” group. The benefit 
of a community-based program is shown by the substantial additional coverage achieved 
through AIN-C for both vitamin A and iron supplementation. Furthermore, the higher 
proportion of children who are fully immunized by 13 months shows another significant 
benefit of the AIN-C program. The fact that there is no difference in deworming coverage is 
not surprising, as deworming has not been a part of the AIN-C program. Deworming is 
offered only at the health center.
Summary of Key Findings: Vaccinations and Micronutrients 

 Significantly more AIN-C children have a record of their immunizations than “No-GMP” 
children. 

 A greater proportion of AIN-C children were fully immunized by 13 months than “No￾GMP” children. 

 Children enrolled in AIN-C were far more likely to receive iron and vitamin A 
supplementation than were “No-GMP” children. 
69
12. Nutritional Status and AIN-C Participation 
All children 0 through 23 months of age were measured using standard anthropometric 
procedures. Weight was measured using hanging scales similar to those used by AIN-C 
program staff and volunteers. Recumbent length was measured using boards created for this 
purpose. Anthropometrists worked as part of each interview team and carried out all weight 
and height measurements. Weight and height data were recorded onto the completed 
questionnaire of each individual child. Weight and height measures were available for all 
children surveyed (1343), although a few outlying z-score values were excluded from the 
analysis. 
12.1 Assessment of Anthropometric Outcomes: Bivariate Analysis 
Descriptive results (mean and median z-scores) for weight-for-height, height-for-age, and 
weight-for-age are presented by age group in Table 12.1. AIN-C participants did not differ 
from the “No-GMP” group in terms of weight-for-height (wasting). Nutrition status as 
measured by height-for-age and weight-for-age show that AIN-C participants had poorer 
nutrition status than the “No-GMP” group in the less than 6 month age group and in the 
combined 0–23 month age group. 
Table 12.1 Mean and Median Values for Anthropometric Measures of Nutritional Status 
among AIN-C and “No-GMP” Children 
Anthropometric 
Measurement 
AIN-C No GMP 
Mean (SD) Median # of 
children
Mean (SD) Median # of 
children 
Weight-for-height 
 < 6 months 0.61 (0.82) 0.54 115 0.48 (0.80) 0.40 132 
 6–11 months 0.11 (0.93) 0.10 171 0.18 (0.96) 0.11 164 
 12–23 months -0.32 (0.87) -0.37 289 -0.34 (0.92) -0.42 315 
 0–23 months -0.00 (0.95) -0.03 575 -0.02 (0.97) -0.05 611 
Height-for-age 
 < 6 months ††† -0.53 (0.88) -0.53 115 -0.20 (0.85) -0.19 132 
 6–11 months -1.09 (1.01) -1.03 171 -0.93 (1.11) -0.94 164 
 12–23 months -1.55 (1.09) -1.59 289 -1.47 (1.10) -1.47 315 
 0–23 months †† -1.21 (1.10) -1.16 575 -1.05 (1.16) -1.04 611 
Weight-for-age 
 < 6 months † 0.10 (0.82) 0.15 115 0.28 (0.82) 0.26 132 
 6–11 months -0.76 (1.07) -0.83 171 -0.60 (1.13) -0.61 164 
 12–23 months -1.15 (1.00) -1.19 289 -1.10 (1.04) -1.14 315 
 0–23 months †
 -0.78 (1.10) -0.86 575 -0.67 (1.16) -0.73 611 
†
p < 0.10; †† p < 0.05; ††† p < 0.01 
70
To interpret the differences in this cross-sectional comparison of these two groups, several 
factors need to be taken into account: 
• As discussed in Chapter 4 of this report, these two groups (AIN-C and “No-GMP”) are not 
equivalent. The AIN-C group is poorer and lives further from the health center. Therefore, 
the two groups differ on factors other than participation in AIN-C that are also related to 
nutritional status. 
• The major difference in nutritional status is in height-for-age under six months. Prior to six 
months of age, height-for-age largely reflects birth weight.21 Many factors influence birth 
weight, including maternal nutritional status at conception, maternal weight gain during 
pregnancy, and maternal height. Each of these maternal factors is in turn associated with 
economic status, and the “No-GMP” group is better-off economically than the AIN-C 
participant group (see Chapter 4). 
• Height is a component of weight, and weight-for-age thus also reflects height-for-age. 
 
• Height-for-age falters globally from birth through approximately 24 months of age.22
Therefore, differences between groups in mean height-for-age would be constant as age 
increases, as long as the rate of faltering is constant for both. Since mean height-for-age 
among 0–5 month-olds is statistically lower in the AIN-C group compared to the non￾participant group, mean height-for-age would be expected to be statistically lower among 
AIN-C participants in the 6–11 and 12–23 month-old groups. However, the data show 
that this is not the case: There are no differences in nutritional status between AIN-C and 
the “No-GMP” groups in the 6–11 and 12–23 month age groups. The lack of a difference 
in mean height-for-age in these age groups may suggest that AIN-C participation protects 
against height-for-age faltering. In other words, AIN-C participant infants may begin life 
shorter-for-age than the “No-GMP” group (perhaps due to higher rates of low birth weight 
births), but they “catch-up” to the “No-GMP” children by the age of 6–11 months, 
indicating more rapid height gain among AIN-C participants. However, this conclusion is 
not definitive, since a pre-and post-intervention comparison of height-for-age status of 
these two groups was not possible. 
12.2 Assessment of Anthropometric Outcomes: Controlling for Confounding Factors 
12.2.1 Differences in Socio-Economic Status 
As mentioned in Chapter 4, the AIN-C and “No-GMP” groups varied in terms of economic 
status; AIN-C participants have lower economic status. Because of the many associations 
between economic and nutritional status, analysis of these outcome data which controls for 
these economic differences, through statistical procedures, may provide additional insights 
into the impact of the AIN-C program. 
No simple and universally effective means exists for controlling economic status. However, a 
number of factors related to economic status were assessed in the caregiver interviews and 
provided the data for creating a socio-economic assets score. Factors were coded so that 
values ascended with higher socioeconomic status (see Table 12.2), and then used to create 
 
21 Some have argued that even at 2 years of age, much of height-for-age is determined by factors 
related to birth weight and maternal nutritional status. 
22 See Shrimpton et al. Worldwide timing of growth faltering: implications for nutritional interventions. 
Pediatrics 2001:107(5). 
71
a standardized assets index score. This score was created by scoring the first component 
from a principal components analysis of all individual records (i.e., both groups combined). 
Table 12.2: Factors Assessed in the Creation of an Asset Score and their Coding 
Factor Low Socio-Economic Status (SES) Higher SES 
Scale Factors (continuum) 
Location/type of water source 
Natural source (river, pond), purchased water /water truck), or “other” 
 Well water  Supply water, distant from homestead  Supply 
water, on homestead 
Number of rooms in dwelling 1 room  2 rooms  3 rooms  4 rooms  5 rooms  6+ rooms 
Distance (in time) to health facility 3+ hours  2 to 3 hours  1 to 2 hours  ½ to 1 hour  < ½ hour 
Categorical Factors 
Sanitation type Non-improved toilet (open pit latrine, 
open air, none, other) 
Improved toilet (indoor 
toilet, water-sealed latrine) 
Cooking fuel Firewood “Modern” (kerosene, 
propane, electricity) 
Dwelling floor type Dirt floor, “other” Wood floor 
Dirt floor, “other” Cement floor 
Dirt floor, “other” Tile floor 
Maternal education No formal education Any formal education 
Maternal employment status Not employed Employed 
Dwelling electrified No Yes 
Possessing radio No Yes 
Possessing television No Yes 
Possessing refrigerator No Yes 
Possessing telephone No Yes 
Possessing motor vehicle No Yes 
There are highly significant differences between AIN-C and “No-GMP” groups in terms of 
asset score, a reasonable proxy for economic status (Table 12.3). 
Table 12.3 Mean and Median Values of the Asset Score Factor used for Estimating 
Economic Status, by Participation Status 
AIN-C No-GMP All 
Statistic Asset 
Score 
Number of 
Households
Asset 
Score 
Number of 
Households
Asset 
Score 
Number of 
Households
Mean †
(standard deviation) 
-0.1637 
(0.9429) 
0.1600 
(1.0287) 
0.0000 
(1.0000) 
Median -0.3136 
557 
0.2009 
570 
-0.1026 
1127 
Note: The score is standardized with mean and standard deviation for both groups combined equal to 
0 and 1, respectively; †
p < 0.001 
72
12.2.2 Differences in Age Distribution 
Additionally, the two groups differ in the distribution of ages of the children. Having groups 
that are comparable in the age distribution is important because of the strong relationship 
between nutritional status and age among children younger than 24 months. Globally, 
infant/child weight-for-length/height deteriorates from approximately 3 months of age through 
approximately 18 months, height-for-age deteriorates from birth through 24 months, and 
weight-for-height deteriorates from 3 months through 12 months. Differences in the age 
distribution of the groups thus can create inherent differences in nutritional status. Table 12.4 
shows the differences between the AIN-C and the “No-GMP” groups in the distribution of 
infant/child age, and the statistically significant association between age and participation 
status. Note that these results suggest that AIN-C participation is more common for younger 
infants/children than for older ones. 
Table 12.4 Age Distribution of Infants/Children according to Participation Status 
AIN-C No-GMP All 
Age group Number of 
Children 
% Number of 
Children 
% Number of 
Children 
% 
0–5 months 115 19.8 132 21.6 247 20.7 
6–11 months 173 29.8 164 26.8 337 28.3 
12-17 
months 
157 27.1 135 22.1 292 24.5 
18-23 
months 
135 23.3 180 29.5 315 26.4 
All ages 580 100.0 611 100.0 1191 100.0 
χ2 for association between age and participation status = 8.324, p=0.040. 
12.2.3 Accounting for Differences 
Two steps were taken to address these differences between the AIN-C and the “No-GMP” 
groups in terms of economic status and the distribution of infant/child ages: To address 
economic status, the analysis was restricted to infants/children from either group whose 
household asset score was equal to or lower than the median value for both groups 
combined—the poorer half of all households. To address differences in age distribution, 
comparisons between groups were controlled for infant/child age by including age in a 
multiple linear regression predicting nutritional status. 
The results of linear regression analysis on the poorest half of households are shown in 
Table 12.5. Infant/child age was highly and negatively associated with all measures of 
nutritional status: with each additional month of age weight-for-height, height-for-age, and 
weight-for-age decreased 0.057, 0.073, and 0.072 Z-score, respectively. AIN-C participation 
was not associated with height-for-age or weight-for-age. No negative association between 
AIN-C participation and height-for-age and weight-for-age is seen in the controlled, restricted 
analysis (uncontrolled analysis had shown such a relationship). A positive and marginally 
significant (p<0.10) effect of AIN-C participation on weight-for-height was observed in the 
controlled, restricted analysis (participation was negatively associated with weight-for-height 
in uncontrolled comparisons). The mean weight-for-height Z score is 0.122 higher among 
AIN-C participants, holding age constant and excluding wealthier households. 
73
Table 12.5 Regression Analysis of Nutritional Status in Relation to AIN-C Participation, 
Controlled for Infant/Child Age (Analysis limited to households with asset scores less 
than or equal to the median asset score) 
Factor n Adj. R2
β Std. 
Error 
t value P 
 
Weight-for-height (F=53.377, 
p<0.001) 
564 0.157 
 Intercept 0.555 0.090 
 Infant/child age (months) -0.057 0.006 -10.130 0.000 
 Participation (AIN-C vs. no 
GMP) 
 0.122 0.074 1.659 0.098 
 
Height-for-age (F=65.593, 
p<0.001) 
564 0.187 
 Intercept -0.612 0.102 
 Infant/child age (months) -0.073 0.006 -11.449 0.000 
 Participation (AIN-C vs. no 
GMP) 
 -0.009 0.083 -0.106 0.916 
 
Weight-for-age (F=69.304, 
p<0.001) 
564 0.195 
 Intercept -0.204 0.099 
 Infant/child age (months) -0.072 0.006 -11.672 0.000 
 Participation (AIN-C vs. no 
GMP) 
 0.090 0.081 1.110 0.268 
12.3 Participation Intensity and Nutritional Status. 
An additional analysis was conducted on the AIN-C participant group to determine whether 
the intensity of program participation was associated with nutritional status, controlling for 
household assets and the infant/child’s age. Participation intensity was defined as the 
percentage of possible weighings that the infant/child had attended. That is, a child whose 
number of lifetime weighings was equal to her/his age in completed months would have 
100% participation intensity, a child whose number of lifetime weighings was half of her/his 
age would have 50% participation intensity. The total number of times that a child was 
weighed was taken from the child’s growth cards shown to the interviewers. Children without 
growth cards were assigned a value of zero for participation intensity. 
Household assets and infant/child’s age were controlled for in this analysis because they are 
associated both with nutritional status and with participation and thus are, by definition, 
confounding factors. Both are negatively associated with participation (i.e., participation rates 
decline with increasing household assets and increasing infant/child age), but this negative 
correlation was weaker for asset score than for age, which would be expected since the AIN￾C group is poorer than the No-GMP group. The Pearson’s correlation coefficients for 
household assets and age with participation intensity were -0.083 (p=0.069) and -0.237 
(p=0.000), respectively. 
With these controls in place, the intensity of participation was positively associated with 
nutritional status for all measures of nutritional status (Table 12.6). The association was 
74
strongest for weight-for-age (p=0.004), less strong but still highly significantly so for weight￾for-height (p=0.011), and marginally significant for height-for-age (p=0.083). The effect of 
participation was greatest for weight-for-age and weight-for-height, at 0.005 and 0.004 Z￾score, respectively. That is, holding infant/child age and asset score constant, for every 1 
percent increase in participation intensity, weight-for-age increased 0.005 Z-score, and 
weight-for-height increased 0.004 Z-score. The model containing participation intensity, age, 
and asset score also explained more than 20 percent of the variability in height-for-age and 
weight-for-age. 
Table 12.6 Results of Multiple Linear Regression Analysis Predicting Nutritional Status 
on the Basis of Participation Intensity, Infant/Child Age and Household Asset Score 
Factor n Adj. R2
β Std. 
Error 
t value P 
 
Weight-for-height (F=25.956, 
p<0.001) 
482 0.135 
 Intercept 0.333 0.153 
 Infant/child age (months) -0.050 0.007 -7.581 0.000 
 Asset score ( X = 0.000 ; 
SD=1.000) 
 0.037 0.043 0.863 0.389 
 Participation intensity (%) 0.004 0.002 2.566 0.011 
 
Height-for-age (F=42.696, p<0.001) 482 0.236 
 Intercept -0.613 0.170 
 Infant/child age (months) -0.061 0.007 -8.318 0.000 
 Asset score ( X = 0.000 ; 
SD=1.000) 
 0.397 0.047 8.368 0.000 
 Participation intensity (%) 0.003 0.002 1.735 0.083 
 
Weight-for-age (F=69.304, p<0.001) 482 0.206 
 Intercept -0.333 0.171 
 Infant/child age (months) -0.063 0.007 -8.458 0.000 
 Asset score ( X = 0.000 ; 
SD=1.000) 
 0.280 0.048 5.858 0.000 
 Participation intensity (%) 0.005 0.002 2.890 0.004 
The interaction of asset score and participation was tested in the models shown in Table 
12.6, and it was statistically significant for weight-for-height (p=0.047) only. Omission of this 
interaction is not strictly proper, because the interaction affects the values and meaning of 
main effects in the regression model. However, inclusion of the interaction makes the 
interpretation of main effects difficult, so the interaction issue is dealt with below by stratifying 
for asset score (Table 12.7). 
Several aspects of the results from stratified analysis are notable: First, age and participation 
intensity explain a much greater proportion of the variability in nutritional status among 
poorer households than they do among more wealthy households. Second, the size of the 
positive effect of participation intensity is 2–3 times greater among poorer households than it 
is among more wealthy households. Finally, the association between participation intensity 
and nutritional status is highly significant for weight-for-height and for weight-for-age among 
75
poorer households, while no association exists between participation intensity and nutritional 
status among more wealthy households. 
Table 12.7 Linear Regression Results for the Prediction of Nutritional Status by 
Participation Intensity and Infant/Child Age, by Asset Score Categories 
Factor Adjusted R2 β
(Std Error) 
P 
Asset score 
≤ median 
(n=283) 
Asset 
score 
> median 
(n=218) 
Asset 
score 
≤ median 
Asset score 
> median 
Asset 
score 
≤ median 
Asset score 
> median 
Weight-for-height 0.185 0.075 
 Intercept 0.250 
(0.193) 
0.424 
(0.239) 
 Infant/child age 
 (months) 
 -0.054 
(0.008) 
-0.043 
(0.010) 
0.000 0.000 
 Participation 
intensity 
 (%) 
 0.005 
(0.002) 
0.002 
(0.002) 
0.005 0.476 
Height-for-age 0.211 0.057 
 Intercept -0.708 
(0.225) 
-0.400 
(0.265) 
 Infant/child age 
 (months) 
 -0.078 
(0.010) 
-0.043 
(0.011) 
0.000 0.000 
 Participation 
intensity 
 (%) 
 0.003 
(0.002) 
0.001 
(0.003) 
0.220 0.610 
Weight-for-age 0.237 0.081 
 Intercept -0.510 
(0.214) 
-0.065 
(0.273) 
 Infant/child age 
 (months) 
 -0.072 
(0.009) 
-0.051 
(0.012) 
0.000 0.000 
 Participation 
intensity 
 (%) 
 0.006 
(0.002) 
0.002 
(0.003) 
0.005 0.400 
12.4 Discussion of Findings 
As discussed in Chapters 2, 4, and 5, the AIN-C and “No-GMP” groups were not equivalent 
in terms of several measures of wealth and education of mothers. “No-GMP” households 
appeared to be significantly wealthier than “AIN-C” households and “No-GMP” mothers 
completed more years of education than “AIN-C” mothers. Further analyses based on an 
“asset score” summarized in this chapter confirm that AIN-C households are poorer than 
“No-GMP” households. Both lower socio-economic status and fewer years of maternal 
education are linked to lower nutrition status. In addition, the age distribution of the AIN-C 
and “No-GMP” samples was significantly different: the AIN-C group had more children 0–11 
months compared to the “No-GMP” group, also related to nutrition status. Analyses of 
76
anthropometric data collected for this evaluation that do not control or account for these 
factors at first glance suggest that AIN-C many not be having a positive impact on nutrition 
status. However, closer examination reveals that while AIN-C children have poorer nutrition 
status in the 0–6 month age-group compared to “No-GMP” children, this difference 
disappears as children get older. This is the opposite of what is expected, as age was highly 
and negatively correlated with all measures of nutritional status. 
When socio-economic and age factors are controlled, the AIN-C program appears to result in 
a marginally positive impact on nutrition status as measured by weight-for-height: the mean 
Z-score was .122 higher among AIN-C children than “No-GMP” children. While the size of 
this overall impact is relatively small, additional analyses that examine the impact on 
households of varying levels of wealth and of intensity of participation in AIN-C show that the 
AIN-C program has a much greater positive impact on the nutritional status of children from 
poorer households, and among those children who have better participation in the program. 
When controlling for the interaction of participation and socio-economic status, the impact on 
nutrition status is two to three times greater among poorer households. The intensity of 
participation in the AIN-C program shows that for every additional month of participation in 
the program, there is a .085 increase in weight-for-age Z-score. 
77
Summary of Key Findings: Nutritional Status 

 Uncontrolled comparisons between AIN-C and “No-GMP” children for nutritional status 
show no positive effect of AIN-C. For some measures of nutritional status, the effect 
actually appears to be negative. However, the AIN-C and non-participant groups are not 
equivalent, and better nutritional status is expected to be higher in the wealthier “No￾GMP” group. 

 Cross–sectional analyses of the anthropometric data show that AIN-C infants may begin 
life shorter-for-age than the “No-GMP” group (perhaps due to higher rates of low birth 
weight births), but they “catch-up” to the “No-GMP” children by the age of 6–11 months, 
indicating more rapid height gain among AIN-C participants. 

 Highly significant differences exist between AIN-C and “No-GMP” groups in terms of 
asset score, a reasonable proxy for economic status. 

 AIN-C and “No-GMP” sample groups differed in the distribution of ages of children. A 
difference in the age distribution of the groups creates inherent differences in nutritional 
status. The AIN-C group included more infants/children in younger age categories. 

 Analyses restricted to the poorest half of households show that infant/child age was 
highly and negatively associated with all measures of nutritional status: with each 
additional month of age weight-for-height, height-for-age, and weight-for-age decreased 
0.057, 0.073, and 0.072 Z-score, respectively. 

 In restricted analyses (controlling for socio-economic status and age) AIN-C participation 
had a marginally positive impact on weight-for-height; mean Z-score was 0.122 higher 
among AIN-C compared to “No-GMP” children. 

 Among AIN-C participants, the intensity of participation is positively associated with 
nutritional status in terms of weight-for-height, height-for age, and weight-for age. The 
strongest impact was on weight-for-age: for every 1 month increase in participation there 
was a .085 increase in Z-score. 

 Controlling for interaction of participation in AIN-C and socio-economic status shows that 
the size of the positive effect of AIN-C participation on nutritional status is two to three 
times greater among poorer than wealthier households. 
78
13. Conclusions 
The AIN-C program in Honduras is an example of a CBGP program taken to scale that 
endured despite a range of “real-life” challenges—reductions in funding, natural disasters, 
and changes in government—and made a significant, positive impact on under-nutrition in 
children 0–2 years of age. Furthermore, the commitment to the program at the community 
level—a by-product of its impact on nutrition in the community—is paralleled at the national 
level, as AIN-C became part of the National Strategy for Nutrition in Honduras. 
This evaluation has pointed to a number of important general and specific design and 
implementation considerations and “lessons learned” for CBGP programs. First, AIN-C is a 
“pro-poor” strategy: the value and impact of the program was show to be significantly greater 
among poorer than more wealthy households. Therefore, the fact that AIN-C was targeted to 
lower socio-economic communities who most need support for nutrition was an important 
factor in its success. A second and related lesson learned is that consistency in participation 
in AIN-C activities is critical to maintaining nutrition status and preventing under-nutrition. 
Children from poor households with the best rates of participation benefited the most from 
AIN-C. 
Beyond these overall program design conclusions—target the poor and ensure consistent 
participation and coverage—this evaluation also highlights a number of more specific 
“lessons learned.” These include: 

 AIN-C had a positive impact on breastfeeding and complementary feeding practices, and 
based on this evaluation appears to have contributed to the improvement in under￾nutrition in children under two. However, the results show that additional efforts are 
needed to expand the reach of improved practices and to achieve more specific changes 
in breastfeeding, complementary feeding, and care practices. 

 The improvement shown in exclusive breastfeeding is encouraging. However, according 
to this evaluation, almost half of AIN-C mothers are not practicing exclusive 
breastfeeding for the full six months. Additional work is needed to identify the barriers to 
continuing exclusive breastfeeding and more effectively support and encourage this 
practice. 

 For complementary feeding practices, AIN-C made a significant impact on the frequency 
of feeding children age 6–11 months. The results of the evaluation also showed some 
improvement in the frequency of feeding in children 12–23 months, but only for about 15 
percent of caregivers. Increasing the frequency of feeding after 11 months deserves 
additional attention in counseling. More notable, however, is that practices related to the 
quantity of food provided to children 6–23 are extremely inadequate. Increasing the 
impact on nutritional status will depend on effectively addressing this practice related to 
the amount of food provided to children in the second year of life. Strategies for both 
increasing skills of and tools for monitoras to more effectively communicate messages 
regarding the quantity of food required for adequate growth and approaches to 
addressing the constraints caregivers face in providing adequate food are needed. 
79

 AIN-C effectively introduced improved care practices for children 0–2 years old suffering 
from diarrhea and ARI, although the evaluation results also indicated areas for 
improvement in home treatment as well as care delivered by health care providers. For 
example, more than half of health providers serving AIN-C participants and more than 
two-thirds of those treating the “No-GMP” children recommended medicine, and none 
promoted continued feeding. Addressing these practices would call for additional training 
to support the appropriate treatment of diarrhea. 

 The evaluation also showed that, overall, the messages related to feeding and care 
practices received by caregivers were somewhat generic, indicating that additional skill￾building in effective counseling techniques would further strengthen the quality of the 
AIN-C program. 
Finally, after this evaluation was completed, a detailed cost-benefit analysis of AIN-C was 
conducted.23 The cost study showed that AIN is a cost-efficient program and highlighted 
those activities that account for the majority of the costs. The long term, annual, recurrent 
cost of the program per child under 5 was estimated at $2.73. The fixed costs, which include 
personnel and non-personnel supervision costs, represent about 40 percent of the costs; the 
long-term, annual, incremental budget required was $4.00 per child under 2, and under $2.00 
per child for children under 5. About 30 percent of the annual recurrent costs were in 
supplies and medicines and 70 percent for personnel and related personnel costs, per diem, 
and transportation. “The average direct cost per child of an AIN-C Program community￾based weighing and counseling session is 10.9 lempiras (US$0.66), just 11 percent of the 
direct cost of a single MOH staff-provided, facility-based, child growth and development 
consultation.” (Fiedler: 2003, p. 63.) 
 
23 Fiedler, J. A Cost Analysis of the Honduras Community-Based Integrated Child Care Program, The 
World Bank: Washington DC, 2003.