Mercy Corps Guatemala
PROCOMIDA Program:
Award Number: AID-FFP-A-09-00005-00
Final External Evaluation Report:
PROCOMIDA’s Community Food Diversification
Program for Mother and Child
(Programa Communitario Materno Infantil de
Diversificación Alimentaria)
December 2016
Evaluation contributors:
i
PROCOMIDA EVALUATION DETAILS
Intervention Title Programa Communitario Materno Infantil de Diversificación 
Alimentaria / Community Food Diversification Program for Mother 
and Child (PROCOMIDA)
Award Number: AID-FFP-A-09-00005-00
Scope: Independent Final Evaluation
Country: Guatemala
Evaluation Date: 10 July 2015 (end of field data collection)
Principle Consultants:
Final Evaluation Re-analysis 
and Re-drafting Consultants:
Evaluation Managers:
PROCOMIDA Approval Date: July 2009
PROCOMIDA End Date: May 30, 2015 (field intervention) 
November 30, 2015 (closing date)
Budget US$ 37.9 million
Key Words: Nutrition, Food aid, Food security, Guatemala, PM2A, FANTA III, 
FFP, USAID, MYAP, M&E, MEL, Final Evaluation
ii
ACKNOWLEDGEMENTS
, acting as an external international evaluation consultant team. The 
consultant team implemented evaluation fieldwork from April - July 2015 under the guidance of 
the PROCOMIDA team, Guatemala. , reviewed endline 
data for statistical relevance and comparison with baseline results, reconfigured data 
presentation, consolidated and integrated reviewer comments, and reorganized and edited the 
final evaluation report.
The evaluation drew from the input provided by a wide range of individuals, both internal and 
external to the PROCOMIDA program. Special thanks goes to the PROCOMIDA management 
team including , to the monitoring and evaluation 
(M&E) team including , and , and to
at Mercy Corps Headquarters.
ABOUT THIS REPORT
This report begins with background information on food insecurity and malnutrition in Guatemala, 
and the PROCOMIDA program. It then presents the endline evaluation methodology. The 
discussion of final evaluation findings begins with program characteristics, and then results 
based on Indicator Performance Tracking Table (IPTT) objectives by the percentage of target met 
in comparison to baseline figures. The report concludes with lessons and recommendations.
3
CONTENTS
List of Figures .............................................................................................................................v 
List of Tables ............................................................................................................................. vi 
Acronyms.................................................................................................................................. vii 
Executive Summary ................................................................................................................... 1
Background ............................................................................................................................ 1
Methodology........................................................................................................................... 1
Results ................................................................................................................................... 2
1 Introduction ............................................................................................................................. 5
1.1 Guatemala Context........................................................................................................... 5
1.2 Program Intervention Logic............................................................................................... 6
2 Endline Evaluation .................................................................................................................. 7
2.1 Purpose and Scope .......................................................................................................... 7
2.2 Methods ........................................................................................................................... 7
2.2.1 Study Design.............................................................................................................. 7
2.3 Quantitative Surveys ........................................................................................................ 8
2.3.1 Sampling.................................................................................................................... 8
2.3.2 Data Collection........................................................................................................... 9
2.3.3 Monitoring and Evaluation Framework ......................................................................10
2.3.4 Analysis ....................................................................................................................10
2.4 Qualitative Fieldwork .......................................................................................................11
2.4.1 Sampling...................................................................................................................12
2.4.2 Data Collection..........................................................................................................12
2.4.3 Analysis ....................................................................................................................13
2.5 Limitations .......................................................................................................................13
3 Findings .................................................................................................................................14
3.1 Program Description........................................................................................................14
3.1.1 Key Baseline Findings...............................................................................................14
3.1.2 Program Design ........................................................................................................15
3.2 Program Research ..........................................................................................................21
3.3 Program Resources.........................................................................................................22
3.4 Program Participation ......................................................................................................23
3.5 Program Monitoring .........................................................................................................24
3.6 Program Sustainability.....................................................................................................24
3.7 Program Results..............................................................................................................25
3.7.1 IPTT..........................................................................................................................25
4
3.7.2 Discussion of Results................................................................................................30
3.7.3 Trigger Indicators ......................................................................................................51
4 Conclusions, Lessons, and Recommendations......................................................................53
4.1 Conclusions.....................................................................................................................53
4.2 Lessons and Recommendations......................................................................................56
References ...............................................................................................................................58
Appendices ...............................................................................................................................60
5
LIST OF FIGURES
Figure 1: Key sampling and sample size process flow ............................................................... 9
Figure 2: Number of rations distributed .....................................................................................19
Figure 3: Participation rates at PROCOMIDA training events (%) .............................................20
Figure 4: Amount of voluntary contributions and use per year...................................................23
Figure 5: Household hunger scale at baseline and endline .......................................................33
Figure 6: Average HDDS at baseline and endline .....................................................................34
Figure 7: Increased mother knowledge of child nutrition at baseline and endline ......................36
Figure 8: Mother knowledge on child nutrition at baseline and endline......................................36
Figure 9: Mother knowledge of danger signs during pregnancy at baseline and endline ...........36
Figure 10: Mother knowledge of danger signs of childhood illnesses at baseline and endline ...38
Figure 11: Mothers receiving prenatal care assistance at baseline and endline ........................39
Figure 12: Mothers receiving minimum recommended antenatal and postnatal care at baseline 
and endline ...............................................................................................................................39
Figure 13: Institutions visited for prenatal care at baseline and endline.....................................40
Figure 14: Place of delivery and assisting staff during delivery at baseline and endline ............40
Figure 15: Prevalence of children immunized and measured for growth at baseline and endline
.................................................................................................................................................41
Figure 16: Essential newborn care at baseline and endline.......................................................42
Figure 17: Children receiving minimum treatment for respiratory problems and diarrhea at 
baseline and endline .................................................................................................................42
Figure 18: Breastfeeding practices among mothers at baseline and endline.............................43
Figure 19: Percentage of children 0-6 months exclusively breastfed at baseline and endline....44
Figure 20: Dietary diversity in children 6-24 months at baseline and endline............................45
Figure 21: Complementary feeding practices at baseline and endline.......................................45
Figure 22: Percentage of children with measurements classified as wasting, stunting, and 
underweight, by age at baseline and endline ............................................................................47
Figure 23: Availability of basic services at CHCs at baseline and endline .................................49
Figure 24: CHC staff knowledge regarding danger signs and ORS use at baseline and endline
.................................................................................................................................................49
Figure 25: CHC staff knowledge of child feeding practices at baseline and endline ..................49
Figure 26: Availability of medical supplies at CHCs at baseline and endline .............................51
6
LIST OF TABLES
Table 1: Description and values of parameters used in the sample size calculation................... 9
Table 2: Distribution of communities and interviewed households per district ............................ 9
Table 3: Evaluation criteria for the exclusion of anthropometric outliers ....................................11
Table 4: Total valid observations in final databases ..................................................................11
Table 5: Sex and literacy levels of FGD participants per community (# and %).........................12
Table 6: Results framework.......................................................................................................16
Table 7: Distribution of monthly rations .....................................................................................18
Table 8: Summary of food aid program resources in US$ .........................................................22
Table 9: Amount and use of financial contributions from participants to PROCOMIDA..............23
Table 10: IPTT – baseline, target, endline, and percent achievement .......................................26
Table 11: Background characteristics of surveyed households, mothers, and children, at baseline 
and endline ...............................................................................................................................32
Table 12: Nutritional status of children at baseline and endline .................................................47
Table 13: Distribution of stunting among children 6-59 months (<-2 z-score height-for-age), by 
sex, at baseline and endline......................................................................................................47
Table 14: Adverse conditions and events affecting Alta Verapaz at baseline and endline .........52
Table 15: Evaluation ratings......................................................................................................53
vii
ACRONYMS
ACF Action Against Hunger
BCC Behavior Change Communication
BL Baseline
BMI Body Mass Index
CC Convergence Center
CDC Centers for Disease Control and Prevention
CHC Community Health Commission
CHV Community Health Volunteer
CHW Community Health Workers
COCODE Consejo Comunitario de Desarrollo / Community Development Council
CRS Catholic Relief Services
CSB Corn-soy Blend
CSI Coping Strategy Index
DFAP Development Food Assistance Program
EL Endline
ENA Emergency Nutrition Assessment
ENSMI Encuesta Nacional de Salud Materno Infantil / National Survey of Maternal 
and Child Health
FAFSA Food Aid and Food Security Assessment
FANTA Food and Nutrition Technical Assistance
FEWSnet Famine Early Warning System Network
FFP Food for Peace
FGD Focus Group Discussion
FY Fiscal Year
GDP Gross Domestic Product
H-A Height-for-age
HAZ Height-for-age z-score
HDDS Household Diet Diversity Scale
HFIAS Household Food Insecurity Access Scale
HHS Household Hunger Scale
HP Health Post
IFPRI International Food Policy Research Institute
IPTT Indicator Performance Tracking Table
IR Intermediate Result
IY Implementation Year
IYCF Infant and Young Child Feeding
KG Kilogram
KII Key Informant Interview
KM Kilometers
LNS Lipid-based Nutritional Supplements
M&E Monitoring and Evaluation
MCU Mother-Child Unit
MCHN Maternal Child Health and Nutrition
MIS Management Information System
MNP Micronutrient Powder (Chispitas)
MOH Ministry of Health and Social Assistance
MT Metric Ton
MYAP Multi-year Assistance Program
N Sample Size
NGO Non-governmental Organization
OECD Organization for Economic Cooperation and Development
ORS Oral Rehydration Salts
PEC Programa de Extensión de Cobertura (Ministry of Health’s Extension of
Coverage Program)
PM2A Preventing Malnutrition in Children Under 2 Approach
PPS Probability Proportional to Size
PREP Pipeline and Resource Estimate Proposal
PROCOMIDA Programa Comunitario Materno Infantil de Diversificación Alimentaria /
Community Food Diversification Program for Mother and Child
Q Quartiles
SAM Severe Acute Malnutrition
SBC Social and Behavior Change
SD Standard Deviation
SESAN Secretaria de Seguridad Alimentaria y Nutricional / Secretariat of Food and
Nutritional Security
SIAS Sistema Integral de Atención en Salud / Integrated Health Attention System
SO Strategic Objective
SOW Scope of Work
SPSS Statistical Package for Social Sciences
STATA Data Analysis and Statistical Software (branch name)
USAID United States Agency for International Development
WAZ Weight-for-age z-score
WHO World Health Organization
WHZ Weight-for-height z-score
Y Year
viii
1
EXECUTIVE SUMMARY
Background
Guatemala has some of the highest rates of chronic malnutrition in Latin America, and in the 
world. Conditions are historically most severe in the northern lowland and highland areas 
converging in and around Alta Verapaz. Within these geographic areas, the most vulnerable and 
affected by food insecurity and malnutrition are i) pregnant and lactating women, and ii) children 
under two years of age. This is particularly true amongst poor, marginalized, and vulnerable 
indigenous and ethnic communities, including the Q’eqchí.
Programa Communitario Materno Infantil de Diversificación Alimentaria (Community Food 
Diversification Program for Mother and Child/PROCOMIDA) was a six-year program1
funded by 
the United States Agency for International Development (USAID) through the Bureau of 
Democracy, Conflict, and Humanitarian Assistance Office of Food for Peace (FFP), and 
implemented by Mercy Corps/Guatemala. PROCOMIDA began in July 2009, and field 
implementation (which began June 2010) concluded in May 2015. PROCOMIDA sought to 
improve the nutritional status of direct and indirect participants in 936 vulnerable communities in 
the Alta Verapaz Department, Guatemala. PROCOMIDA directly reached 60,605 mother-child 
units
2
in 1,180 vulnerable communities. The total number of direct and indirect beneficiaries was
266,000 community and family members, with an average family size of 5.3.
PROCOMIDA worked with non-governmental organizations (NGO) and Ministry of Health and 
Social Assistance (MOH3
) units to strengthen community-level health structures (convergence 
centers/CC) to improve culturally and technically appropriate health service provision, and to 
stimulate demand for these services. It provided participants at 356 CCs with monthly training 
sessions, health and nutrition counseling, and additional household visits combined with the 
provision of one of five different types of food rations.
Methodology
This final independent evaluation coincided with the conclusion of PROCOMIDA’s field 
implementation (service and food delivery) on 31 May 2015. The purpose was to evaluate the 
performance of PROCOMIDA in targeted populations, assessing for relevance, effectiveness, 
impact, and sustainability. This implies evaluating the program’s achievements in meeting 
objectives and targets set against baseline values. We intend lessons and recommendations for 
improvement of future Title II Development Food Assistance Programs and similar activities.
The evaluation design was steered by general evaluation criteria from the Organization for 
Economic Cooperation and Development (OECD) Development Assistance Committee, 
guidelines for the evaluation of Preventing Malnutrition in Children under Two Years of Age 
Approach (PM2A) interventions, by USAID, and by the Food and Nutrition Technical Assistance 
(FANTA) project.
4
Field data collection occurred between April - July 2015. Activities included: i) a document review; 
ii) quantitative surveys; iii) anthropometric measurements; and iv) qualitative fieldwork. Analysis 
occurred in fall 2015, with an initial evaluation submitted in November 2015, and a re-drafted final 
evaluation report submitted in March 2017.
1 PROCOMIDA was originally defined as a ‘project.’ Given the scope and complexity of the intervention, it has the characteristics of a program,
and thus this report refers to PROCOMIDA as a ‘program.’
2 PROCOMIDA used a measurement unit of ‘Mother-Child Unit’ (MCU), which uniquely identified the beneficiary through the 1000 days
approach (pregnant mother, lactating mother, child six to 24 months).
3 Local acronym is MSPAS.
4 The program implementation started during the FANTA II project and ended during the FANTA III project. Hence, in this report both project
will be referred to as ‘FANTA project’.
2
Results
PROCOMIDA improved the nutritional status and health of women and children vulnerable to food 
insecurity in northern Guatemala (program goal).
PROCOMIDA’s two strategic objectives were:
SO.1: Pregnant and lactating women, children under two, and malnourished children under five 
in program areas have improved and sustainable health and nutrition status
Activities were appropriate, technically sound, and efficiently implemented. Activities and 
results materialized as expected, primarily operating through consistent nutritional 
supplementation of children and their mothers, along with increased dietary diversification 
among mother-child units (MCU). Core nutritional outcomes and their level of achievement 
against target include:
 Indicator 1: % of children 0-59 months underweight – baseline 13.3%; endline 14.0% 
(71% achievement) – not statistically significant, p=0.2570.
 Indicator 2: % of children 6-59 months stunted – baseline 59.2%; endline 53.5% (99%
achievement) – statistically significant, p=0.002.
 Indicator 3: Average Household Dietary Diversity Score (HDDS) – baseline 7.19; endline
7.26 (91% achievement) – not statistically significant, p=0.2899.
 Indicator 5: % of children 6-24 months with minimum dietary diversity – baseline 60.1%;
endline 71.6% (90% achievement) – statistically significant, p<0.0001.
 Indicator 7: % of caregivers demonstrating increased nutritional knowledge – baseline
0%; endline 52.6% (75% achievement)
 Indicator 9: % of children 0-6 months who were exclusively breastfed – baseline 65.1%;
endline 75.7% (89% achievement) – statistically significant, p=0.0411.
SO.2: Healthcare service providers at community through municipal levels have improved 
service quality and delivery
PROCOMIDA’s activities focused on strengthening the quality and availability of primary 
healthcare services in rural areas through a model of NGO service delivery through the MOH 
Programa de Extensión de Cobertura (Extension of Coverage Program/PEC). Communities 
benefitted from strengthened local structures for self-management (establishment of health 
commissions and emergency funds; developed local health planning capacity) and improved 
infrastructure (store rooms, primary healthcare centers).
Intervention success depended largely on the PEC’s budget and MOH provisions, which were 
outside of PROCOMIDA’s control. Issues of local and national governance interrupted primary 
healthcare service delivery during implementation periods, limiting the program’s ability to 
achieve outcomes, and exposing the fragility of the PEC model.
The themes addressed under these two objectives were highly relevant, despite activities being 
limited to those pre-determined by the donor.
Despite challenges, PROCOMIDA managed efficient product and service delivery in a complex 
geo-cultural environment, benefitting thousands of community members. While program 
participation was voluntary, 88.9% of eligible mothers and children benefitted from their 
engagement in PROCOMIDA during the implementation period. Key achievements include:
Improved nutrition. PROCOMIDA presented an opportunity for participating children to improve 
their future potential. Econometric outcome assessment and undernutrition trends showed high 
long-term impact for children who participated, with positive results from nutritional improvements 
felt well beyond program graduation, specifically:
3
 Improved maternal nutrition contributing to the improved growth and development of their 
children – over 34,000 pregnant and lactating women received food rations
 Improved and consistent access to diverse and nutritious foods for children 6-24 months
– 71.6% obtained a minimum dietary diversity; 57,590 received food rations
 Reduced levels of stunting in children under five years, with enduring benefits through 
their lives
The difference between participant and non-participant results is statistically significant. 
Households that participated in the program did so for a range of six to 30 months. Graduated
households demonstrated increased efforts to reduce their own vulnerability to food insecurity:
maintaining a diverse diet, improving household environments, and adopting behaviors related to 
improved infant and young child feeding (IYCF) and caring practices. Participating families 
expressed a desire for even further information on nutrition, childcare, illness warning signs, and 
family planning. They demonstrated an eagerness to continue using knowledge gained to benefit 
their current and future children.
PROCOMIDA concentrated on support (skills, knowledge, resources, structures) at the local level 
and within existing health systems, and deliberately promoted the adoption of actionable 
behaviors to address barriers. While there is strong probability that nutritional gains for children 
under two years who participated in the program will endure, resulting in healthier and more 
productive adults, there is only moderate likelihood that: i) mothers and families will function in 
environments enabling application of program knowledge and skills (health behavior and 
capacities) to effectively protect future pregnancies and early childhood development; and ii) the 
public health sector will be able to continue program improvements given existing structural 
barriers (current health system, skills, resources). Increased food consumption ended with 
program graduation, and improved knowledge does not guarantee the adoption of desired 
behaviors, particularly in areas with extreme poverty and income insecurity. A stronger focus on 
community structures and exit strategies, that strengthened and complemented existing health 
structures may have compensated for this.
Decreased stunting. PROCOMIDA achieved a decrease in annual stunting of 1.14 percentage 
points per year, a considerable improvement when compared against Alta Verapaz national 
survey results from 1999-2009 showing an average decrease of only 0.17%/year,5 and an overall 
national decrease of 0.55%/year during PROCOMIDA’s implementation period in survey results 
from 2014-2015.
6
A 2014 national survey conducted by SESAN for the Zero Hunger Pact in 166 prioritized 
municipalities (including all PROCOMIDA municipalities) found 60.6% stunting (H/A, rural areas
65.4%), 1.3% wasting (W/H, rural areas 1.4%), and 14.3% underweight (W/A, rural areas 15.6%)
in children three to 59 months of age.
7 Although the latter age range is slightly different, the 
variance in stunting levels is indicative. When compared to the national and regional levels,
PROCOMIDA was able to double the national annual stunting decrease. The average reduction
for Latin American is 1.54%/year, and globally it is 1.32%/year.
8
5 MSPAS (2011). Encuesta Nacional de Salud Materno Infantil 2008 (ENSMI-2008/09). Ministerio de Salud Pública y Asistencia Social
(MSPAS)/Instituto Nacional de Estadística (INE)/Centros de Control y Prevención de Enfermedades (CDC). Guatemala
6 MSPAS (2015). VI Encuesta Nacional de Salud Materno Infantil (ENSMI-2014/15). MSPAS/INE/Secretaría de Planificación y Programación 
de la Presidencia (Segeplán). Guatemala.
7 SESAN (2015). Tercera Encuesta de Monitoreo en los 166 municipios priorizados; Tomo I: Estado Nutricional de Menores de Cinco Años y
Mujeres en Edad Fertil. Guatemala.
8 van Haeften, R, et.al. (2013). Second Food Aid and Food Security Assessment (FAFSA-2). Washington, DC: FHI 360/FANTA.
4
Healthy Timing and Spacing of Pregnancies. PROCOMIDA contributed to improved healthy 
pregnancy timing and birth spacing among program participants. Household survey data 
measured this, and sources that participated in qualitative fieldwork confirmed it. Coordinated and 
simultaneous interventions of other organizations (family planning, contraception distribution) 
complemented PROCOMIDA’s family planning activities in implementation areas.
Gender empowerment. As a crosscutting programmatic theme, gender results show greater 
female participation, as learned primarily from qualitative fieldwork. In particular, women were 
more involved in decision-making around health and nutrition. Increases were noted in mother’s 
knowledge on child nutrition, danger signs of child illness, and danger signs during pregnancy. 
The development of home gardens and the overall increased role in food production enabled 
them to control dietary diversification for their families. Pertaining to gender equity, results were 
weaker. The almost exclusive concentration on the participant MCU limited involvement of other 
family members, including fathers, in-laws, and siblings. This was a lost opportunity to reinforce 
the drivers of change, only realized in the late stages of implementation when the program did 
initiate work with men and key family members including mothers-in-law.
Strengthened community arrangements. Communities benefitted from strengthened 
community-based structures (health commission, emergency fund, health planning mechanisms) 
and improved infrastructure (program store rooms, health centers). It is unclear how well new first 
level health services will engage with community arrangements strengthened by the program. 
Community leaders intend to maintain and utilize infrastructure, co-opting them to serve as 
community centers with an aim of furthering programmatic messages and behaviors.
5
1 INTRODUCTION
Food insecurity disproportionately affects the poor, indigenous, smallholder farmers, and landless 
populations in rural communities. In Latin America, including in Guatemala, food insecurity results 
in high rates of chronic child malnutrition. Programa Communitario Materno Infantil de 
Diversificación Alimentaria (Community Food Diversification Program for Mother and Child, or 
PROCOMIDA), a six-year program (2009 – 2015) funded by USAID through FFP, and 
implemented by Mercy Corps/Guatemala, aimed to improve the nutritional status and health of 
women and children vulnerable to food insecurity in northern Guatemala.
1.1 Guatemala Context
Guatemala presents with some of the worst rates of chronic malnutrition in the world. Despite a 
strong institutional framework and established coordination structures supporting food security 
programming, Guatemala has the highest prevalence of stunting (49.8%) in children under five in 
Latin America, and among the worst rates of chronic malnutrition globally.
9 Conditions are most 
severe in the western highlands and the northern lowland and highland areas converging in and 
around the Alta Verapaz Department. Within this geographic area, the most vulnerable and 
affected by food insecurity and malnutrition are pregnant and lactating women, and children under 
two years of age, exacerbated amongst poor, marginalized, and vulnerable indigenous and ethnic 
communities, including the Q’eqchí.
10
In Alta Verapaz, malnutrition is especially harmful to 
children under two as they can experience lifetime impairment due to weak physical and cognitive 
development. Undernutrition causes 45% of child deaths globally.
11,12 In Alta Verapaz, 60% of 
children under five are chronically malnourished, while the infant mortality rate for 2014-2015 was
29 per 1,000 live births, putting Alta Verapaz at a ranking equivalent of 131
st (from best to worst)
out of 195 countries.13 Overall, the maternal mortality ratio in Guatemala was 140 per 100,000, 
ranking equivalent to 122
nd out of 184 countries.
14,15
Evidence suggests that investments in nutrition improve national productivity and economic 
performance (gross domestic product/GDP per capita) by countering direct productivity losses, 
losses via poorer cognition, and losses via reduced schooling.
16 Despite macroeconomic stability 
and 6.5% GDP growth in Guatemala between January 2010 and December 2014, fiscal health 
structures remained unchanged, and poor and marginalized populations underserved.
17
Populations in Alta Verapaz largely pursue subsistence agriculture (corn and beans) with some 
cash crops (coffee, cocoa, cardamom), which suffered from pest-related diseases during periods 
of PROCOMIDA implementation. According to the coping strategy index (CSI),
18 by program end,
9 FAO (2014). Guatemala Country Fact Sheet on Food and Agriculture Policy Trends: Socio-economic Context and Role of Agriculture. 
http://www.fao.org/3/a-i4124e.pdf.
10 FEWSnet Guatemala Country Profile Website; FAO Socioeconomic Studies of Vulnerable Groups in Guatemala.
11 WHO (2016). Media Center: Children: Reducing Mortality Fact Sheet. http://www.who.int/mediacentre/factsheets/fs178/en/.
12 Black, R.E. et al., (2013). Maternal and child undernutrition and overweight in low-income and middle-income countries. The Lancet 382 
(9890) pp. 427-451.
13 Murphy, L (2010). Mercy Corps and USAID mark one year of malnutrition prevention program in Guatemala. Guatemala. 
https://www.mercycorps.org/press-room/releases/mercy-corps-and-usaid-mark-one-year-malnutrition-prevention-program-guatemala
14 MSPAS (2015). VI Encuesta Nacional de Salud Materno Infantil (ENSMI-2014/15). MSPAS/INE/Secretaría de Planificación y Programación 
de la Presidencia (Segeplán). Guatemala.
15 World Bank Open Data. http://data.worldbank.org/indicator/SH.STA.MMRT and http://data.worldbank.org/indicator/SP.DYN.IMRT.IN
16 Black, R.E. et al., (2013). Maternal and child undernutrition and overweight in low-income and middle-income countries. The Lancet 382 
(9890) pp. 427-451.
17 Cabrera, M., N. Lustig and H. Morán, Fiscal Policy, Inequality and Ethnic Divide in Guatemala CGD Working Paper No. 397, March 2015, 
page 22 - 23. http://www.cgdev.org/sites/default/files/CGD-Working-Paper-397-Cabrera-Lustig-Moran-Fiscal-Policy-Inequality-Ethnic-Divide￾Guatemala.pdf.
18 The CSI enumerates all consumption-related coping strategies commonly used by a population. It enumerates both the frequency and 
severity of coping strategies of households faced with short-term insufficiency of food. Hence, the CSI is a measure of food insecurity. The 
higher the score, the greater the food insecurity (see http://www.fsnnetwork.org/coping-strategies-index-field-methods-manual). Following the
6
populations were facing five times more difficulty and stress securing short-term food security 
than at program start.
1.2 Program Intervention Logic
PROCOMIDA is part of FANTA Project research on Preventing Malnutrition in Children under 2 
(PM2A) – a food-assisted approach to reducing the prevalence of child malnutrition in children 
under two years by targeting a health and nutrition intervention package to all pregnant women, 
lactating mothers of children from birth to six months, and children under two years, in food￾insecure areas, regardless of nutritional status. PM2A is built upon the premise that these women 
and children are the most nutritionally vulnerable members of the population, and that the ‘first
1,000 days’ represent the greatest potential for growth and development. Actions to increase 
household knowledge of nutrition and care, and strengthen health services, further bolster the 
universal approach of PROCOMIDA.
To validate the PM2A approach, USAID’s FFP funded two parallel research Title II programs, in 
Burundi and Guatemala. They chose Mercy Corps to implement PROCOMIDA in Guatemala as 
a multi-year assistance program (MYAP).
Mercy Corps began working globally in 1979 to alleviate suffering, poverty, and oppression by 
helping people build secure, productive, and just communities. In the mid-1980s, Mercy Corps 
began operations in Central America, and in 2001 expanded programming to northern 
Guatemala, targeting poor areas in Alta Verapaz. Mercy Corps pursued PROCOMIDA at a time 
when more than 60% of children under five in Alta Verapaz were chronically malnourished. 
Program implementation initially covered the period July 2009 – June 2014. USAID officially 
extended implementation to May 2015 through an agreement modification, signed June 2012 (the 
final modification covered through November 2015).
rationale of the CSI, which by definition does not have a fixed maximum level (defining levels of low, moderate or severe food insecurity are 
context specific), the initial value measured for Alta Verapaz can be considered as low, while the final average level of 40 points would be 
considered to be at least at the upper level of a moderate food insecurity.
7
2 ENDLINE EVALUATION
2.1 Purpose and Scope
The purpose of this final PROCOMIDA evaluation is to assess impact and sustainability potential. 
Specifically, it seeks to determine: i) results; ii) effectiveness and relevance of activities and 
implementation approaches; and iii) extent crosscutting themes (community participation, gender, 
and environment)
19 were integrated (see scope of work/SOW, Annex 1, A). Towards this, 
performance ratings are provided to compare achievements against the results framework (goal, 
objectives, performance, impact indicators: see Table 10). Achieved values are compared with
targets and baseline findings to determine to what extent USAID-supported interventions
contributed to reducing malnutrition and related health conditions. This evaluation further
establishes plausible links between activities, outputs, and outcomes, identifying lessons and 
recommendations for the improvement of future Title II Development Food Assistance Programs
(DFAP) and similar activities.
General evaluation criteria of the OECD Development Assistance Committee, guidelines for the 
evaluation of PM2A interventions, and guidance from USAID and FANTA inclusive of rigorous 
statistical analysis, conclusions, and recommendations, led the evaluation approach. It covers the 
period July 2009 – May 2015.
The SOW for re-drafting the final evaluation report (Annex 1, B) specified the need to: i) review 
endline data for statistical relevance in comparison with baseline results; ii) integrate reviewer 
comments; and iii) reorganize and edit the final report to comply with industry standards.
2.2 Methods
2.2.1 Study Design
Comparability between the 2010 baseline and 2015 endline samples was assured by: i) 
comparable sampling frameworks and criteria; ii) data collection techniques that employed 
triangulation and stakeholder participation; iii) identical questionnaires; and iv) survey 
implementation at the same time/season and in the same geographic areas.
20 Note that the 
midterm evaluation and annual M&E surveys are not comparable with the baseline or endline, as 
their sampling frameworks were participant-based rather than population-based.
Final evaluation activities covered two phases, over the period April – November 2015.
Phase 1: Data Collection (April – July 2015) was conducted by a consortium between ACF 
(Action Against Hunger) and ECI Consultores, international and local consultant firms whose 
combined technical study team included eight members with backgrounds in economics, nutrition, 
management and public policy, social science, and knowledge management, and a further 6 field 
teams of 5 people each. Field team members spoke Spanish. Local translators assisted them 
where local Q’eqchí language was required. A team of local enumerators, all native Q’eqchí
speakers, translated surveys from Spanish to Q’eqchí, and collected field data. The study team 
trained all enumerators in using tablets for data entry, and in using anthropometry equipment.
19 BCC is generally considered another crosscutting theme for Title II programs. However, BCC was a core PROCOMIDA component.
20 This refers to the original intervention areas and does not cover expansion sites.
8
ACF/ECI study team members, with Mercy Corps staff, developed quantitative endline data 
collection tools based on baseline tools, and designed by the International Food Policy Research 
Institute (IFPRI), to answer evaluation questions (see Annex 5, 6, 7). All field team members 
attended a two-week long training workshop to: i) become acquainted with the tools; ii) ensure 
uniform field application; iii) understand the study rationale and protocol; iv) receive advice on 
data collection procedures; v) practice conducting anthropometric measurement standardization 
tests with 12 voluntary children and their mothers; and vi) review digital data collection procedures 
using electronic tablets.
21 Data collection activities consisted of:
Document review
Quantitative field surveys
 Household (participant level)
 Knowledge survey (health workers, volunteers, health center – community level)
 Health center (convergence center – cluster level)
Anthropometric measurements
 Children (0-59 months)
 Mothers (caregiver) of same children
Qualitative fieldwork
 In-depth key informant interviews (KII)
 Focus group discussions (FGD)
 Direct observation (site visits)
Phase 2: Data Cleansing, Analysis, and Reporting (August – November 2015) were completed 
by ACF/ECI, and reanalysis completed by Health and Development Consultants International, 
LLC (October 2016). The HDCi team included one global public health specialist, and one 
epidemiologist/biostatistician tasked with redrafting the final external evaluation report.
2.3 Quantitative Surveys
The endline household survey was a cross-sectional study, comparable with the 2010 baseline.
2.3.1 Sampling
Since the household survey aims to show results at the level of the entire population, the sampling 
frame was comprised of all households with children between zero and 59 months of age in March
2015, within selected clusters. The study team defined clusters using probability proportional to 
population methods equivalent to nearly one CC per cluster. They based the geographical
sampling frame on the same 100 clusters used during the baseline.
22 Following a census 
conducted by PROCOMIDA staff within these 100 clusters, they defined the total sample size as
12,844 households
23 having children within the inclusion criteria age range.
A sampling frame of 46 CCs (inter-cluster correlation = 0.01) were selected randomly using the 
probability proportional to size (PPS) algorithm (described in FANTA-2 guidelines)
24 out of the
100 clusters, taking into consideration that two CCs were larger communities that covered two 
clusters (thus, 46 CCs covered 48 clusters). The sample additionally represented with high fidelity 
the true geographic distribution of PROCOMIDA participants and non-participants.
Standard statistical software (ENA
25
for cluster selection and STATA
26
for household selection) 
ensured that each household had the same probability of being selected. For a number of 
indicators, the survey only needed to interview a sub-population and not the entire sample of
21 Cybertech 8170 Tablets were used, and the operating system was Java for Android 2.2.
22 Refers to baseline survey stratum A to E. Baseline control groups and any alternative control groups were not included.
23 Census data revealed that non-participation was found to be 10.8% of households in the 100 clusters.
24 USAID (2009). Food and Nutrition Technical Assistance II Project-Annual Program Report Project Year One. June 2, 2008 – 30 Sept. 2009.
25 Emergency Nutrition Assessment for SMART, see www.nutrisurvey.net/ena/ena.html.
26 Data Analysis and Statistical Software (branch name).
9
households. The sampling framework was designed to produce statistically significant results for 
all sub-populations, specifically children between six and 24 months, and their mothers.
The change in population prevalence of malnutrition was the core indicator used in the baseline 
study and hence determined the sample size requirement for the endline. In accordance with
FANTA guidelines,
27
the study team used following equation (for follow-up surveys):
n = D[(Zα + Zβ)² * (P1(1-P1) + P2(1-P2)) / (P2 – P1)²]
Table 1: Description and values of parameters used in the sample size calculation
Variable Description Value
D Design effect (default value) 2
P1 Baseline prevalence (baseline) 0.566
P2 Endline expected prevalence (target) 0.504a
Zα Z-score of the degree of confidence to ensure that the change occurred because of the 1.645 
intervention (statistical significance
Zβ Z-scored of the degree of confidence to ensure that the change has actually occurred 0.84 
(statistical power)
n Sample size (number of valid household interviews required) 1,592
a The expected decrease in malnutrition prevalence was 0.062. This was accurate with Mercy Corps.
In calculating the minimum required sample size, the Type 1 error (α) was set at 0.05, and the 
power of study (1-β) at 0.80.
Figure 1: Key sampling and sample size process flow
1,592
Minimum required sample size (of 
households) based on calculation
Actual number of households 
sampled
a
No. of households interviewed
(99 non-response, 5%)
Final valid household dataset
(23 with data issues, excluded)
1,843
1,744
1,721
1,830
1,850
Minimum required, accounting for
15% potential non-response
Sampling strategy: 37 households x
50 clusters
a One cluster only contained 30 households compared to the targeted 37, corresponding to a smaller community.
2.3.2 Data Collection
Quantitative survey fieldwork occurred over several weeks between April 21 - May 18 in the 
municipalities of Cahabón, Cobán, Lanquín, and San Pedro Carchá, of the Alta Verapaz 
Department. All survey data was collected through a java-based digital application for Android on 
tablets, originally developed by Mercy Corps Staff,
28 and continuously crosschecked for accuracy. 
Field team surveyors sent survey and anthropometric data directly to field editors who used ENA 
software to check for consistency, and who returned Z values to the surveyors. Z values were 
double-checked during data analysis.
Table 2: Distribution of communities and interviewed households per district
District # of selected clusters # of selected CCs # of households sampled
Cahabón 5 5 175
Cobán 11 10 376
27 USAID (1999). Fanta Sampling Guide. http://pdf.usaid.gov/pdf_docs/Pnacg172.pdf.
28 Original applications were designed for annual monitoring and for the Midterm Evaluation. Baseline data collection was held on a PDA
platform through Mobile Data.
10
Lanquín 3 3 105
San Pedro Carchá 31 30 1088
TOTAL 50 48a 1,744b,c
a Two CCs did not have any staff available for interviews, thus final sample size for the CC survey was 46.
b Datasets from 23 households were incomplete, thus final sample size for household analysis was 1,721.
c Four households had 2 mothers among the members, thus final number of mothers sampled was 1,748.
The field teams also collected survey data from sampled healthcare facility staff, and community 
health workers (CHW) and volunteers (CHV) – from the same CCs as the household surveys.29
2.3.3 Monitoring and Evaluation Framework
All surveys and information collected addressed the following topics (see Annex 5, 6, and 7 for 
more details):
Household survey
 Household composition
 Dietary diversity (HDDS and Household Food Insecurity Access Scale/HFIAS)
 Mother’s knowledge
 Utilization of antenatal and post-natal care
 Infant and young child feeding (IYCF practices)
 Vaccination and growth monitoring
 Field and household management plan
 Annual monitoring (CSI)
Anthropometric measurements
 Mother (or caregiver) of 0-59 month old children (only one per household)
 Children 0-59 months old (only children of the mother/caregiver)30
Knowledge survey of healthcare staff (CC level)
 Education and basic knowledge
 Pregnancy and complication
 Antenatal and postnatal care
 Maternal nutrition
 Infant and Young Child Feeding and caring practices
Health center survey
 Responder information
 Medicines and materials
 Administration
 Care services provided
The study team did not conduct CC equipment surveys, which were part of the baseline. Mercy 
Corps/Guatemala negotiated the elimination of this indicator in the Fiscal Year (FY) 2014 pipeline 
and resource estimate proposal (PREP). USAID approved this in Modification 10.
2.3.4 Analysis
The study team conducted initial data analysis in fall 2015 using Statistical Package for Social 
Sciences (SPSS). The HDCi team re-verified findings in fall 2016. Data and documentation were 
available in the original Spanish, and in some cases English translations. The review and analysis 
process included the following:
 Comparison of IPTT data entry
 Review of SPSS datasets, and cross-checking these against survey questionnaires
29 The healthcare center survey could not be carried out in Chinaichab, district of Cobán (HC code 73) and San Antonio I, district of San Pedro 
Carcha (HC code 280). The CHW was identified as an interview partner for the community healthcare center surveys. In both of these 
communities, the CHW was absent during the days of field visit in the respective communities. The information recollected for the final survey 
in 46 community health centers allows for full comparability with the baseline study of 45 community health centers.
30 Pregnant women were excluded from anthropometry.
11
 Translation Spanish-English, where required
 Review and re-running of the previous SPSS syntax files, where possible
 Finalizing IPTT indicator list and presentation
 Harmonizing any conflicting findings from draft reports
 Generating correlation graphs for selected indicators
 Running adding additional summary tables and data analysis, where relevant
For anthropometric data, the study team used the World Health Organization’s (WHO) Growth
Standard – 2006 31 data cleaning criteria.
Table 3: Evaluation criteria for the exclusion of anthropometric outliers
32
Dimension of under nutrition Exclusion Criteria
Wasting [-5 SD; +5 SD]
Stunting [-6 SD; +6 SD]
Underweight [-6 SD; +5 SD]
For household survey variables, the study team applied usual data consistency controls,
33
providing a final and validated number of observations.
Table 4: Total valid observations in final databases
Sampling population and unit sub-strata Observations Minimum required observations
Households with children 0-59 months 1,721 1,592
Children 0-6 months 189 176
Children 0-23 months 924 892
Children 6-23 months 735 694
Children 6-59 months 2,211 2,144
Children 0-59 months 2,406 2,365
CC’s (equipment and services available) 46 45
CC’s (health staff knowledge) 46 45
The results contain descriptive and econometric analysis of endline survey data in comparison 
with the baseline. Towards providing full methodological comparability
34 between baseline and 
endline surveys, all baseline data were re-estimated, with the following differences observed:35
 The correct baseline value for I3 is 7.19 (was reported as 5.8 in 2010)
 The correct baseline value for I13 is 50.0 (was reported as 40.5 in 2010)
 The correct baseline value for I17 is 73.2 (was reported as 85.9 in 2010)
PROCOMIDA discussed these differences with IFPRI (baseline implementer), and presented 
their agreement to USAID/FFP on October 8, 2015, after which IFPRI published an addendum to 
their baseline report.
Throughout the analysis and presentation of findings, the team compared average values 
between two groups using the t-test. When analyzing the relationship between two categorical 
variables, they used the chi-square test. For the analysis of the percentage change among linked 
or paired data, they used the McNemar test. In all statistical analyses, a 95% significance level 
and a p-value of less than 0.05 indicates a statistically significant association.
2.4 Qualitative Fieldwork
As per requirements, the study team complemented quantitative data analysis with qualitative 
field research for more comprehensive understanding of overall findings. Qualitative research
31 WHO (2006). WHO Child Growth Standards, Geneva. http://www.who.int/childgrowth/standards/Technical_report.pdf?ua=1.
32 Ibid.
33 Error detection through consistency checking using the consistency matrix tool.
34 In the sense of having used exactly the same definitions and steps for computing variables.
35 The evaluation team shared and discussed these findings intensively with both Mercy Corps and IFPRI (who conducted the baseline survey). 
IFPRI accepted the revised values and issued an errata for the baseline report.
12
allowed for triangulation and interpretive perspectives regarding PROCOMIDA, and the 
sociocultural and economic contexts of implementation geographies. The study team defined 
semi-structured key informant interview (KII) questionnaires and a thematic focus group 
discussion (FGD) list (see Annex 7) after reviewing the preliminary findings from the three surveys 
(household, health center services, and health center staff knowledge).
2.4.1 Sampling
The study team identified 14 communities (see Annex 2) for qualitative research, from among the
46 CCs identified during quantitative surveys. They applied the following selection criteria (in no 
particular order): i) randomized selection; ii) geographic distribution proportional to all 
communities assisted by PROCOMIDA; iii) proportional distribution between communities nearby 
and far away from district capitals; and iv) proportional distribution between big and small 
communities.
2.4.2 Data Collection
The study team utilized a variety of data collection techniques.
Key Informant Interviews to gather perceptions from key stakeholders and participants included: 
i) PROCOMIDA team (management, M&E unit, nutritionists, other sector staff; n=11); ii) local 
authorities (head or members of the Consejo Comunitario de Desarrollo (Council for Local 
Development/COCODE; n=13); iii) CHWs (n=14); iv) midwives of visited communities (n=14); and 
v) peer counselors or instructors (n=3), for a total of 55 interviewees. Study field teams conducted 
community-level interviews in both Spanish and Q’eqchí (with local translators), while they 
conducted interviews with PROCOMIDA staff in Guatemala City and Coban in Spanish or English. 
Each interview lasted approximately 60 minutes (see Annex 7 for qualitative questionnaires).
Focus Group Discussions were completed with 373 family members (Table 5), including 300 
program participants. The goal was to gather information on behaviors, expectations, and 
satisfaction levels from: i) mothers receiving rations during PROCOMIDA’s field implementation 
(through May 2015); ii) mothers who ‘graduated’ from the program; iii) mothers with children under 
five who did not participate in or benefit from PROCOMIDA; and iv) fathers of participating 
families. In one community, a fifth group was comprised of five mothers who graduated from the 
program, and were also heads of household.
36
Table 5: Sex and literacy levels of FGD participants per community (# and %)
Community Females Female
Illiteracy (%)
Males Male
Illiteracy
(%)
TOTAL Female
Participation
(%)
Male
Participation
(%)
1 21 50 7 21.4 28 75 25
2 21 57.1 7 14.3 28 75 25
3 21 42.9 6 0 27 77.8 22.2
4 20 70 7 0 27 74.1 25.9
5 20 50 6 50 26 76.9 23.1
6 19 15.8 7 42.9 26 73.1 26.9
7 20 20 6 16.7 26 76.9 23.1
8 21 57.1 6 0 27 77.8 22.2
9 20 50 7 57.1 27 74.1 25.9
10 19 10.5 7 0 26 73.1 26.9
11 20 55 7 28.6 27 74.1 25.9
12 24 45.8 2 0 26 92.3 7.7
36 In comparison to this group, all mothers from other groups live together with their spouses or partners, who are considered heads of 
household. The group of female single headed households was established in order to be able to determine if there were any important 
differences in their experiences and considerations regarding PROCOMIDA. There were no important differences found, but the evaluation had 
to check on this and could not just assume that experiences and considerations would be the same.
13
13 21 42.9 6 16.7 27 77.8 22.2
14 19 10.5 6 0 25 76 24
TOTAL 286 43 87 24.1 373 76.7 23.3
Discussions included seven participants on average, and lasted approximately 90 minutes. 
Themes discussed varied between participant cohorts.
Direct Observation allowed for primary data collection across a wide range of contexts, captured 
a variety of community and local interactions, and facilitated on-site verification of KII and FGD 
findings. For example, the field teams observed hygiene conditions at community health centers 
(CHC), and confirmed the existence of home gardens. The field study teams visited 14 CHCs, 
two community centers, two schools, and three home gardens.
2.4.3 Analysis
Translators transcribed KIIs and FGDs directly in the field by recording Q'eqchí interview notes in 
Spanish. They later coded discussion content by category for analysis using Atlas.ti
37 software. 
We highlight themes with greater frequency in this evaluation report.
2.5 Limitations
There were a number of limitations in the endline evaluation process:
 Delays to the contracting process required two rounds of solicitations. During the first 
round, Mercy Corps received only two solicitations, and neither passed the initial technical 
review. The second round involved an intensive process of promoting and sharing the 
scope of work with multiple firms. The process faced additional delays during the 
USAID/FFP approval process for the selected consultant.
 There were multiple linguistic challenges with the need to cross-translate between 
Spanish, Q’eqchí, and English, both during field data collection, data management, and 
re-analysis of documents and datasets originally recorded in Spanish and ultimately 
needed in English.
 Although the study team completed the evaluation with professionalism, strong statistical 
capacity, attention to detail, and an intensive feedback mechanism, USAID/FFP rejected 
the final report two times due to readability issues. Since Mercy Corps/Guatemala 
refrained from any significant direct editing of the document, to assure it maintained 
independency, final delivery of the document and approval of the final evaluation was 
delayed for over a year between reviews of FFP and responses from Mercy Corps. 
Ultimately, Mercy Corps contracted a separate consultant (HDCi) to rewrite the document 
and re-verify that all indicator calculations were correct, given the need ensure accurate 
and consistent data representation, use of professional language, and a polished 
presentation.
 The virtual absence of services in 2014/2015, and changes to the MOH 
intervention/service model in early 2015, seriously affected outcomes at the community 
level and outcomes related to MOH services. This affected the search of qualified MOH 
staff at the CC level during the final evaluation process. In some communities, there were 
no CHVs, and in others, the study field teams had to return several times to find a newly 
hired CHV.
37 The Atlas.ti version 6 from 2011 was used, http://atlasti.com/.
14
3 FINDINGS
3.1 Program Description
PROCOMIDA sought to improve the nutritional status of pregnant and lactating women and 
children under two in 936 vulnerable communities in selected municipalities
38 classified by the 
Famine Early Warning System Network (FEWSnet) as having poor food security and livelihoods, 
and some of the country’s highest stunting and infant and maternal mortality rates.
39
To ensure that PROCOMIDA’s activities strengthened existing health service structures, Mercy
Corps worked at different levels in collaboration with the following implementing partners:
 Local health commissions (community level)
 NGOs implementing PEC40
(CC level)
 Municipal health commissions and other indigenous programs (municipal levels)
 MOH Secretaria de Seguridad Alimentaria y Nutricional (Secretariat for Food Security and 
Nutrition/SESAN41 – department level, and central level for coordination of overall 
strategies and work plans)
These collaborations extended through two presidential administrations,42
four MOH minister 
turnovers, four MOH departmental director turnovers, and three SESAN secretary turnovers. It 
witnessed the breakdown of primary healthcare services under PEC in late 201443
through early
2015, when the MOH cancelled all NGO/private service provider contracts (sub-grantees of 
PROCOMIDA) in Alta Verapaz. Ramifications on the health and well being of local populations 
from this interruption were devastating, and affected several of PROCOMIDA’s expected results. 
Despite this, the implementation period coincided with increasing political attention and 
prioritization of nutrition and health issues.
44 This included Guatemala joining the Scaling Up 
Nutrition movement in December 2010, and the Presidential Initiative ‘Zero Hunger Pact’ in 2010,
which aimed to reduce malnutrition in children under five by 10%. PROCOMIDA activities were 
included in the Zero Hunger Pact, resulting in intensified coordination with MOH/SESAN.
3.1.1 Key Baseline Findings
To ensure guiding principles and actions were appropriate and operational for the specific local 
context, formative research (baseline) was conducted in early 2010 in 15 Alta Verapaz health 
districts to: i) understand food, nutrition, and health conditions and practices; and ii) identify key 
barriers that could prevent adoption of promoted practices. The baseline suggested that extreme 
poverty, large family size, limited land access, and limited income generation and employment 
opportunities created challenges for most households to access necessary food quality and 
quantity. Only limited fruit and vegetable production occurred in home gardens, and was sold
38 Implementation originally covered the municipalities of Cobán, San Pedro Carchá, Lanquin, Cahabon, and Senahú, expanded in August
2012 to include Chahal, Fray Barolomé de las casas, and Chisec were added.
39 WFP (2011). Evaluación de la seguridad alimentaria de los pueblos afrodescendientes, en las comunidades de las yungas de La Paz. 
http://es.wfp.org/sites/default/files/pma-afrobolivianosyyungas.pdf.
40 In Guatemala, the majority of rural communities access health services through the government System of Integrated Health Care, Coverage
Extension Program (PEC). Instituted in 1996 as part of the Peace Accords, PEC aims to reach rural areas and improve primary healthcare. 
PEC is operated by MOH-contracted private service providers/NGOs, working with community support but paid with public funds to provide 
primary health services within communities. PEC provided services on an itinerant basis, with a doctor or nurse visiting each CC monthly.
41 The Law on Food and Nutritional Security System established SESAN as coordination structure that involves all relevant government 
ministries, donors, and civil society at national and local levels.
42 Government of Alvaro Colom (January 2008 to January 2012) and Otto Perez (January 2012 to January 2016).
43 Beginning in late 2012, reform processes in the health sector sought increased control of service provision through public institutions in an 
effort to achieve universal coverage. In early 2014, ministerial staff were expected to gradually take over services provided by NGOs through 
PEC. This transition process failed. In November 2014 NGO service provider contracts were canceled by the government in various 
departments, including Alta Verapaz, leaving the populace without primary healthcare during important parts of 2014 and early 2015.
44 The Guatemalan Constitution through Articles 2, 94, 97 and 99, provides specific provisions to everyone’s right to food. Similarly, Article 43 
of the Health Code states that actions to ensure the availability, production, consumption, and adequate biological utilization of food should be 
promoted in order to support achieving food and nutrition of the Guatemalan population.
15
rather than consumed. General knowledge regarding optimal health and nutrition was limited. The 
suboptimal utilization of food, household resources, and health services contributed towards 
overall food insecurity. Several factors linked to the problem of undernutrition focused around two 
recurrent themes: i) behaviors that influence nutritional status; and ii) limited access to healthcare 
and related services.
While household hunger was uncommon, household and child diet quality was poor, and IYCF 
practices suboptimal. Suboptimal growth and development of children under two years of age 
was compounded by repeated illness, poor hygiene and sanitation, and suboptimal IYCF. Low 
quality household and infant diets meant that children lacked essential micronutrients. Stunting 
was a major problem, with average height-for-age z-scores were significantly lower among boys 
compared to girls.45 While there was almost no wasting, and the prevalence of underweight was 
moderate, the majority of underweight was likely explained by high stunting.
The baseline study further found that two-thirds of mothers were illiterate and did not speak 
Spanish, with a high prevalence overweight and obese (average body mass index/BMI 24.8). 
Mothers were short in stature, which places their fetus at increased risk for growth retardation in 
utero, and implies they experienced stunting during their own early childhoods (inter-generational 
pattern). Knowledge of optimal breastfeeding practices was high, but only half of mothers could 
identify the correct age to introduce complementary foods to infants. Early introduction of 
inappropriate foods and liquids, along with poor food hygiene, poor water quality, and poor food 
diversity, were a concern. These are associated with poor infant nutritional status and morbidity, 
and high rates of stunting and diarrhea. Feeding frequency was inadequate in many children, 
pointing to the fact that mothers’ improved knowledge did not always correspond to changes in 
actual practice. Few mothers knew the danger signs of pregnancy or childhood illness. While 
most attended a sufficient number of prenatal visits, postnatal care appeared inadequate with less 
than half of recommended visits attended. Growth monitoring and promotion in children 12 to 23 
months was nearly universal, yet less than half of children received a mega-dose of vitamin A, 
and only half received iron and folic acid. Barely two-thirds of children 18 to 23 months were fully 
immunized, with most vaccinations received later than recommended. Only 60% of the baseline 
sample of children under two consumed four or more food groups, and only 52% consumed iron￾rich or fortified foods in the previous 24 hours.
3.1.2 Program Design
PROCOMIDA followed recommended nutrition guidelines, sustained by technical and empirical 
evidence in line with global protocols on child nutrition. The innovative program employed a 
preventive methodology based on research and results from World Vision’s PM2A approach, 
which focused on protecting children at the most critical period in their development. To do this, 
PROCOMIDA worked with NGOs and the MOH to establish and/or strengthen community 
structures to improve the provision of culturally and technically appropriate health services, and 
specifically adequate maternal and child health and nutrition (MCHN) service coverage and 
quality. PROCOMIDA further provided nutrition and health education to empower mothers
46 to 
adopt best practices, including seeking health services, towards realization of the program goal 
and strategic objectives (Table 6).
45 There was no significant difference in the prevalence of stunting between boys and girls – boys were shorter in general, but no more were
categorized as stunted.
46
‘Mother’ is understood to be either the biological mother, or the primary caregiver.
16
Table 6: Results framework
Goal: Improved nutritional status and health of women and children vulnerable to food insecurity in
northern Guatemala
Strategic Objectives Intermediate Results Activities
1. By 2015, pregnant and
lactating women, children 
under 2, and malnourished
children under 5 in program 
areas have improved and 
sustainable health and 
nutrition status
1.1 Increased consumption of
energy and nutrient dense foods in 
households with pregnant/lactating 
women, a child 6-24 months of age, 
and/or a child 24-59 months with 
acute malnutrition
1.2 Improved household knowledge,
attitudes and practices for key care 
behaviors
1.3 Increased engagement of
households in reducing vulnerability 
to food insecurity
 Consultation with government and 
civil society leaders
 Establish / strengthen community
health commissions
 Disseminate criteria and selection 
of beneficiaries
 Provide preventative rations to all
pregnant or lactating women and 
all children aged 6-24 months
 Management of acute malnutrition
 Develop behavior change 
communication (BCC) messages
 Enhance the capacity of 
community outreach actors 
(CHCs, CHVs, etc.) to deliver
BCC
 Form mothers’ groups
 Improved practices reinforced 
through home visits, ration
distribution and radio
 Establish community emergency 
funds to facilitate transportation to 
health services
 Connect with ongoing Mercy
Corps food security programs
2. By 2015, healthcare service
providers at community 
through municipal levels have 
improved Service quality and 
delivery
2.1 Increased communication and
interaction between vulnerable 
populations and health service 
providers
2.2 Increased and improved
provision of minimum standards in 
health services to women and 
children
2.3 Increased promotion and
planning by healthcare providers to 
meet community healthcare needs
 Enhanced technical capacity of 
health service providers to provide
strong MCHN care
 Improve MOH extension services 
through building capacity of SIAS 
(Sistema Integral de Atención en 
Salud / Integrated Health Attention 
System) NGOs
 Strengthen recognition and 
treatment of children with severe
acute malnutrition (SAM)
 Facilitate regular meetings among 
community health commissions,
community leaders, health service 
providers to plan around 
community health-related priorities 
and concerns
 Facilitate orientation visits to CCs,
health posts and health centers by 
community members
 Strengthen health facility capacity 
to provide culturally and
linguistically appropriate care
 Build the capacity of SIAS￾implementing NGOs and MOH 
health post staff to prioritize, plan
and advocate for community 
health priorities with key 
municipal, departmental and 
national actors and decision￾makers
17
In line with their results framework and proposed preventive activities, PROCOMIDA based 
program design on the following basic lines of action:
 Universal coverage
 Food distribution
 Behavior change promotion
 Health service strengthening
Universal coverage. Under the universal preventive approach, all households with i) pregnant 
mothers, ii) lactating mothers of children under six months, and/or iii) children under two were 
targeted, not just those demonstrating signs of malnutrition. This approach is intended to improve 
growth and development within the first 1,000 days period (PM2A), or to prevent long-term 
consequences (diminished psycho-motor skills, work capacity, IQ, earning potential).
47 While the 
program defined the mother-child pair as its main participant beneficiary, the involvement of 
mothers was necessary at all levels of intervention, so the program measured beneficiaries in 
MCUs. A fourth participant group consisted of mothers with children between two and five years 
of age with acute malnutrition, who were temporarily enrolled for three-month periods when 
presenting with severe acute malnutrition (SAM). All participants were provided unique 
registration codes (ultimately assigned to each household), and received participant cards that 
they brought to trainings and food distribution points, allowing PROCOMIDA to track participation. 
Each participated for a minimum of six months and a maximum of 30 months. As of project 
implementation end (May 2015), 1,180 communities had participated in the program, reaching
60,605 direct participants, the vast majority children between six and 24 months of age. Though 
participation was voluntary, the final evaluation found that 88.9% percent of households surveyed 
had eligible mothers and children and had participated at some point in PROCOMIDA (between 
July 2010 - May 2015).
During qualitative fieldwork, many participants expressed frustration that the program focused on 
children under two years. Former health interventions, from both the MOH and other NGOs, 
focused on children under 5 years, hence the reference to this age range. “I would say that 
(PROCOMIDA staff) should give us food until our children are five years old because when they 
turn two years old, this is when they (really) start eating” (graduated mother, Jobchacob). While 
program implementers and funders understand the scientific argument for focusing on children 
under two years, Mercy Corps staff could have better communicated the logic of the PM2A 
approach (and first 1,000 days) to beneficiary households and mothers.
While PROCOMIDA’s design is based on the first 1,000 days concept (PM2A), the program was 
limited by the timely identification, enrollment, and retention of pregnant women. According to 
data provided by the program, the average duration of participation per MCU was 567 days. Had 
the project offered a suite of multi-sectoral activities, PROCOMIDA could have achieved more for 
these mothers and children.
Food distribution. PROCOMIDA provided participants with a balanced ration to overcome food 
insecurity, differing according to cluster: individual rations of corn-soy blend (CSB - four kg), lipid￾based nutritional supplements (LNS), or micronutrient powder (MNP), and household rations of 
rice (six kg), pinto beans (four kg), and vegetable oil (two liters). The program provided individual 
rations to mothers during pregnancy and lactation until the child reached six months, after which 
it was transferred to the child until 24 months of age. Food supplementation was an incentive for 
participation, and a source of macro- and micro-nutrients for children between six and 24 months, 
and pregnant and lactating women. Because of the research focus, PROCOMIDA distributed five
47 FHI 360 (2010). FANTA-2 - Title II Technical Reference Materials. TRM-01: Preventing Malnutrition in Children Under 2 Approach (PM2A): A 
Food-assisted Approach. Washington, DC
18
different monthly rations: one standard ration and four alternate rations. For the research, the 
program randomly selected 120 CCs from the initial program intervention area. For each research 
arm, they selected 20 CCs and an additional 20 CCs as a control group. The latter did not receive 
any program interventions, including training or food rations. All non-research CCs received the 
standard ration (Ration A in Table 7).
48
Table 7: Distribution of monthly rations
Research Arm A B C D (LNS replaces E (MNP replaces F (No
CSB) CSB) intervention)
# of CC’s 20 20 20 20 20 20
Ration Type Commodity KG KG KG KG KG KG
Individual CSB 4 4 4 LNS MNP 0
Ration
Family Ration Rice 6 3 0 6 6 0
Beans 4 3 0 4 4 0
Veg Oil 1.85 0.925 0 1.85 1.85 0
TOTAL 15.85 10.93 4 11.85 11.85 0
PROCOMIDA was highly effective at providing all goods and services under their control, 
including financial and material (food), on time. At program start, there were problems regarding 
micronutrient availability, which resulted in some delays.
49 Program staff quickly resolved them.
Most Title II programs distribute commodities in 50 kg or 25 kg bags to distribution sites, where 
staff hand them out to participants who bring their own containers to transport home. Borrowing 
from learnings in social marketing, PROCOMIDA chose to pre-package food rations into individual 
household packages to better organize distributions, provide accurate weight and hygiene 
controls, and facilitate a user-friendly approach to distribution. They provided each household a
‘master’ ration bag allowing them to carry their food home in the traditional manner, using a strap 
around their forehead with the bag on their back. They printed 29 key messages (pictograms
without text, consistent with the BCC strategy, see below) on the pre-packaged monthly ration 
bags, as daily program message reminders (Annex 9).
The repacking operation occurred at a specially designed warehouse in Coban, Alta Verapaz, 
with 2,400 square meters of floor space to store and repack an annual average of 3,000 metric 
tons of donated food. Initially, the program distributed vegetable oil in four-liter bottles packaged 
in the United States and distributed every other month, based on a monthly ration of two liters. 
When the program’s field team noted that monthly attendance dropped on the months they did 
not distribute oil, PROCOMIDA recruited a local technical university to design and build a machine 
to bottle the oil into two-liter containers for monthly distribution.
PROCOMIDA delivered 967,708 rations consisting of 15,172 metric tons of food donated during
58 months of uninterrupted monthly distributions (Figure 2). The monthly distributions used 48 
transport routes during the first 16 workdays of each month. Field staff informed participants of 
the exact date for the next month’s distribution during the monthly BCC educational session. To 
ensure that food arrived on time, they dispatched three trucks each day from the central 
warehouse in Coban during the monthly distribution period. These trucks traversed an average 
of 5,726 kilometers (with a peak of 6,309 km) a month over mostly unpaved roads to deliver the 
rations, for a program total of 337,851 km covered. At the conclusion of the distribution period,
48 Initially the program had larger family rations (12 Kg Rice, 8 Kg beans, and 3.7 Kg vegetable oil for ration A), but after the first year they were 
reduced to the here mentioned ration sizes, as agreed with IFPRI and USAID/FFP.
49 IFPRI was responsible for the purchase and import of micronutrient, delays in distribution was due to delayed approval of the human 
research trails by MOH.
19
only 65 metric tons of different food items remained, which Mercy Corps donated to another Title
II partner (Catholic Relief Services).
30,000
Figure 2: Number of rations distributed
25,000
20,000
15,000
10,000
5,000
0
Jul-2010 Jul-2011 Jul-2012 Jul-2013 Jul-2014
Behavior change promotion. To overcome identified gaps in knowledge and health seeking 
behavior amongst participating mothers, PROCOMIDA developed and implemented a 
comprehensive and targeted BCC component. PROCOMIDA utilized the BEHAVE Framework to 
design interventions, which consisted of both health education and the deliberate promotion of 
specific behaviors at household and community levels. Seventy-two key messages (included in 
Annex 9) were identified and organized around five themes: (i) food and health; (ii) exclusive 
breastfeeding; (iii) care of pregnant and lactating mothers; (iv) care of children from six to 24 
months; and (v) sick and malnourished children. Materials included flipcharts, messages printed 
on ration bags, and pamphlets. Program field staff utilized them during monthly trainings (with the 
aid of a training guide). All training materials were pictorial with little or no text, and the BCC 
strategy used adult literacy techniques focused on collective participative learning through 
positive deviance. Enrollment and receipt of rations was conditional. Pregnant and lactating 
women were required to participate in trainings in order to qualify for rations (Figure 3). When 
food packages arrived in a community, they were stored in the storehouse until the training
activity took place, and only distributed after the training’s conclusion. Depending on climate
conditions and the scope of training activities, the time between food arrival and distribution
averaged one to three days. While it was expected that the reception of rations would be the
strongest incentive to high training participation, the qualitative endline research suggests that
common situations made participation impossible, including mothers with ill children combined
with the absence of fathers from homes (periodic labor migration). Participation rates averaged 
78.6% between November of 2011 and May 2015.
50
50 The first month of food distribution was July 2010, and the last month was May 2015. Participation rates in training events have been 
measured systematically only since November 2011.
20
100
90
80
70
60
50
40
30
20
10
0
Figure 3: Participation rates at PROCOMIDA training events (%)
Assistance level Absenteeism
To ensure fluent communications between program staff and participants, as well as gender 
balance, PROCOMIDA organized field staff/educators into trainer teams of one woman and one 
man, speaking the native language or bilingual in Q’eqchí-Spanish. Trainers divided participants 
into four identified groups according to interest and cultural necessity: (i) pregnant women; (ii) 
lactating mothers with children from zero to six months of age; (iii) mothers with children from six 
to 24 months; and (iv) mothers of sick infants and children under 2. Trainings focused on a key 
message, coordinated with existing MOH messages. Training teams facilitated 60,995 short 
educational sessions with participating mothers over program implementation.
PROCOMIDA utilized the Care Group approach and positioned women as change agents in their 
communities, enlisting throughout the duration of the program 2,549 mothers as Mother Leaders 
to model behaviors – starting in their own homes. Mother Leaders were equipped with knowledge 
and interpersonal communication skills towards promoting hygiene, health, and nutrition through 
specific behaviors, counsel, and advocacy. Educational sessions implemented by program staff 
included monthly recipe demonstrations for Mother Leaders to replicate with up to 20 peer 
neighbor participants assigned to each of them, ensuring transfer of knowledge to promote more 
balanced, nutritious, and diversified diets to participating families. Field staff and Mother Leaders 
conducted a combined 93,685 cooking demonstrations.
Field and PEC staff complemented trainings with household visits, stressing the importance of 
attending trainings, strengthening key messages, monitoring food ration use, and following up on 
pregnant women and malnourished children. Over program implementation, they conducted
74,017 visits. They introduced the Household Action Plan during these home visits. The
Household Action Plan is a pictorial diagram illustrating activities around their home to be
21
maintained or improved to ensure proper hygiene and sanitation, as well as home gardens to help 
diversify family diet by growing vegetables, fruit, and herbs. By program end, Mother Leaders 
implemented 11,587 home gardens. However, most of these home gardens provided vegetables 
during only one season given that seeds (purchased locally by communities themselves) were 
hybrids and thus not usable for seed reproduction. Field and PEC staff also promoted homemade 
hand washers called Ch’ajleb-uq’ (tippy taps) made from the program’s reused oil bottles during 
visits. Ch’ajleb-uq’ use only small amounts of water and help make hand washing fun and habitual.
Health service strengthening. At the community and healthcare services level, staff 
implemented PROCOMIDA through CCs, which are rallying posts located in rural aldeas (small 
rural villages or communities) that do not have any other health services. CCs were part of the 
MOH first level of attendance, and managed by local health implementing NGOs contracted by 
the MOH. The main CC task included carrying out the Integrated Care of Prevalent Childhood 
Illnesses and Integrated Health Care for Women and Children at the Community Level Program, 
Guatemala’s integrated healthcare strategy established by the MOH. This program provided: i) 
integrated case management of childhood illness; ii) monthly growth monitoring and promotion 
for children; iii) standard preventive care for pregnant and lactating women; and iv) vaccinations 
and micronutrient supplements for pregnant and lactating women and children under five years 
of age. CCs were essential to the Zero Hunger Pact, but were limited in their ability to supply 
quality healthcare due to lack of necessary medical equipment and supplies (including drugs), 
and limited health staff knowledge. At its peak, staff implemented PROCOMIDA through 358
Training and Distribution Posts, including 344 CCs and 14 health posts.
PROCOMIDA enhanced CC activities by providing trainings for CHVs working in intervention 
areas in themes such as nutritional evaluation, nutrition, and malnutrition. Staff invited 
decentralized health service providers (local NGO health implementers) and their teams to attend 
all trainings. The program extended additional institutional strengthening to these providers to 
improve their data management, reporting, financial, and organizational skills. All program field 
staff themselves received monthly trainings in either i) BCC, ii) nutrition, iii) monitoring and 
evaluation (M&E), iv) data management (management information system/MIS), v) taking 
standardized anthropometric measurements, vi) use of data collection forms, or vii) identification 
and referral of SAM. All of these activities strengthened local capacity and contributed towards 
sustainability.
3.2 Program Research
In addition to direct activity implementation, PROCOMIDA incorporated a research component as 
part of FANTA’s research to validate the PM2A approach in two countries (Guatemala and 
Burundi). IFPRI implemented and coordinated the research in Guatemala with Mercy Corps, and 
in Burundi with Catholic Relief Services (CRS).
The basic research question tried to identify the most effective mix of different foods in the monthly 
ration to be distributed, for reducing undernutrition at the lowest possible cost.
51
IFPRI randomly 
selected six clusters of 20 communities (total of 120 research sites), with one cluster as a control, 
and five covered by the program and receiving family and/or individual food rations (three arms 
of CBS, one arm of LNS, and one arm of MNP). Given the operations research focus, all data 
collected by PROCOMIDA was quantitative, though qualitative monitoring for specific areas 
(quality of services, adoption of practices) was conducted in years four through six.
Towards a rigorous research design, USAID and the FANTA Project intentionally limited 
PROCOMIDA to a set of activities that excluded livelihood security (agriculture, income 
generation, livelihoods, water sanitation, hygiene) on the grounds that their inclusion would
51 This evaluation does not address the results of these research questions, answered through a separate report prepared by IFPRI.
22
complicate research analysis by adding too many variables. Thus, Guatemala and Burundi 
programs are not directly comparable to other Title II programs.
The rigorous research component itself restricted the benefits that PROCOMIDA was able to 
offer, and when compared to programs of wider scope, was less able to meet local needs or fulfill 
individual and community expectations. This was difficult for participants to understand, despite 
PROCOMIDA’s explicit efforts to communicate the nature of the program to authorities and 
beneficiaries. To address participant confusion on the limits of the program, without compromising 
the provision of fully valid research results, PROCOMIDA implemented Household Action Plans 
in 2012, and took great effort to communicate the nature of the program to authorities and 
participants and support them in more sustainable ways. It was not a perfect solution, and some 
participants maintained a desire that the program should have provided more support in terms of 
agriculture and livelihood activities, especially considering the increasing violence and insecurity 
in Alta Verapaz.
3.3 Program Resources
USAID, through the Bureau of Democracy, Conflict, and Humanitarian Assistance / Office of FFP, 
used monetization proceeds and Section 202(e) to fund PROCOMIDA. Overall resources 
mobilized throughout the entire program by FY reached a life-of-award budget of approximately 
inclusive of more than 35,000 metric tons (MT) of direct distribution and 
monetization commodities.
Table 8: Summary of food aid program resources in US$
Fiscal Year
Commodities Called
Forward
Commodities for
Direct
Distribution (MT)
Commodities for
Direct Distribution
(cost & freight) US$
Commodities
for Monetization
(MS)
Commodities for
Monetization
(cost & freight) 
US$
A unique feature of Title II programs in Guatemala is voluntary contributions made by program 
participants, normally used to cover food commodity transportation costs. In the case of 
PROCOMIDA, these Community Funds remained within the communities for local development 
needs, and managed by local health committees trained by the program.
Communities held assemblies at program start to determine the nominal contribution amount per 
month, set at 10 and 15 Quetzales (US$1.25-2.00). Health committees then planned for annual 
Community Fund use, providing a fund report at quarterly public assemblies. Each community set 
aside a portion of their fund to pay for emergency health expenses, particularly those related to 
pregnancy and child illness. Access to the Community Fund was available to all community 
members during an emergency, regardless of program participation.
When used, households received emergency funds as a loan to be repaid within three to six 
months. Participating communities collected over 12 million Quetzales (over US$1.6 million) over 
the life of the program towards their Community Fund, demonstrating that local communities are 
capable of raising, managing, and utilizing funds for their own development needs (Figure 4). This 
amount equals an impressive 4.5% of the overall program resources (commodities plus
23
operational costs), and coming from poor and extremely poor households, is a strong indication 
of the program’s importance to these communities. Of this Community Fund, 5% was dedicated 
to emergency funds, 63% to improve community infrastructure (CCs, CHCs, and local storerooms 
at distribution points used to temporarily store PROCOMIDA rations), and the remaining 32% 
covered operating costs (recipe ingredients; home garden seeds - Table 9). By program end, 632 
communities had active emergency funds, and 1,490 emergency cases utilized funds. When fund 
management issues emerged (over contribution amounts, transparency, loan repayment), field 
staff mediated solutions with community managers.
52
“We had to pay Q15 on a monthly basis, 
which was used for improving the health center and for the emergency fund. This fund is very 
helpful, but there is a problem that once a family gets an emergency loan, it is very hard for them 
to pay it back. But we have to insist. We have still not benefited from the fund and we need it to 
be operational” (mother, Seconon).
Figure 4: Amount of voluntary contributions and use per year
4,000,000.00
3,500,000.00
3,000,000.00
2,500,000.00
2,000,000.00
1,500,000.00
1,000,000.00
Storeroom Other 
Infrastructure 
Emergency Fund
500,000.00
0.00
IY2 (2010/11) IY3 (2011/12) IY4 (2012/13) IY5 (2013/14) IY6 (2014/15)
Qualitative evaluation findings confirm that communities and participants appreciated 
infrastructure improvements, which remained with the communities after the program ended. 
They also appreciated the existence of the emergency funds, encouraged by the program and 
which continue to exist in a majority of the communities under local management.. “I think it was 
good to give the contribution, because the whole community benefited. And I would like to see 
that at least the emergency fund continues” (mother, Seconon).
Table 9: Amount and use of financial contributions from participants to PROCOMIDA
Year Emergency Fund Infrastructure
53 Storeroom Other
54 Total
IY2 (2010/11) Q 28,830.43 Q 269,726.88 Q 555,772.82 Q 417,223.11 Q 1,271,553.24
IY3 (2011/12) Q 52,647.61 Q 492,551.78 Q 1,014,904.02 Q 761,896.59 Q 2,322,000.00
IY4 (2012/13) Q 76,636.05 Q 716,978.91 Q 1,477,336.61 Q 1,109,048.43 Q 3,380,000.00
IY5 (2013/14) Q 80,218.44 Q 750,494.49 Q 1,546,395.54 Q 1,160,891.53 Q 3,538,000.00
IY6 (2014/15) Q 348,465.52 Q 1,021,302.13 Q 236,621.03 Q 752,794.92 Q 2,359,183.60
Total Q 586,798.05 Q 3,251,054.19 Q 4,831,030.02 Q 4,201,854.58 Q 12,870,736.84
% of total 4.5% 25.3% 37.5% 32.6% 100%
52 There was one case where the food distribution was suspended for two months until the community reached an agreement and could 
continue within the program.
53 Investment in improvement of the convergence centers.
54 Costs of training events (inputs for cooking recipes) and procurement of home garden seeds.
24
Total in US$
55 $74,750 $412,142 $615,410 $535,261 $1,639,560
3.4 Program Participation
By program end, 88.9% of all households surveyed participated in the program, on a voluntary 
basis. For some remote households, the considerable distance to distribution points was a reason 
for non-participation. For others, the financial contribution to the Community Fund was an 
obstacle. “I did not participate in the program due to a lack of money. Participants had to pay a 
monthly contribution of Q.15,
56 and so I was unable to get in” (non-participant mother, Sacristal). 
Despite the voluntary and unconditional nature of contributions to reception of monthly food 
rations program, insufficient feedback mechanisms from program staff to community members 
and health commission staff regarding financial contributions resulted in confusion and led to 
avoidable non-participation.
PROCOMIDA dealt with a range of challenges regarding the participation of women, including 
high levels of illiteracy and the cultural appropriateness of their participation. PROCOMIDA 
designed the BCC strategy to train and equip women in particular with knowledge and skills to 
improve their well being and the well being of their children. Where the program fell short 
regarding equitable gender relationships was underappreciating the value of such issues as the 
division of labor in child rearing, decision-making power, and access to and control of resources, 
to name a few. The baseline identified that culturally the Q’eqchí of Alta Verapaz allowed young 
women and mothers only limited autonomy. Respondents reiterated this during endline FGDs. 
Many participants noted that efforts to include fathers and other family members in trainings did 
not meet expectations, and made behavior change more difficult. “It would be good if training 
sessions include husbands, so that they can also hear and learn what we are learning” (graduated 
mother, Seconon).
A major barrier to father participation was their lack of availability. Most men engage in 
subsistence agriculture. They were thus unable to participate in trainings that took place during 
farming hours. They also periodically migrate away from home in search of seasonal and other 
temporary work to supplement household incomes. Many fathers stated during qualitative 
fieldwork that they regretted not being able to participate more in PROCOMIDA activities, and 
wished trainings had been held in the evenings or on weekends.
3.5 Program Monitoring
Mercy Corps/Guatemala designed a results based M&E framework to measure performance 
indicators, improve program implementation, and report the degree to which results were 
achieved. In addition to the 2010 baseline, a 2012 midterm review, and this 2015 endline, 
PROCOMIDA’s M&E unit carried out annual participant surveys to verify progress made against 
indicators in the IPTT. The program used results from these surveys, reviews, and other formative 
research to inform, and subsequently adjust, activities towards improved effectiveness and 
efficiency and, ultimately, results. Among the most important adjustments were:
 Family ration size reductions
 Message and BCC distribution channel changes for all training sessions
 Household Action Plan implementation for all participating households
 Home garden introduction
 Oil ration repackaging
3.6 Program Sustainability
Sustainability may be measured by whether partners (public sector and civil society sector)
could i) continue to provide similar inputs to sustain the processes leading to desired results, and
55 Using exchange rate US$ 1 = Q 7.4955 as of August 30, 2015, www.oanda.com.
56 Q 15.00 = US$ 1.91, at an exchange rate of US$ 1 = Q 7.4955 as of August 30, 2015, www.oanda.com.
25
ii) ensure that results achieved under the program would be upheld and eventually replicated by 
new inputs. Towards this, partners must possess:
 Institutional capacity (organizational structure, knowledge, expertise)
 Institutional mandate and normative framework
 Financial capacity to conduct expected tasks
 Sufficient ownership and appropriation
Regarding the public sector, this evaluation was unable to assess in any detail whether the local 
healthcare system, currently being rebuilt in Alta Verapaz, would be able to assume the functions 
listed above. The breakdown of local healthcare systems not only compromised PROCOMIDA’s 
outcomes, but also the possibility to successfully complete its comprehensive exit strategy. 
PROCOMIDA’s training efforts during the final months of fieldwork were immense, as shown by 
monitoring indicator 27 (health facility staff trained) in health and nutrition best practices 
(cumulative), which increased from 1,138 to 3,012 persons between June 2014 and June 2015. 
Only since May 2015 have health authorities been sufficiently organized to allow PROCOMIDA 
to provide some basic training to the new local healthcare staff. It is too early to determine whether 
this intense training effort will pay off, yet the new local health system - as designed - could meet 
its goals of providing quality services at reduced costs, if properly implemented. In order to 
increase the sustainability potential, PROCOMIDA transferred its training methods and materials 
to MOH and SESAN national, regional, and local staff, as well as local NGOs, during the last two 
years of the program.
A third partner to sustainability is the community (household) sector, measured by how well 
they were able to maintain and share promoted behavior change. While evaluation findings 
demonstrate that community knowledge and capacity is improved, most households lacked the 
resources to apply knowledge gained through the program. Concerning nutritional improvements 
and undernutrition trends, optimal nutrition and healthy growth for children in the first 1,000 day 
period will have enduring benefits through their lives. Despite food security situations in 
households changing considerably in the absence of food rations, 41% of surveyed households 
(graduated from the program, and with children 24 to 59 months of age who were program 
participants at some point during their first two years of age) were able to maintain a positive 
nutritional status even in the absence of program interventions (rations). Findings further prove 
that knowledge increased significantly (irrespective of whether all targets were reached), but there 
was a delay in subsequent behavior change. Qualitative fieldwork suggests that the process of 
behavior change was further weakened following PROCOMIDA exit.
Few structures were left behind for local multiplication of new knowledge from graduated 
participants to new mothers, and these are weak. Mothers rely primarily on midwives, future 
programs could increase replication of knowledge by including midwives at the design 
stage.PROCOMIDA staff anticipated that women involved in Care Groups would continue to 
practice the behaviors they promoted or adopted during the project period, and to influence other 
women. They hoped that future local public health staff will support their sustained performance, 
but there is no available information to allow for an informed judgment.
3.7 Program Results
3.7.1 IPTT
This final evaluation implemented a series of household, knowledge, and healthcare center 
endline surveys to estimate achievement of the overall program goal and specific targets. 
PROCOMIDA provided other indicators related to the accomplishment of the program’s delivery.
26
Table 10 below shows the list of impact, monitoring, and trigger indicators with respect to the 
baseline, endline (achieved), and final target values.
Table 10: IPTT – baseline, target, endline, and percent achievement
Indicators Baseline
(2010)
Target (+/-
direction)
Endline
(2015)a
% of 
PROCOMIDA 
target 
achieved
b
# Indicator
Objective 1:
By 2015, mothers have the capacity to improve and maintain the health and nutritional status of household 
members, particularly for children less than 2 years of age, acutely malnourished children under 5 years of 
age, and pregnant and lactating women
#1
Impact Indicator 1:
13.3%
10% 
(-)
% children 0-59 months underweight 14.0% 60%*
(<-2 z-score weight-for-age)
#2
Impact Indicator 2:
59.2%
53% 
(-)
% children 6-59 months stunted
57 53.5% 99%
(<-2 z-score height-for-age)
#3
Impact Indicator 3:
7.19
8 
(+) Average Household Dietary Diversity Score 7.26 91%
(HDDS)
Intermediate Result 1.1:
Increased consumption of energy and nutrient dense foods by pregnant and lactating women, all children
6-24 months of age, and acutely malnourished children between 24-59 months of age
#4
Monitoring Indicator 1:
0
18,900 
(+) # pregnant and lactating women receiving 34,337 182%
food rations (accumulative)
#5
Monitoring Indicator 2:
60.1%
80% 
(+) % children 6-24 months with minimum 71.6% 90%
dietary diversity
#6
Monitoring Indicator 3:
0
35,000 
(+) # children aged 6-24 months receiving food 57,590 165%
rations (accumulative)
Intermediate Result 1.2:
Improve household knowledge, attitudes and practices for key care behaviors
#7 Monitoring Indicator 1: 0.0% 52.6% 75%
57 Note that the sample size for stunting is representative for both age ranges, 0-59 months and 6-59 months. We measured 0-59 months for 
comparison with the baseline reported by IFPRI, but the IPTT was 6-59 months from PROCOMIDA startup.
27
% mothers demonstrating increased 
nutritional knowledge
70% 
(+)
#8
Monitoring Indicator 2:
75.0%
85% 
(+) % 69.5% 82%* of newborns who receive essential 
newborn care
#9
Monitoring Indicator 3:
65.1%
85% 
(+) 75.7% 89%
% children 0-6 months exclusively breastfed
#10
Monitoring Indicator 4:
5.9%
40% 
(+) % mothers that know the danger signs of 41.7% 104%
pregnancy
#11
Monitoring Indicator 5:
4.9%
45% 
(+) % mothers with proper identification of 23.3% 52%
childhood illness warning signs
#12
Monitoring Indicator 6:
26.3%
70% 
(+) % children aged 0-23 months with diarrhea 44.6% 64%
that received adequate treatment
#13
Monitoring Indicator 7:
50.0%
70% 
(+)
% 13.8% 20%* children aged 6-23 months with 
respiratory diseases that received adequate 
treatment
#14
Monitoring Indicator 8:
82.2%
90% 
(+) % m 80.5% 89%* others receiving minimum 
recommended antenatal care
#16
Monitoring Indicator 10:
26.9%
70% 
(+) % mothers receiving minimum 30.7% 44%
recommended postnatal care
#17
Monitoring Indicator 11:
73.2%
95% 
(+) 10.0% 11%*
% children receiving full vaccinations
#18
Monitoring Indicator 12:
18.3%
75% 
(+) 16.2% 22%*
% children receiving routine health services
Intermediate Result 1.3:
Increased engagement of households in reducing vulnerability to food security
#19
Monitoring Indicator 1:
1.8%
60% 
(+) % households with Household Actions 37.4% 62%
Plans
#20
Monitoring Indicator 2:
0
358 
(+) # Training and Distribution Points with 211 59%
emergency funds
Objective 2:
By 2015, healthcare service providers at the community through municipal levels have improved service 
quality and deliver
#21
Impact Indicator 1:
6.5%
40% 
(+)
% 0.0% 0%* health facility staff and community 
volunteers able to identify minimum number
of core health and nutrition practices
#22
Impact Indicator 2:
0.0%
40% 
(+) 0.0% 0%
% of training and distribution points that 
offer mothers pre- and post-natal services
and children growth monitoring and general 
health services
28
#30
Impact Indicator 3:
% of local CHVs who meet minimum 
standards/thresholds for performance
0.0%
30% 
(+)
97.7% 326%
Intermediate Result 2.1:
Increased communication and interaction between community members and health service providers
#23
Monitoring Indicator 1:
0
# health commissions with regular meetings
Monitoring Indicator 2:
358
(+) 211 59%
#24
#25
# health commissions demonstrating 
progress on action plans
Monitoring Indicator 3:
0
358 
(+)
0
2,100
211 59%
#26
# of pregnant women in health facility 
orientation visits
Monitoring Indicator 4:
% deliveries at health facilities
35.9%
(+) 2,124 101%
45%
(+) 43.6% 97%
Intermediate Result 2.2:
Increased and improved provision of minimum standards in health services to women and children
#27
Monitoring Indicator 1:
# health facility staff trained in health and 
nutrition best practices
Monitoring Indicator 3:
0
330 
(+)
95%
3,012 913%
#29 % of detected SAM cases referred per MOH
protocols
N/A (+) 98.7% 104%
Intermediate Result 2.3:
Increased promotion and planning by healthcare providers to meet community healthcare needs
#31
Monitoring Indicator 1:
# persons trained in planning and advocacy 
around food security and health
0 2,600 5,387 207%
Trigger Indicators*:
Trigger Indicator 1:
#32
#33
Wasting: % children 6-59 months who are 
wasted
(<-2 z-score weight for height)
Trigger Indicator 2:
Change in food prices
Trigger Indicator 3:
-
<4.0% 1.3% n/a
- 8.74 3.32 n/a
2,351
#34
Irregular rainfall
Trigger Indicator 4:
- 2,792 n/a
#35
#36
Security
Trigger Indicator 5:
Coping strategies
- 35.7 53.7 n/a
- <10 40.1 n/a
a Achievements that match or exceed the targeted values are shown in green while those that are below target are 
shown in red.
b The endline value is compared to the target value set by the program, to show the percentage of the target
achieved. Indicators that show an achievement of 90% and above are shown in this column in green, while those 
that are below 90% are in red.
29
* Denotes indicators that have endline values less than the baseline - showing an opposite directional change 
compared to the intended target.
At the time of the endline evaluation, PROCOMIDA achieved 12 out of 29 impact and monitoring 
indicators (defined as having values within at least 90% of the target set, regardless of the 
baseline). Out of these, eight exceeded the target values.
 IR 1.1: The distribution of food rations to pregnant and lactating women (indicator 4) 
reached close to 35,000 – nearly doubling the target set of 18,900. The distribution of food 
rations to children aged six to 24 months (indicator 6) exceeded the endline target by 65% 
(57,590 compared to the target of 35,000). Following agreed-upon ration reductions (after 
it was found that there were leftovers), the program was able to expand its interventions, 
hence increasing coverage. The program learned important lessons about strategic return 
on investment, minimum ration requirements, and balancing quantity and reach/coverage.
 IR 1.2: Among indicators for household knowledge, attitudes, and practices, only the 
percentage of mothers with knowledge on the danger signs of pregnancy reached their 
targets (slightly exceeding), demonstrating that while the program had the potential to 
transfer knowledge to participating mothers as a result of the BCC strategy, more could 
have been done to improve attitudes and practices.
 IR 1.3: The number of CCs and health posts (HP) with emergency funds (indicator 20) 
and the percentage of households with Household Action plans (indicator 19), only 
achieved about 60% of the target set.
 SO 2: One of three indicators met the target. The target set for local CHVs who met 
minimum standards/thresholds for performance was 30%, but the achievement was close 
to 100%. It is worth noting, however, that the percentage of training and distribution points 
that offer pre- and post-natal services actually achieved the midterm target of 20%, but by 
the end of the program was reduced back to 0%. This is possibly (but not conclusively) 
linked to the observed last minute dip in post-natal care received (indicator 16).
 IR 2.1: Both the number of health commissions with regular meetings (indicator 23), and 
those demonstrating progress on Household Action Plans (indicator 24), achieved only
59% of their targets. CHCs meet monthly with PROCOMIDA staff, and had half-yearly
assemblies to present their progress towards annual work plans to their constituent
communities.
 IR 2.2: Indicator 29 (percentage of detected severe acute malnutrition, or SAM cases 
referred per MOH protocols) met the set target, likely due to the expansion and inclusion 
of educators and staff at HPs. The number of healthcare facility staff trained in health and 
nutrition best practices (indicator 27) exceeded the set target by about ten-fold, reported 
at 3,012 versus the targeted 330.
 IR 2.3: The number of persons trained in planning and advocacy around food security and 
health (indicator 31) doubled the target, related directly to the expansion of the program.
 SO 1: None of the anthropometric impact indicators met or exceeded target. The 
percentage of children 6-59 months stunted dropped from 59.2% to 53.5%, very narrowly 
missing the target of 53.0%. The percentage achievement, however, was 99%. The 
average HDDS also did not reach the targeted score, but did achieve 91% of the target.
 IR 2.1: The percentage of deliveries at health facilities (indicator 26) increased significantly 
from baseline, but was slightly lower than the target set, at an achievement percentage of
97%.
30
Eighteen indicators did not reach the targets set, including seven indicators that showed values 
below those at baseline (marked * in the IPTT table) and clearly did not achieve endline targets. 
Of these, four indicators had unexpected results that were substantial.
 SO 1: Although statistically insignificant, the percentage of children 0-59 months who were 
underweight increased from 13.3% at baseline to 14.0%, compared to 10.2% for Alta 
Verapaz, and 12.6% nationally (MSPAS 2015). The SESAN Zero Hunger Pact monitoring 
report for 2014 similarly reported 14.3% of children 3-59 months for all 166 prioritized 
municipalities were underweight. This worsening direction of change occurred in the last 
two years of the program; at midterm (year 3), the percentage underweight reduced to
12.3%,The national health system breakdown directly influenced the jump up to 14.0% in 
the second half of the program, and is an important finding for the country.
 IR 1.2: Out of fifteen indicators, only two achieved target. The reduction in the MOH 
budget, issues of supply chain management, the MOH revamp of primary care delivery 
that caused gaps in health services at the community level for up to six months, all lowered 
coverage in Alta Verapaz for full and timely vaccinations and other interventions, and may 
explain this decrease. The breakdown of community health services through PEC severely 
hampered seven indicators. Other indicators had lower levels of achievement or reversed 
direction due to partial CC closeout, including: i) the percentage of newborns receiving 
essential newborn care (dropped by about 5%); ii) the percentage of children aged six to
23 months with respiratory diseases that received adequate treatment (dropped by more 
than 36%); and iii) the percentage of children receiving full vaccinations (dropped 
drastically from 73.2% to just 10%). Only indicators 7 (52% achievement) and 11 (52% 
achievement) were under the direct influence of the program. It is worth noting that, 
although the indicator on the percentage of mothers receiving minimum recommended 
antenatal care dropped only by about 2%, the indicator moved from a baseline figure that 
was already within 90% of achieving the target at the start of the program, to a final figure 
that fell just under 90%. It therefore failed to achieve the target.
 SO 2: The percent of healthcare facility staff and CHVs able to identify a minimum number 
of core health and nutrition practices (indicator 21), and the percent of healthcare facilities 
meeting minimum standards for health and nutrition services and practices, was 
significantly unable to meet the endline target, and in fact dropped from baseline. This is 
linked intrinsically with changes in the MOH intervention strategy, and interim breakdown 
of the healthcare system around the time of this endline survey’s data collection period.
Trigger indicators were measured through secondary data sources, mostly from government 
institutions, with the exception of indicator 36 (coping strategies), which was calculated using the 
CSI and measured through the annual household survey. Neither trigger indicators nor non-trigger 
indicators are linked directly to specific program activities, and therefore change cannot be 
attributable to either one. Information on security was unavailable for this final evaluation 
(indicator 35) from the Governor of Alta Verapaz. Coping strategies (indicator 36) increased, 
meaning that people in intervention areas had to implement more strategies to cope with food 
insecurity, even though they received food rations. Additionally, there were changes in food prices 
(indicator 33).
3.7.2 Discussion of Results
The presentation of results provides a snapshot of two moments in time: April/May 2010, and 
April/May 2015. Between April/May 2010 and the end of 2013, there were some improvements in 
local healthcare services, reported in the program’s ARR through PROCOMIDA’s M&E unit. The 
task of this endline report, however, is to compare the snapshots including the virtual absence of 
services in 2014/2015.
31
Outcomes and results reference the objectives and indicators listed in the IPTT as the contextual 
framework for discussion. As such, Objective 1, which relates to parameters focused on the 
population-level indicators, covers household characteristics, dietary diversity, mothers’ 
knowledge of feeding care and health, childhood illness preventive healthcare practices, IYCF 
practices, and nutritional status of children, begins with the characteristics of households and 
mothers, while Objective 2, which concerns service providers, covers convergence centers, and 
community healthcare.
Objective 1: By 2015, mothers have the capacity to improve and maintain the health and 
nutritional status of household members, particularly for children less than two years of 
age, acutely malnourished children under five years of age, and pregnant and lactating 
women
Household Characteristics. Although no specific poverty indicators were assigned for the 
PROCOMIDA baseline and endline surveys, general characteristics of the population show that 
PROCOMIDA reached both poor and very poor households, and that the overall household 
characteristics did not change substantially during the implementation period. Table 11 presents 
several key household characteristics.
In 2015, the average household size was 4.8 people, lower than in 2010, with an average of 2.4 
members under the age of 18, and 1.4 children under the age of 60 months – a sampling effect 
since the household inclusion criteria was based on the household having at least one child aged 
zero to 59 months. Of all households, 11% had a child under six months of age, and 52% had a 
child between six and 23 months of age. Among the 2,419 sampled children in the zero to 59 
months age group, 208 (8.6%) were from non-participant households, possibly due to the high 
PROCOMIDA participation rate.
The percentage of pregnant women among the mothers sampled was lower in the endline survey 
at 5.2% compared to 10.6% in 2010. However, in both baseline and endline surveys, pregnancy 
among adolescents (pregnancy among mothers under 19 years of age) was at about 5%. A 
comparison of the age distribution of mothers between 2010 and 2015 shows that the share of 
pregnancies decreased for the age group 19 to 29, but remains unchanged for mothers under 19 
and over 29 years of age.
Family planning - Findings from the qualitative evaluation suggest that PROCOMIDA’s capacity 
building in family planning improved pregnancy timing and spacing. PROCOMIDA actively 
coordinated its activities with healthcare providers, NGOs, and others working in implementation 
areas. A reproductive health NGO, WINGS, trained PROCOMIDA field staff in family planning, 
enabling them to promote healthy timing and spacing of pregnancies among mothers reached by 
the program through family planning messages within the BCC component.
Key informants and FGD participants noted that most mothers were planning and spacing their 
pregnancies, in agreement with their partners. This is important considering the strong community 
tradition of allowing pregnancies to occur naturally. Interviewed mothers appreciated their new 
family planning knowledge, highlighting the important role of midwives in this regard.
Endline survey findings suggest that family planning efforts and contraceptive use was most 
successful for women between 19 - 29 years of age. Pregnancy among adolescents, however, 
continued to be an issue, indicating a need to focus on younger adolescent girls. Previously, 
public health institutions and NGOs in Alta Verapaz distributed contraceptives free of charge or 
at low cost, but the availability of free contraceptives through the public healthcare system 
decreased throughout 2014 and stopped completely towards the end of that year. Adolescents 
who were sexually active but not financially independent would not likely have had other means 
of obtaining contraceptives. “There are girls aged 13-14 already with their own babies, while they
32
themselves are still just little girls; but they have no alternative. Once they get together with a 
young man they quickly get pregnant, and these are often risky pregnancies because of their 
young age” (local midwife).
33
Table 11: Background characteristics of surveyed households, mothers, and children, at 
baseline and endline
Characteristics Baseline Endline
(2010) (2015)
Household
Average household size Mean: 5.3 Mean: 4.8
Average age of head of the household
(years) Mean: 32
Percentage of pregnant women 10.6% 5.2%
Number of children (siblings) in the 0-
59 months old age range (% n = 1,736 %
households)
1 1,106 63.7
2 579 33.4
3 49 2.8
4 2 0.1
Mothers
Total number of mothers 1,307 1,748
Average age of mother (years) 28 (± 7.2 SD) 28 (±7 SD)
Percentage of mothers with male
partners
93% 92%
Mothers’ anthropometric classification
(%)
Overweight 35.8% 27.6%
Underweight 0.7% 1.7%
Children (0-59 months)
Total number of children 1,840 2,419
Boys (%) 907 (49.3%) 1,234 (51.0%)
Girls (%) 933 (50.7%) 1,185 (49.0%)
Average age (months) 29.7 (±16.7) 30.3 (±16.6)
Age groups (months)
0-5 (%) 189 7.8%
6-23 (%) 737 30.5%
24-59 (%) 1493 62.7%
Household conditions - In 2015, housing conditions remained of poor quality for the vast majority, 
and access to basic services was poor. The majority had dirt floors (86% - a strong proxy for 
poverty in Guatemala), wooden walls (69%), and aluminum roofs (78%). The most common water 
source was rainwater (47%) and spring or river water (33%). Only 20% of households had access 
to an improved source of water (faucet, bottled, protected well, reservoir). The majority of 
households (92%) used latrines, an increase compared to 82% at baseline. This increase is likely 
due to campaigns to build or improve latrines by the municipality and NGOs in the area. Although 
the percentage of households with electricity increased considerably from 16% in 2010 to 30% in
2015, 70% of the population still did not have access to this critical service.
Qualitative field visits confirmed the precarious living conditions reflected through the household 
survey. Only one of 14 communities visited had piped water. In the remaining communities, 
households collected rainwater or had to walk for 30 minutes to a nearby spring for water, often 
several times a day. This situation represented a severe impediment for the implementation of 
hygiene habits promoted by the program in order to provide proper and healthy conditions in the 
household. “There are children who often become sick. Even if we are being trained regarding 
hygiene habits, we have problems because of water shortage. There are children who do not 
bathe; the main problem is hygiene because of the lack of water” (CHW, Chajixim).
34
91.2% 87.7%
11.6%
0.7%
8.8%
Per
c
enta
g
e (%)
Household hunger scale - The Household Hunger Scale (HHS) is a household food deprivation 
scale based on perceptions and experiences of interviewed households, regarding their sensation 
of hunger.
58 At baseline, only 8.8% of households reported moderate hunger levels, and none for 
severe hunger. By 2015, although there is a slight shift of 11.6% reporting moderate hunger and
0.7% (12 households) severe hunger, the increase was not statistically significant.
In the four weeks preceding the household survey, 28.7% of households reported that there were 
periods with no food in the house,
59 while 8.8% reported that at least one household member 
went to bed hungry, and 6.3% reported that at least one household member had not eaten for an 
entire day.
Figure 5: Household hunger scale at baseline and endline
100%
80%
60%
40%
Little or no hunger 
Moderate hunger 
Severe hunger
20%
0%
Baseline 2010 Endline 2015
The qualitative evaluation confirmed that although most families do not suffer hunger for most of 
the year and have on average an acceptable dietary diversity, the poor economic conditions of 
most communities and families created important obstacles to people's access to safe and 
adequate food. The most dramatic cases are those of families who had no land available for 
growing food and consequently relied exclusively on income to feed their families. This situation 
may be aggravated if women are single heads of household. "In my case, my husband works 
away from home for 30 days at a time. For me it is complicated to be alone when my husband is 
away, because I have no money and I must find a way to feed my children" (graduated mother, 
Samanzana). Further, "Being a single mother is very difficult because we cannot find a husband 
and must always respond to what our fathers or siblings want. This is not right; we have to work 
at some point in order to have some Quetzales. When we grow older we will have to have our 
own household and home for our family; we cannot stay like this forever" (female head of 
household, San Antonio).
58 USAID (accessed October 2016). Household Hunger Scale (HHS): Indicator definition and Measurement Guide. 
http://www.fantaproject.org/monitoring-and-evaluation/household-hunger-scale-hhs.
59 Of the 28.7%, none of the respondents were active in PROCOMIDA at the moment of the survey. Some were participants in the program 
before and others never participated.
35
Average
Household Diet
ary
Diversity
Score
Dietary Diversity. The HDDS, used as a proxy measure of household food access, counts the 
number of different food groups consumed rather than the number of different foods consumed. 
In 2015, households consumed on average 7.26 food groups out of 12 in the 24 hours before 
interview, only a slight increase from the 7.19 recorded in 2010 (Figure 6).
The highest consumption (amount of food) and the most frequently consumed food groups were 
cereals, sugar, the miscellaneous group (junk food, instant food, coffee, tea, condiments), oils 
and fats, pulses and legumes, and eggs. Less than 40% of interviewed households reported the 
remaining food groups. The consumption of micronutrient-rich foods was low, with less than 40% 
of households having consumed meat, and the proportion of households reporting the 
consumption of fruits less as than 20%. Less than 10% of households reported the consumption 
of vegetables and dairy products.
Figure 6: Average HDDS at baseline and endline
8.2
8
7.8
7.6
7.4
7.2
7
6.8
6.6
Target
Baseline survey Final evaluation
FGDs suggest that the main driver of improved dietary diversity was behavioral change regarding 
food consumption and optimal nutritional knowledge. New and additional food types were 
provided via home gardens and the spending of diverted household income. PROCOMIDA’s food 
ration package did not contribute to changes in diversity, in terms of their composition. Anecdotal 
reports suggest that the implementation of home gardens by the program contributed to improved 
dietary diversity, but these could not be confirmed econometrically since the home garden 
component was systematically implemented only during the second half of the program. 
Discussions also revealed that for most participants, they used home gardens during one season 
only and apparently did not extend them beyond that due to the use of hybrid seeds (purchased 
directly by communities, and likely unknowingly).
FGDs revealed that measurements of dietary diversity might not apply consistently throughout 
the year. Measurements made in April (2010 and 2015) were at the end of yearly periods still 
characterized as modestly abundant, whilst there were at least three months per year (June to 
August) characterized by very low diversity and occasionally even a scarcity of food. As such, 
measurements of HDDS in this evaluation may not reflect the true dietary diversity of the 
population throughout the year.
Program participants expressed their appreciation of new nutritional knowledge and recipes to 
cook more diverse foods, with the caveat that they were concerned about: i) the availability of 
ingredients, particularly once their participation in PROCOMIDA ended; and ii) their inability to 
use the cook book provided as it was prepared in Spanish. Mothers cited that they were only able 
to understand Q’eqchí and had to learn the recipes by memorizing them.
36
Meal times - Most FGD participants mentioned having a routine of three meals per day (breakfast, 
lunch, dinner) for the whole family, the most important one being breakfast, which occurred 
between five and seven a.m. Preparing large quantities of food in the morning demanded a 
considerable amount of time from women and girls, who were traditionally in charge of preparing 
food. “We have so much work to do that a day is simply not enough time for carrying out all our 
tasks, we would like to have more time, but it is hard for us because we have to take care of the 
house and the children” (graduated female, Jobchacob).
Mothers with young children spent additional time preparing snacks between meals, while still 
complementing with breastfeeding. Older children may have received food at school, which 
reduced the pressure on households. Even though the coverage of school feeding programs (not 
part of PROCOMIDA) was officially countrywide, many schools, particularly those in remote 
areas, in reality did not have the program in place.
Most FGDs indicated that at dinner, a light meal was prepared or leftovers from lunch consumed. 
The combination of a lack of electricity, durable goods, poor housing conditions, and often very 
low levels of cash income led to considerable limitations regarding the storage of food and hygiene 
when appropriate storage facilities were not available.
Mothers’ Knowledge on Feeding, Care, and Health. The main aspects of breastfeeding learned 
or reinforced through PROCOMIDA, according to FGDs, were knowledge of the nutritional value 
and other properties of breast milk and colostrum versus bottle feeding, and about the amounts 
of food and suitable timing for feeding babies and small children at different stages in life. 
“Previously, I would give solid food to my three-month old baby, but when I entered the program 
I learned that I should not give him any solid foods before six months” (recently graduated mother, 
San Antonio I). Further, "The knowledge we obtained during the trainings includes the preparation 
of different foods, such as banana porridge during its season, and how to feed children and how 
to implement a home garden (...) and to not feed the kids only with what is in the shops, such as 
biscuits. We learned to wash the children's as well as our own hands and the importance of 
keeping them clean" (graduated mother, Secoyou). Lastly, "Knowledge improved regarding the 
quantities of food that should be given to babies and to feed when the baby wants to as opposed 
to only during fixed family mealtimes and to place the spoon in his mouth until he learns" (CHW).
FGD participants noted that although they learned about childhood nutrition and 
appropriate feeding practices, they frequently encountered obstacles implementing the suggested 
practices. Many families had only one cooking pot, did not have the time (or firewood) to prepare 
three different meals a day, or simply did not have any food to prepare. PROCOMIDA achieved 
considerable improvements regarding mother knowledge on infant feeding (Figure 7). Although a
high proportion of mothers had knowledge of each of the breastfeeding and feeding practices
questions asked (ranging from 67% knowing that children less than six months should not 
receive semi-solid foods, to 92% knowing that children less than
37
%
mothers de
m
onstrating
incr
e
ased nutritional kno
wledge
six months should not be given colostrum), just over 50% had knowledge of all four infant feeding 
indicators.
Figure 7: Increased mother knowledge of child nutrition at baseline and endline
80
70
Target
60
50
40
30
20
10
0
Baseline survey Final evaluation
Figure 8: Mother knowledge on child nutrition at baseline and endline
Children <6 m should not receive semi-solid foods
Children <6 m should not drink other liquids
67.0%
83.3%
82.0%
Children <6 m should be given colostrum
71.6%
92.0%
86.3%
Children should be breastfed less than one hour after 
birth
Mothers who have good knowledge of infant feeding
35.0%
53.0%
89.0%
82.5%
Endline 2015 Baseline 2010
The baseline and endline surveys asked for nine different danger signs for pregnancy. Mother 
knowledge on proper identification of any four of these nine danger signs during pregnancy 
represents ‘proper identification’ of danger signs, and is presented in Figure 9 as an additional
38
comparison bar. In fact, proper identification of danger signs increased seven-fold during the 
project period. Only one danger sign (persistent backache) showed a decrease, most likely 
related to the fact that during the program period, Guatemalan health authorities removed
‘backache’ from the official catalogue of pregnancy danger signs (and thus the program no longer 
promoted this sign). Knowledge regarding all remaining danger signs improved, and these
observed improvements are statistically significant at the level of 0.03 or below.
Figure 9: Mother knowledge of danger signs during pregnancy at baseline and endline
Trouble for breathing 
No fetal movement after the 5th month 
Continous contraction before the 37th wks 
Stronng headache and blurred vision 
Hands, face and body swelling
Vaginal discharge of clear liquid 
Persistent backache 
Strong stomachache
Haemorrhage and vaginal bleeding
Mother with proper identification of pregnancy danger signs ( 4 
or more)
9.0%
3.0%
4.0%
2.0%
3.0%
4.0%
6.0%
18.0%
24.0%
21.0%
30.0%
26.0%
31.0%
43.0%
41.0%
41.0%
42.0%
57.0%
76.0%
79.0%
Endline 2015 Baseline 2010
The baseline and endline surveys asked for nine different danger signs for pregnancy. Mother 
knowledge on proper identification of any four of these nine danger signs during pregnancy 
represents ‘proper identification’ of danger signs, and is presented in The baseline and endline 
surveys asked for nine different danger signs for pregnancy. Mother knowledge on proper 
identification of any four of these nine danger signs during pregnancy represents ‘proper 
identification’ of danger signs, and is presented in Figure 9 as an additional comparison bar. In 
fact, proper identification of danger signs increased seven-fold during the project period. Only one 
danger sign (persistent backache) showed a decrease, most likely related to the fact that during 
the program period, Guatemalan health authorities removed ‘backache’ from the official catalogue 
of pregnancy danger signs (and thus the program no longer promoted this sign). Knowledge 
regarding all remaining danger signs improved, and these observed improvements are statistically 
significant at the level of 0.03 or below.
Figure 9 as an additional comparison bar. In fact, proper identification of danger signs increased 
seven-fold during the project period. Only one danger sign (persistent backache) showed a 
decrease, most probably related to the fact that during the program period, Guatemalan health 
authorities removed ‘backache’ from the official catalogue of pregnancy danger signs, and the 
program consequently no longer promoted this sign. Knowledge regarding all remaining danger 
signs improved, and these observed improvements, are statistically significant at the level of 0.03 
or below.
39
The results show that even if the PROCOMIDA trainings had emphasized knowledge regarding 
all of these signs, success in converting information (training) into knowledge is very different for 
different signs. PROCOMIDA was less successful in increasing knowledge regarding continuous 
contractions or lack of fetal movements. The qualitative evaluation was unable to clearly identify 
the underlying causes of these different learning outcomes, yet it appears plausible that the 
messages used for transmitting this specific knowledge on contractions or fetal movements was 
more difficult to understand or less convincing for participants compared to training on other 
danger signs.
Childhood Illness. Regarding mother’s knowledge on proper identification of danger signs in 
childhood illness, the endline survey shows a high level of mothers with adequate capacity to 
recognize the classic signs of fever and vomiting (whose recognition was already high at 90%, 
before the program) (Figure 10). However, less than 30% of mothers were able to properly identify 
other symptoms, such as respiratory problems, although improvements were achieved for most 
signs compared to levels shown for 2010. All observed improvements are statistically significant 
at least at a 0.05 level. The symptom ‘Sicker, or not better’ was the only one with an observed 
decline at endline, but the difference was not statistically significant. When combining knowledge 
of all these signs into one indicator (the correct recognition of any three out of six childhood illness 
danger signs), it was observed that the knowledge level increased almost five-fold, from 5% to
23%.
Figure 10: Mother knowledge of danger signs of childhood illnesses at baseline and endline
Blood in stool 
Trouble to breath 
Rapid breathing 
Fever and vomiting
Sicker or not better
5.0%
5.0%
6.0%
11.0%
24.0%
28.0%
23.0%
90.0%
90.0%
13.0%
No food - consumption (breastfeeding, drinking 
and eating) 10.0%
24.0%
Mother with proper identification of children illness 
warning signs (3 or more) 5.0%
23.0%
Endline 2015 Baseline 2010
Public healthcare system - The achievement of the outcomes related to changes in minimum 
childcare practices depended exclusively on the performance of the public healthcare system 
(with a very small exception regarding the performance of midwives). As a result, the performance 
of public institutions affected all findings reported in this chapter.
In November 2014, the MOH decided to discontinue contracts with local healthcare service 
providers, without being able to immediately reestablish services for ministerial healthcare staff 
to conduct. Services in many communities were virtually nonexistent for several months and were 
only slowly reestablished from March 2015 onward, and even later in some communities. This 
represented a virtual breakdown of all primary public healthcare services in the entire territory of 
Alta Verapaz during a key programmatic implementation period. Prenatal and postnatal services
40
% of 
mothers re
c
eiving the
minimum recomm
ended 
care
Figure 11 shows that, in spite of the breakdown in the healthcare system, the indicators on mothers 
who attended four prenatal care visits, and those who received a tetanus shot, supplements, and 
blood pressure control, remained virtually unchanged. The sample included roughly 10% of non￾participants as well as roughly 42% of participants already graduated (with children who had 
turned two years of age by March 2015 or before). Most of these pregnancies took place before 
the eventual breakdown in the healthcare system. Even for the 48% of still active participants, a 
large share of their pregnancies and deliveries might have taken place before November 2014.
Figure 11: Mothers receiving prenatal care assistance at baseline and endline
Mother who received immunization and nutritional 
supplements
66.0%
67.0%
Mother who attended prenatal care visits
81.0%
82.0%
Mother who received minimum prenatal care
57.0%
57.0%
Endline 2015 Baseline 2010
Figure 12: Mothers receiving minimum recommended antenatal and postnatal care at baseline and 
endline
100
90
80
70
60
50
40
30
20
10
0
Antenatal
Postnatal
Target
Target
Baseline survey Final evaluation
Both the indicators on percentage of mothers receiving minimum recommended antenatal and 
postnatal care showed poor performance over the program duration. Although already relatively 
high at baseline (above 80%), the percentage of mothers receiving minimum antenatal care 
(defined as attending at least four antenatal visits) dropped by about 2% in 2015, and did not 
achieve the program target of 90%. The indicator on postnatal care also did not achieve the target 
set, and remained low at just 30.7% in 2015.
Figure 13 shows that the community healthcare center (CHC) was by far the most important 
reference point for both 2010 and 2015, which is in line with the MOH’s recommendation to always
41
seek the nearest possible service point. There is also an interesting increase in visitations for 
prenatal services from public service providers, which presumably took place before the 
aforementioned breakdown in local healthcare services in November 2014. An additional increase 
in demand (albeit less strong) for private services can be observed (private hospitals, private 
clinics, others). Presumably, this increase took place only recently, in response to the crisis of 
public healthcare services. The same would apply to the observed increase of services sought 
from midwives or CHWs.
According to mothers interviewed in FGDs, access to institutionalized antenatal and postnatal 
care and delivery was related directly to household income, notably the family’s ability to pay for 
transport from remote rural areas to healthcare facilities in urban centers – despite the Community 
Fund being implemented to counter this. “I had my delivery at the hospital because the midwife 
told me to go there. I went there by bus but had to walk two hours in order to get to a point where 
I could take the bus. I would not have been able to pay for an ambulance since this is very 
expensive, about 300 Quetzals” (mother, Chitoc).
Figure 13: Institutions visited for prenatal care at baseline and endline
Others 
Private Clinic 
Private Hospital 
APROFAM
Social Security ( IGSS) 
Public Hospital
Health Center
3.0%
7.0%
4.0%
3.0%
2.0%
2.0%
1.0%
1.0%
2.0%
2.0%
14.0%
27.0%
Health Post
Community Health Center
Midwife or Community Health Facilitator
3.0%
1.0%
7.0%
3.0%
91.0%
91.0%
Endline 2015 Baseline 2010
In 2015, 44.0% of deliveries occurred at healthcare centers, a significant increase from 34.0% in
2010. The remaining 56.0% of deliveries took place at home. Figure 14 shows differences in main 
assisting personnel during deliveries.
Figure 14: Place of delivery and assisting staff during delivery at baseline and endline
42
Delivery with local health volunteers 
Delivery with midwife 
Delivery with doctor or nurse 
Delivery at health facilities
Delivery at home
3.0%
7.0%
13.0%
35.0%
44.0%
34.0%
58.0%
56.0%
65.0%
84.0%
Endline 2015 Baseline 2010
Interestingly, although a large (albeit decreasing) share of home deliveries continued to occur, 
assistance of these deliveries by more qualified medical staff seemed to be more frequent. 
Despite only 44% of deliveries taking place within healthcare facilities, either a doctor or a nurse, 
including ‘roving doctors and nurses,’ assisted 84% of all deliveries. Notably, the use of midwives 
or local CHVs decreased sharply.
Preventive Healthcare Practices. Regarding the preventive care of children under two years, 
the proportion of children with proper immunization and their attendance at periodic growth 
monitoring and promotion were analyzed. The MOH, through local healthcare service providers, 
administered immunization services. Growth monitoring and promotion was conducted through a 
mixed system of local service providers assisted by the CHWs or CHVs from the community.
Figure 15: Prevalence of children immunized and measured for growth at baseline and endline
Number of children who attended growth monitoring 
and promotion
16.0%
18.0%
Children 18 - 24 months fully immunization
10.0%
78.0%
Children 12 -17 months fully immunization
10.0%
69.0%
Children < 24 months fully immunization
10.0%
73.0%
Endline 2015 Baseline 2010
Poor preventive services are a clear result of the decay (and subsequent breakdown) of 
community healthcare services provided by MOH, whether directly or indirectly.
FGDs confirmed that participants of PROCOMIDA’s emergency funds viewed these funds as 
one way to access institutionalized healthcare services in cases of emergency, through the loan
43
for transportation and other costs. Regarding the care of newborns, the proportion of mothers 
who reported that their child had received all components of essential newborn care decreased 
overall (
Figure 16). For the purpose of the IPTT and program monitoring, PROCOMIDA defined
essential newborn care through a composite indicator including drying the newborn immediately 
after birth, wrapping the newborn in a warm cloth or blanket immediately after birth, and putting
the newborn to the breast within one hour of birth. The only newborn care practice that 
increased between 2010 and 2015 was early initiation of breastfeeding, which was widely
promoted through PROCOMIDA.
The overall decrease of newborn care suggests unsatisfactory performance of public healthcare 
staff. Thus, while institutional deliveries and professional assistance during deliveries (even for 
home deliveries) increased, it appears that the quality of newborn care decreased. As these 
activities form part of the national public health mandate, PROCOMIDA will not likely be 
accountable for these poor findings.
The percentage of children under two years receiving the minimum treatment for diarrhea 
increased two-fold between 2010 and 2015, meanwhile the percentage of children receiving 
minimum treatment for respiratory problems decreased dramatically. Data analyzed indicate that 
both changes observed are statistically significant at the 0.01 level.
Figure 16: Essential newborn care at baseline and endline
Newborns wrapped immediately after birth 88.0%
99.0%
Newborns dried immediately after birth 86.0%
98.0%
Newborns breastfed within one hour of birth
Newborns received essential care
81.0%
74.0%
70.0%
75.0%
Endline 2015 Baseline 2010
Figure 17: Children receiving minimum treatment for respiratory problems and diarrhea at 
baseline and endline
44
Children 6 - 23 months receiving minimum 
treatment for respiratory problems
Children 6 - 23 months receiving minimum 
treatment for diarrhea
14.0%
26.0%
45.0%
50.0%
Endline 2015 Baseline 2010
Despite the breakdown in primary healthcare provision, changes were observed during the 
implementation of PROCOMIDA regarding increased treatment of diarrhea,. A key factor in these 
treatments consisted of the provision of oral rehydration salts (ORS) to families even without the 
presence of community center healthcare staff. The qualitative evaluation was able to confirm the 
availability of ORS in many of the community healthcare centers visited. In the absence of 
healthcare staff (in early 2015), CHWs administered ORS. Opposite to this example, adequate 
assistance in cases of respiratory problems required greater support from qualified staff. The 
combination of lower mother knowledge about respiratory problems (understood as a lower 
demand for respiratory therapeutic services), and the breakdown of local primary healthcare 
services, most likely contributed to decreases in this indicator.
As in the previous examples regarding service demand (prenatal and postnatal control, 
institutionalized delivery, newborn care), data indicated that the training and sensitization efforts 
carried out by PROCOMIDA had the expected positive effect, increasing demand for these 
services. Local healthcare services were unable to meet increased demand in early 2015 given 
the systematic crisis in the healthcare sector. "Whatever affects our children, whether it is fever, 
cough, grains, diarrhea, flu, conjunctivitis, headache, we would visit the community health center, 
but there are shortages of medicines. For the last six months there was no health attention at all, 
including weighing children” (non-participating mother, Chitoc). Further, “The community feels 
abandoned in certain areas, such as health" (advisory mother, San Antonio 1).
Infant and Young Child Feeding Practices. Figure 18 shows the comparisons between 
breastfeeding practices surveyed in 2010 and in 2015, with exclusive breastfeeding under six 
months rising from 65.1% to 75.7%. At baseline, a high proportion of children already received 
continued breastfeeding until one year of age. Three quarters of newborns received early initiation 
of breastfeeding. Statistics between the baseline and the endline increased regarding 
breastfeeding soon after childbirth. The baseline survey of 2010 indicated that 74% of babies 
were breastfed within one hour after birth, whereas the endline survey of 2015 suggests that 81% 
of babies were breastfed in the same timeframe, indicating a 7% increase (Figure 16). In the 24
hours before the interview, 75.7% of children under six months were exclusively breastfed. At
the age of two, 88.5% of children continued to breastfeed, as recommended. Breastfeeding
practices were extensive (98.7%), and the use of bottle feeding was low (14%). “We breastfeed
all around the clock, because there is no specific time for breastfeeding. When (the babies) are
already grown up we start to give them food and drinks but do not use bottle-feeding because
PROCOMIDA taught us that we should not use bottles because children get sick” (graduated
mother, Sacoyou). The program for continued breastfeeding at the age of two achieved an even
greater improvement, which increased from approximately 77% to
89%.
Figure 18: Breastfeeding practices among mothers at baseline and endline
45
%
children
0-6
m
onths
exclusivel
y breastfed Early initiation of breastfeeding (0-23 mths) 
(n=719/n=897)
78.1%
75.1%
Exclusive breastfeeding under 6 mths
(n=152/n=171)
75.7%
65.1%
Continued breastfeeding at 1 yr
(n=120/n=135)
97.0%
97.5%
Continued breastfeeding at 2 yrs
(n=128/n=165)
88.5%
77.3%
Endline 2015 Baseline 2010
FGDs confirmed that early initiation of breastfeeding (but not necessarily exclusive) is traditionally 
practiced, but that the program successfully encouraged mothers to extend the amount of time 
they breastfed their children. “Before participating in PROCOMIDA, we stopped breastfeeding our 
babies as soon as we got pregnant again, even if they had still not reached two years of age. We 
thought breastfeeding during pregnancy would harm the unborn baby. Now we learned that this 
is not the case and continue breastfeeding” (graduated mother, Sexucti). Further, "Before the 
program we knew that we should breastfeed the child at birth. The only difference was that we 
did this less frequently. With the program we learned that we must breastfeed the child exclusively 
during the first six months of life, and that there is no fixed schedule for feeding a child" (graduated 
mother, Chitoc).
Figure 19: Percentage of children 0-6 months exclusively breastfed at baseline and endline
90
80 Target
70
60
50
40
30
20
10
0
Baseline survey Final evaluation
46
%
with
minimum diet
ary
diversity (children
6-24
m
onths)
The reduced use of bottle feeding was the indicator that demonstrated the greatest improved 
performance since baseline. The proportion of bottle fed children decreased from 49% in 2010 to
14% in 2015, with strong statistical significance (chi-square p-value < 0.0001).
A high proportion of children (84.2%) were introduced to soft or semi-solid foods at the age of six 
months, and solid foods at nine months, as recommended. Minimum meal frequency was 
appropriate for 70.9% of children, and minimum dietary diversity was appropriate, with the 
inclusion of more than four food groups, in 63.8% of children. About two-thirds (66.7%) of children 
consumed iron-rich foods in the 24 hours before interviews.
"What we learned is that when we start to offer solid food to children of six months of age, we 
should begin with soft foods, mashed, or broths. We did not know this until the program came, 
and we are now applying what we have learned. From six months on we start giving food to 
children but always continue with breastfeeding" (graduated mother, Chitoc). Further, “Previously 
I thought a small child eats three times a day, just like us. With PROCOMIDA, I learned that my 
young son should eat five times a day and that even the amount of food is important and that I 
have to take my time and make him eat the entire ration every time so that he can gain weight 
and grow up” (graduated mother, Pinares).
Figure 20: Dietary diversity in children 6-24 months at baseline and endline
90
80 Target
70
60
50
40
30
20
10
0
Baseline survey Final evaluation
Figure 21: Complementary feeding practices at baseline and endline
47
Introduction of solid, semi-solid or soft foods
(6-8 months) (n=96/n=39)
Minimum dietary diversity (6-24 months) 
(n=572/n=700)
60.1%
71.6%
84.2%
89.6%
Minimum meal frequency (6-24 months) 
(n=551/n=718)
63.8%
73.3%
Minimum acceptable diet (6-24 months) 
(n=570/n=718)
46.7%
42.3%
Consumption of iron-rich or iron-fortified foods
(n=570/n=718) 51.7%
66.7%
Endline 2015 Baseline 2010
By 2015, the majority of indicators for complementary feeding practices improved, namely, 
minimum dietary diversity, minimum acceptable diet, and consumption of iron rich foods (Figure 
21). The percentage of children 6-24 months with minimum dietary diversity (71.6%) was higher 
than in 2010 (60.1%), but the difference was not statistically significant (chi-square test p- value =
0.1865).
Only two indicators worsened, compared to 2010 (though the baseline was quite high): i) the 
proper introduction of solid and semisolid foods, and ii) meal frequency. The observed decrease 
in the percentage of children given a timely introduction of solid food is not statistically significant. 
The observed decrease in minimum meal frequency is significant (at 0.05 level). Both 
phenomena, however, could be inter-related.60
Nutritional status of children. Stunting, the main impact indicator defined for PROCOMIDA, 
decreased by 5.7 percentage points between 2010 and 2015. The average height-for-age z￾scores (HAZ) dropped from -2.24 z-scores in 2010 to -2.09 z-scores in 2015 (Table 12). The
observed decrease is seen in both boys and girls, but is only statistically significant among boys
(Table 13). Nevertheless, the general level of stunting remained high at over 50%.
60 The observed decrease of timely introduction of solid food is significant at a 0.14 level, which according to handbooks should not be 
accepted as statistically significant. Our understanding, however, is that given the sensitive nature of the observed topic (behavioral change) 
we should be more flexible regarding the interpretation of results. In other words, we could say that “in case the PROCOMIDA program would 
be repeated 100 times, 86 times a decrease in timely introduction of solid food would be observed”, which at least following common sense
seems to be a result of important scope. Hence, if we accept that the introduction of solid food being postponed (for whatever reason, is a valid 
observation, such a behavior could very well lead to a decrease in the minimum meal frequency indicator, as observed in PROCOMIDA.
48
The prevalence of acute malnutrition measured by weight-for-height (WHZ) < -2 SD remained low 
and relatively stable for children six to 59 months, at 1.1% in 2010 and 1.3% in 2015.
Table 12: Nutritional status of children at baseline and endline
Nutritional status of Baseline Boys Girls Endline Boys Girls
children (2010) (2015)
Underweight (<-2 z-score weight-for-age)
Age 0-59 months
N 1,838 907 931 2,406 1,119 1,085
WAZ (mean ± SD) -1.01 (± - - -1.05 (±
0.96) 0.89)
- -
Percentage underweight 14.4% 13.5%
(confidence interval) 13.3% 14.6% 11.2% 14.0%
[11.4 – [12.5 –
14.4] (12.4-17.0) (9.3-13.4) 15.5]
(12.4- (11.4-
16.7) 15.9)
Stunting (<-2 z-score height-for-age) 
Age 6-59 months
N 1,680 838 842 2,204 1,119 1,085
HAZ (mean ± SD) -2.24 (± -2.09 (±
1.01)
- -
0.95)
- -
Percentage stunted 57.5%
(confidence interval) 59.5% 61.5% 53.5% 54.7% 52.3%
[57.1 – [50.9 –
61.8] (58.1-64.7)
(54.1- 56.0] (51.6-57.8) (49.0-55.5)
60.8)
Wasting (acute malnutrition)
Age 6-59 months
N 1,680 838 842 2,211 1,124 1,087
WHZ (mean ± SD) 0.28 (0.88) - - 0.15 (0.89) - -
Percentage wasted 1.4% 0.7% 1.2% 1.4%
(confidence interval) 1.1% 1.3% 
[0.7 – 1.7] [0.9 – 1.9]
(0.8-2.5) (0.3-1.5) (0.7-2.2) (0.8-2.4)
Table 13: Distribution of stunting among children 6-59 months (<-2 z-score height-for-age), by sex, 
at baseline and endline
Baseline Endline
(age 6-59 months) (age 6-59 months)
n % stunted n % stunted
(<-2 z-score h-a) (<-2 z-score h-a)
Total 1,680 1,000 (59.5%) 2,204 1,179 (53.5%)
Boys 838 515 (61.5%) 1,119 612 (54.7%)
Girls 842 484 (57.5%) 1,085 567 (52.3%)
In
Figure 22, the average values for underweight, stunting, and wasting were plotted for each month 
of age of the children. Trend lines were then plotted based on 5-month moving averages. These 
trends were then compared between baseline and endline.
49
Trends regarding stunting and the comparison between 2010 and 2015 suggested that the 
prevalence of stunting at birth is lower in 2015. Further, the period of instability and steadily 
increasing levels of stunting (denoted by a pink background in the figures above) is shorter in
2015, stabilizing at an average of 20 months, compared to 38 months in 2010. Overall duration 
of the increase in stunting shortened as well, from 33 months to 20 months.
Figure 22: Percentage of children with measurements classified as wasting, stunting, and 
underweight, by age at baseline and endline
Objective 2: By 2015, healthcare service providers at the community through municipal 
levels have improved service quality and delivery
Baseline 2010 Endline 2015
Convergence Centers. PROCOMIDA selected CCs using a two-step selection methodology. 
First, at the municipal level, a list of CCs was compiled from those identified by the government 
as being in food insecure communities (prioritized through SESAN). In consultation, SESAN and 
their Municipal Food Security Classification system, and the MOH, selected four municipalities of 
Alta Verapaz for implementation. Together with PEC, 221 extremely poor rural communities with 
year-round access to five-ton trucks were selected to participate. Then, the locations of these 
CCs were assessed for their accessibility by the delivery trucks, necessary for regular (monthly) 
ration distribution. PROCOMIDA later refined the methodology in consideration of smaller delivery 
trucks, reducing the issue of road access and increasing the number of eligible CCs including in 
remote locations to 358 covered by the program. In implementation year five, the program initiated 
a phased closeout, starting with closure of 141 CCs (July 2014). In the last implementation year,
211 CCs were attended, which explains the low coverage measured in the final evaluation. During 
the length of the program, the target of 358 CCs was achieved. Since each CC generally covers
more than one community, note that many program participants still walk from surrounding 
communities to the CCs to attend training sessions and receive rations.
Regarding household engagement in the effort to reduce vulnerability to food insecurity, one of 
two monitoring indicators was achieved. In Y5, all 358 CCs covered by the program had 
emergency funds. PROCOMIDA made good progress under the second indicator (Household 
Action Plans), but the target was not fully achieved since promotional activities regarding 
Household Action Plans were started in November 2011. During the baseline survey in 2010, only
1.8% of households reported having Household Action Plans, but the figure increased to 37.4%
by 2015, achieving 62% of the targeted figure of 60%.
50
Community Healthcare. Although the target was to have 30% of local CHVs meet minimum 
standards of performance, the figure at endline showed more than 300% achievement of the 
target, with 97.7% of local CHVs meeting such standards. Due to changes in the basic healthcare 
service model, many of these CHVs are actually new and received training directly from the MOH 
before deployment. PROCOMIDA supported the MOH in these trainings, but these high levels 
cannot be exclusively attributed to the program. Two other indicators in Table 10 were recorded 
as 0% achieved in 2015. These were: i) the percentage of healthcare facilities meeting minimum
standards for health and nutrition services and practices (Figure 23); and ii) the percentage of 
healthcare facility staff and CHVs able to identify a minimum number of core health and nutrition 
practices (Figure 24: CHC staff knowledge regarding danger signs and ORS use; and Figure 25:
CHC staff knowledge regarding child feeding practices). These indicator results were likely driven
by the way the indicators were defined, as both indicators refer to service provision and the
knowledge of local healthcare staff. Taking into consideration the list of minimum knowledge and
services to be available, failure in only a single element of the list brings the indictor down to 
zero.
Figure 23: Availability of basic services at CHCs at baseline and endline
Growth monitoring
98.0%
93.0%
Preventive care (Children)
22.0%
40.0%
Prenatal service
11.0%
20.0%
Postnatal service
0.0%
2.0%
Endline 2015 Baseline 2010
51
Figure 24: CHC staff knowledge regarding danger signs and ORS use at baseline and endline
Use of ORS 
Of Dehydratation 
During Childbirth 
In Newborn
In Children
During postpartum
During pregnancy
5.0%
15.0%
12.0%
13.0%
22.0%
29.0%
48.0%
60.0%
56.0%
56.0%
67.0%
80.0%
74.0%
89.0%
Endline 2015 Baseline 2010
Figure 25: CHC staff knowledge of child feeding practices at baseline and endline
Avoid semi-solid foods < 6 months
Avoid other liquids < 6 months
Avoid water < 6 months
71.0%
81.0%
81.0%
84.0%
93.0%
Use of colostrum
80.0%
85.0%
96.0%
Early Initiation Breastfeeding
93.0%
86.0%
Endline 2015 Baseline 2010
Local CHC staff knowledge on the following topics was considered a minimum requirement: i) use 
of ORS; ii) danger signs of dehydration; iii) danger signs during childbirth; iv) danger signs in 
newborns; v) danger signs in children/childhood diseases; vi) danger signs during postpartum; 
and vii) danger signs during pregnancy. Eighty percent of staff reported having received training 
on these topics. The trainings that received the highest attendance were care of sick children 
(80%) and issues regarding pregnancy (75%). A high turnover of MOH personnel limited 
PROCOMIDA’s ability to contribute to an increase in their knowledge. Overall, trends are positive.
By aggregating the different categories into a single indicator, none of the health staff in 2015 
were able to recognize all of the danger signs, or correctly describe ORS use. The corresponding 
baseline value in 2010 was also low at 6.5%. In 2015, the highest levels of knowledge measured 
were on ORS use (89% of staff), danger signs of dehydration (80%), and danger signs during 
childbirth (74%).
High levels of knowledge around the initiation of breastfeeding and exclusive breastfeeding were 
observed in 2015 (over 90% of the staff could adequately identify these recommended practices).
52
Per
c
enta
g
e
O
wn Building
Bathroom
Electricity
ORS
A
moxicilin
Trimetroprim-Sulfametoxazole
Penincilin
Albendazole
Vitamin A
Chispitas
Iron
Folic Acid
Prenatal supplements
The proportion of staff capable of properly identifying that water, other liquids and semi-solid foods 
should not be introduced in children under six months was higher than in 2010. Since 2010, 13% 
of CHCs improved their infrastructure or their access to basic services. By 2015, 96% of CHCs 
were located within their own buildings/property, and 89% had a toilet, although only 37% had 
electricity (Figure 26).
In 2015, due to MOH limitations beyond the program’s influence, only 11% of CHCs had an 
available stock of nutritional supplements (micronutrients) for children. Supplements for mothers 
were more available (57% iron, 50% folic acid), however, levels of all supplements were only half 
(or even less) of those registered in 2010. The exception was vitamin A, whose availability in 2015 
was higher than in 2010, but was still only available at 11% of centers.
The same situation applied regarding the availability of medical supplies, which were far below 
2010 levels. The most frequently encountered product, ORS, was only available in 50% of the 
CHCs visited. Medical supplies required for basic treatment of the most prevalent childhood 
diseases (diarrhea, respiratory infections) were available in fewer than 30% of CHCs visited.
Figure 26: Availability of medical supplies at CHCs at baseline and endline
100.0%
80.0%
60.0%
40.0%
20.0%
0.0%
Infrastructure Amenities and Medicines supply Children
Supplements
Supply
Women Supplements Supply
Baseline 2010 Endline 2015
Regarding the availability of service supplies, the following minimum services for pregnancies and 
childcare were considered:
 Antenatal services – TT vaccination (Tetanus), nutritional supplements, diphtheria
vaccination, guidance on danger signs during pregnancy and finding appropriate care, 
promotion of optimal nutritional practices, hygiene, maternal nutrition, safe sex, family 
planning, and breastfeeding
 Postnatal services - home visits, nutritional supplements, and prenatal services listed
above
53
 Growth monitoring and promotion services - anthropometric measurements, 
development of growth curves, guidance on nutrition and health, immunizations
 General preventive care services - centers offering all of the following services were 
considered: de-worming, counseling, anthropometric measures
In 2015, CHCs offered a variety of services. While they did not offer post-natal care services, 20% 
of CHCs (almost twice that in 2010) did offer antenatal care services, and 22% offered general 
preventive childcare (much lower than in 2010). The availability of these services depended 
directly on the MOH.
Growth monitoring and promotion services promoted and supported by PROCOMIDA (through 
trainings, achieving commitments from volunteers, and providing measurement equipment) were 
available in 98% of CHCs.
3.7.3 Trigger Indicators
Trends in vulnerability contexts
Indicators used by the program to reflect security and coping strategies suggest that the external 
living conditions for Alta Verapaz communities during the implementation period were less than 
favorable. Even though food prices showed a decrease at endline compared to baseline, pests 
and diseases affecting cash crops considerably reduced people’s ability to earn cash income, 
especially in recent years. Additionally, although average precipitation was normal, most of 
southern Alta Verapaz suffered from insufficient rainfall. An exploration of adverse conditions and 
events may provide a better understanding of the scope of the decline in external conditions, and 
the potential impact this had on poor rural households.
Table 14: Adverse conditions and events affecting Alta Verapaz at baseline and endline
Year Event/Conditions Description/Consequences
2010 Tropical storm
Matthew
Caused flooding in several municipalities northern Alta Verapaz. The damage 
was considerable in Coban, Carchá, Santa Cruz Verapaz, Raxruhá, Chisec,
and San Cristobal Verapaz, with flooded areas and damaged crops.
2012/2013 Crisis of coffee rust
plague
20% of the 2012/13 crop was lost, representing nearly 19% of income lost.
Landless workers and small farmers were the most affected. If crop losses 
associated with grain quality are added, the income loss rises to 30% - 70%.
June -
July 2012
Prolonged drought A ‘prolonged heat wave’ resulted in significant rainfall deficits in municipalities 
in southern Alta Verapaz. San Cristobal, Santa Cruz, Chamelco, Tactic,
Tamahú, and Tucurú were affected by a ‘strong’ drought. The impact was
quantified as a 40%-50% loss of staple food production (grains, corn). Lower 
demand for labor and increased food prices (13%) resulted.
2014 Recidivism of 
prolonged heat 
wave
A lack of rain of up to 29 to 45 consecutive days was recorded in July 2014, 
causing losses between 60% (corn) and 80% (beans). A lack of drinking water 
was also reported, and an increase of acute malnutrition in children under five
years. Among all 22 Guatemalan Departments, the highest increase was 
reported in Alta Verapaz, with 554 cases.
Lingering effects of 
coffee rust:
Wages in the sector suffered a 50% decrease, and the levels and quality of the 
coffee harvest was poor. Alta Verapaz was declared to be in a state of ‘food 
insecurity crisis.’ Households relying on the coffee harvest (15%-20%) saw 
their revenues cut by 50%-55%. SESAN began distributing food rations and 
the World Food Program and Save the Children begin distributing cash 
transfers.
Cardamom price 
drop
The price of cardamom plummeted due to overproduction and the negative 
effect of a thrips outbreak. The incidence of the outbreak reached 40% of
production, and 300,000 small producers. The effect of the outbreak reinforced
smallholder problems – smallholders who had already suffered from a price 
drop resulting from overproduction by large-scale farmers. The average price
per ‘quintal’ (100 pound bag) dropped from US$645-193.
2015 Coffee does not 
recover
Despite projected increases in coffee rust control, the harvest continued to 
suffer. The sector showed a harvest below normal production for the third
54
consecutive year, with production 16% lower than the harvest recorded in 2012
- before rust infestation. The area remains in a situation of ‘food insecurity.’
55
4 CONCLUSIONS, LESSONS, AND RECOMMENDATIONS
4.1 Conclusions
It is the conclusion of this evaluation that, through analysis of key final evaluation findings, linked 
to strategic objectives and expected results, PROCOMIDA achieved its’ goal of improving the 
nutritional status and health of women and children vulnerable to food insecurity in northern 
Guatemala. Despite external challenges (economic, systemic, political), PROCOMIDA staff 
managed to deliver efficiently both products and services in a complex geo-cultural and logistical 
environment, benefitting tens of thousands of community members. While program participation 
was voluntary, 88.9% of eligible mothers and children ultimately engaged in program activities.
The following evaluation scores are assigned to PROCOMIDA in view of endline results.
Table 15: Evaluation ratings
Evaluation
Criteria
Assigned Score Comment
Thematic
relevance and 
design
Relevant The thematic relevance was extremely high, but given the research 
approach selected by the donor for this PM2A research program, 
the scope of the thematic intervention was limited.
Regarding design, the program defined a set of targets for local 
healthcare service performance and the knowledge and capacities 
of local healthcare staff, which were dependent on MOH 
interventions outside program control. This evaluation understands 
that the target levels defined were desirable to support other 
outcomes of the program. From a technical point of view, the 
evaluation team considers it a mistake to have fixed targets for 
outcomes that were not within the program’s control. The 
program’s intention was to improve availability and quality of MOH 
services at the primary (community) level. That this was 
unsuccessful was outside of the program’s control – and indicates 
that defined targets should have been more under the program’s 
control.
Effectiveness
Satisfactory (minor
shortcomings) for all 
outcomes within 
PROCOMIDA’s control.
Unsatisfactory for the 
non-achievement of other 
important targets outside 
PROCOMIDA’s control.
The expected results for required activities and effects for reduced 
undernutrition fully materialized for all activities fully controlled by
PROCOMIDA (food distribution, trainings, improvement of local 
health services up until the second semester of 2014).
It is plausible that without the substantive worsening of external 
conditions, effects on undernutrition would have been stronger.
From a technical point of view, the program was less effective in 
achieving outcomes and targets in the sector of supporting public 
healthcare service provision. These results were not under the 
program’s control, and while PROCOMIDA’s support was 
technically sound and most likely had the potential to achieve 
defined targets, the non-achievement of the targets must be rated 
as unsatisfactory.
Influence Significant Econometric analysis confirmed the significant influence on 
improving undernutrition levels and trends by contributing to: i) 
reduction of stunting at birth; ii) reduction of the growth faltering 
period under six months; and iii) reduced level of stunting for 
children under five years with an additional increased probability 
that much of the achieved improvement would endure over time. 
The difference between participants and non-participants 
regarding these results is statistically significant.
For benefiting children, participation in PROCOMIDA represented 
an opportunity for their future, not likely possible in the absence of 
the program.
Efficiency Highly satisfactory (no
shortcomings)
PROCOMIDA staff were highly efficient managers of product and 
service delivery in a complex environment (and topographic area)
56
managing extreme logistical challenges in hundreds of
communities for tens of thousands of participants. Activity delivery 
was accomplished without shortcomings (which does not imply 
there were no problems, but rather that all problems were 
satisfactorily addressed and resolved in a timely manner). The 
program achieved participation rates above average for this kind of 
intervention.
Sustainability Moderately likely
(moderate risks)
There is a high possibility that nutritional outcomes achieved for 
beneficiary participants under two years of age will endure into the
future. There is, however, only a moderate likelihood that mothers
and families might experience conditions to productively use their 
acquired knowledge (which translates into capacities and
behavioral changes in a minor degree compared with what was 
anticipated by the program) in order to protect future pregnancies 
and early childhood development effectively.
In addition, it is only moderately likely that the public sector will be 
able (given existing instructional capacities and available 
resources) to sustain service provisions, which enabled 
achievement of PROCOMIDA’s outcomes. To a certain degree, 
PROCOMIDA did not anticipate the need for deeper engagement 
through capacity building interventions and exit strategies with 
health districts, and concentrated too heavily on supporting local 
levels. This evaluation is aware that the fragility and virtual 
breakdown of primary healthcare provision in late 2014 widely 
determined this situation.
Gender Significant Gender results were significant regarding empowerment, which 
was the main goal of the program. Gender results were, however,
much weaker regarding equity results and inclusion of fathers and 
in-laws, including mothers-in-law.
Overall Satisfactory (minor
shortcomings)
Overall, PROCOMIDA was relevant, well designed technically, and 
efficiently implemented. Results and activities materialized as
expected and proved to have the power to achieve the expected
scope of nutritional core results. This happened under an adverse 
external environment. Without the presence of PROCOMIDA, the
nutritional situation in Alta Verapaz would likely have been worse.
The program was committed to the achievement of results that 
depended on the performance of the public health system, which 
was not controlled by the program. The evaluation team 
considered this fact as most certainly an unavoidable yet important 
shortcoming, technically speaking, since the achievement of these 
outcomes most probably would have reinforced program outcomes 
even more.
More specifically, PROCOMIDA:
 Validated the PM2A approach, even when they implemented only health and nutritional 
interventions. Reductions in stunting, achieved through access to supplementary food 
rations and improved childcare and household environments, confirm this.
 Improved nutrition and reduced undernutrition. Econometric assessment of 
outcomes, and undernutrition trends, show impact from nutritional improvements felt well 
beyond program graduation. There was statistical significance between program 
participants and non-participants in this area. A slight increase in the percentage of 
underweight children birth to 59 months was seen (13.3% baseline; 14% endline), yet the 
percentage of children six to 59 months presenting stunted fell (59.2% baseline; 53.5% 
endline), and the HDDS score improved (7.19 baseline; 7.26 endline). The annual 
decrease in stunting (1.14 percentage points) was a considerable improvement when 
compared against the greater Alta Verapaz Department national survey results, which
57
show only an average decrease of 0.17%/year (1999 to 2009),
61 and against overall
national results during PROCOMIDA’s implementation period of 0.55%/year (2014-
2015).
62
While there is high sustainability potential for children and households who participated in 
the program to maintain nutritional outcomes, and become healthier and more productive 
adults, there is only moderate likelihood that: i) mothers and households will function in 
environments enabling application of program knowledge and skills to effectively protect 
future pregnancies and the early development of current and future children; and ii) the 
government’s primary healthcare sector will be able to maintain and advance program 
improvements given existing structural barriers (skills, resources).
 Achieved key Objective 1 outcomes including improved household knowledge.
Nutrition- and training-related activities were technically sound and, based on formative 
research, both necessary and culturally appropriate. PROCOMIDA staff implemented 
them efficiently and effectively. Consistent nutritional supplementation of children and/or 
their mothers, knowledge and skills on key hygiene practices and behaviors, and dietary 
diversification all contributed to varying degrees to the improved nutritional status of 
participating children. Participating families demonstrated improved knowledge regarding 
childcare, illness warning signs, family planning, and nutrition. Households that 
participated in program activities demonstrated increased efforts to reduce their own 
vulnerability to food insecurity through continued efforts to improve household conditions, 
diversify their diets, and adopt behaviors related to IYCF (exclusive breastfeeding, active 
feeding) independent of financial ability.
 Achieved many Objective 2 outcomes including strengthened local structures and 
infrastructure. PROCOMIDA staff provided high quality technical support, and were 
largely on track to reach targets through FY 2014. Activities focused on strengthening the 
quality and accessibility of primary healthcare services in rural areas. Communities 
benefitted from strengthened local structures for self-management (establishment of 
health commissions and emergency funds; developed local health planning capacity) and 
improved infrastructure (store rooms, primary healthcare centers). Implementation, 
however, occurred in a context of increasingly critical external threats, which ultimately 
exposed the fragility of the PEC delivery model. When conditions beyond program control 
interceded, the PEC system broke down and many primary healthcare-related activities 
were suspended.
While some indicators suggest that at least minimum health standards still increased 
(regarding service provision), after the redesign of the public healthcare system in Alta 
Verapaz (re-launched late spring 2015), only limited capacity and skills provided to local 
healthcare staff by the program remained given new service delivery structures. It is not 
yet possible to realistically speculate on how well new primary healthcare services might 
engage with community structures (health commissions, emergency funds) strengthened 
under PROCOMIDA. During evaluation fieldwork, community leaders expressed their 
intention to maintain and utilize infrastructure, co-opting them to support community 
health-related activities, and were optimistic that new primary healthcare services would 
engage with community structures initiated and/or strengthened by PROCOMIDA.
61 MSPAS (2011). Encuesta Nacional de Salud Materno Infantil 2008 (ENSMI-2008/09). Ministerio de Salud Pública y Asistencia Social
(MSPAS)/Instituto Nacional de Estadística (INE)/Centros de Control y Prevención de Enfermedades (CDC). Guatemala
62 MSPAS (2015). VI Encuesta Nacional de Salud Materno Infantil (ENSMI-2014/15). MSPAS/INE/Secretaría de Planificación y Programación 
de la Presidencia (Segeplán). Guatemala.
58
 Promoted healthier pregnancies. This was an important result of family planning BCC, 
leading to healthier birth spacing and pregnancy timing among program participants, as 
measured through household survey data and confirmed by qualitative fieldwork. 
Coordinated and simultaneous interventions of other organizations (family planning, 
contraception distribution) complemented the program’s family planning activities in 
implementation areas.
 Gender empowerment. A crosscutting programmatic theme, gender results showed 
greater female participation. In particular, women were more involved in decision-making 
around health and nutrition. Increases were noted in mother’s knowledge on child nutrition, 
and danger signs of child illness and during pregnancy. Home gardens enabled women 
to be more involved in food production and dietary diversification decisions for their 
families.
Pertaining to gender equity, results were weaker. The almost exclusive concentration on 
the participant mother-child unit limited involvement of other family members, including 
fathers, in-laws, and siblings. This was a lost opportunity to reinforce the drivers of change, 
only realized in the late stages of implementation when the program did initiate work with 
men and key family members including mothers-in-law.
4.2 Lessons and Recommendations
We provide lessons and recommendations for the improvement of future Title II DFAP and similar 
activities.
 The BCC strategy was crucial to achieving program outcomes. It was important that it 
was sensitive to local sociocultural and gender contexts, and flexible to periodic 
adjustment throughout implementation, based on new learning, monitoring evidence, and 
necessary environmental changes. That participating women benefited (e.g., healthier 
birth spacing and pregnancy timing) supports the importance of including social and 
behavior change (SBC) efforts in future Title II PM2A programs on topics of reproductive 
health and family planning.
Information and messages validated over application of the program’s BCC strategy, as 
well as information gathered through M&E processes, has value and considerable 
potential for further program design. The initial BCC strategy included too many key 
messages, and the curriculum to cover all spread over two years. At midterm, the program 
identified that this was seriously affecting appropriation of these messages by participants, 
and decided to reduce the messages to half. In addition, the training sessions became 
routine, and active participation by women began to decline. More participatory processes 
put into place included Mother Leaders and Household Action Plans. CHCs, and CHVs, 
were more prominently involved, increasing retention of messages amongst participants. 
Future strategies should use a more inclusive SBC focus, prioritize more strictly key 
messages, and involve community members in SBC sessions.
 Awareness-raising efforts were continuous throughout implementation regarding 
program scope and research components. Despite these efforts, however, participant 
expectations were higher than what was possible to deliver. Participants found it difficult 
to accept that: i) different communities received different food packages; ii) food rations 
were complementary (and hence did not cover full household food needs); iii) food rations 
ceased when children reached two years of age; and iv) outside of home gardens (which 
were not implemented through the entire program period), there were no livelihood 
components. This was particularly true for participants with exposure to other food aid 
programs. Enhanced communications, particularly with Mother Leaders and other village
59
opinion leaders, might have overcome some of these unmet expectation-related 
challenges.
Of note, for the home gardens to be successful, communities must purchase non￾hybridized seeds to ensure seed saving opportunities.
 Endorse the PM2A approach (predecessor to the first 1,000 days). PROCOMIDA added 
to evidence from similar programs, confirming that nutrition-specific interventions should: 
i) target women and children in the first 1,000 days; ii) deliver preventive nutrition-specific 
interventions; and iii) emphasize promotion and support of IYCF practices through BCC 
strategies based on formative research. Disclosure and dissemination of IFPRI research 
findings is a key step for further consolidation of PM2A. Given evidence already gathered 
on timing and duration of this approach, the details of which type of ration and food should 
be delivered is key to reaching greater efficiency and affordability.
 Capacity strengthening interventions for community members, inclusive of a Household 
Action Plan, home garden promotion, and cooking demonstrations, should be 
implemented throughout program periods. Trainings and education should not only target 
mothers, but fathers, in-laws, and other caregivers, to promote across households the 
concept of responsible caregiving and parenthood, childcare, health, and nutrition. Special 
attention should be made to ensure that training materials are linguistically and literacy￾level compatible with target populations. Capacity strengthening for local public healthcare 
service units should focus on deeper involvement of mid-level administrators and 
healthcare service providers.
60
REFERENCES
Black R.E. et al. (2013). Maternal and child undernutrition and overweight in low-income and
middle-income countries. Maternal and child nutrition series. Lancet, 382:9890, 427-451.
Black, R.E., et.al. (2008). Maternal and child undernutrition: global and regional exposures
and health consequences. Lancet, 371(9608), 243-60.
Cabrera, M., et.al. (2015). Fiscal policy, inequality, and ethnic divide in Guatemala. CGD
Working Paper No. 397, 22-23. (http://www.cgdev.org/sites/default/files/CGD-Working-Paper￾397-Cabrera-Lustig-Moran-Fiscal-Policy-Inequality-Ethnic-Divide-Guatemala.pdf)
CONEVAL (2010). Dimensiones de la seguridad alimentaria: Evaluación Estratégica de 
Nutrición y Abasto. Mexico DF 
(http://www.coneval.gob.mx/rw/resource/coneval/info_public/PDF_PUBLICACIONES/Dimensi
ones_seguridad_alimentaria_FINAL_web.pdfhttp://www.coneval.gob.mx/rw/resource/coneval/
info_public/PDF_PUBLICACIONES/Dimensiones_seguridad_alimentaria_FINAL_web.pdf)
Del Ángel-Pérez, A., J. Villagómez-Cortés (2014). Food, health and poverty in urban marginal 
areas: a case study in Veracruz-Boca del Rio, Mexico. Estud. Soc,
vol.22 no.44, México jul./dic (http://www.scielo.org.mx/scielo.php?pid=S0188-
45572014000200001&script=sci_arttext).
FAFSA-2 (2013). Second food aid and food security assessment. Washington DC, FHI 360.
Fanta-2 Project (2009). Alternative sampling designs for emergency settings: A guide for
survey planning, data collection and analysis. FANTA-2, Washington DC, Academy for
Educational Development.
FAO (2014). Measurement of dietary diversity for monitoring the impact of food based 
approaches. In Thompson, B., and L. Amoroso (Eds), Improving diets and nutrition: Food 
based approaches, 284-289. Rome: FAO.
FEWSNET. Food security outlook 2011, 2011. Guatemala.
MFEWS, FAO, USAID (2009). Guatemala: Perfil de Medios de Vida.
Moratorell, R. (2012). Interventions and policy options for combating malnutrition in
Guatemala. Inter-American Development Bank. (www.iadb.orghttp://www.iadb.org)
MSPAS, INE, CDC (2010). Encuesta nacional de salud materno-infantil 2008 (ENSMI￾2008/2009). Guatemala.
Oxfam (2012). Evaluación rápida. Impacto de la sequía meteorológica de 2012 en la
seguridad alimentaria nutricional en el corredor seco de Guatemala. Guatemala.
Palmieri, M., H.L. Delgado (2010). Análisis situacional de la malnutrición en Guatemala: sus
causas y abordaje. Programa de las Naciones Unidas para el Desarrollo, Guatemala.
Richter, S., et.al., (2010). Strengthening and evaluating the MP2A Approach in Guatemala: 
Cross-sectional baseline report PROCOMIDA. Mercy Corps, Guatemala.
Swindale, A. (2004). The impact of Title II MCHN programs on the nutritional status of 
children USAID office Food for Peace. Occasional Paper No. 4.
UN (2014). Guatemala: Sequía por canícula prolongada 2014. UN. Oficina de la 
coordinadora residente reporte de situación No. 01, Guatemala.
61
USAID (1000). Fanta Sampling Guide. 
(http://pdf.usaid.gov/pdf_docs/Pnacg172.pdfhttp://pdf.usaid.gov/pdf_docs/Pnacg172.pdf)
WFP (2011). Evaluación de la seguridad alimentaria de los pueblos afrodescendientes, en 
las comunidades de las yungas de La Paz. (http://es.wfp.org/sites/default/files/pma￾afrobolivianosyyungas.pdf)
World Bank (2009). Guatemala Poverty Assessment. WB Report No.43920-GT, Washington 
DC. 
(http://siteresources.worldbank.org/INTLACREGTOPPOVANA/Resources/GuatemalaPoverty
AssessmentEnglish.pdf)
Additional PROCOMIDA documents reviewed:
 All ARRs
 All PREPs
 PROCOMIDA Proposal
 Midterm Evaluation
 BCC Strategy
 Formative Research Study (Baseline)
62
APPENDICES
Annex No. Title Page
1 Scope’s of Work (A, 2014; B, 2016) 62
2 List of PROCOMIDA staff and key stakeholders interviewed 86
3 Methodological aspects of the Impact Estimation 87
4 PROCOMIDA Results Framework and Nutritional Indicators 90
5 Survey Methodology and Sampling 95
6 Quantitative Survey Questionnaires 103
7 Qualitative Questionnaires 146
8 Data Collection Agenda 154
9 List of PROCOMIDA Key Messages 157