A retrospective review of the Honduras AIN-C program guided by a community health worker performance logic model

RESEARCH Open Access
A retrospective review of the Honduras
AIN-C program guided by a community
health worker performance logic model
Daniela C. Rodríguez1*
and Lauren A. Peterson2
Abstract
Background: Factors that influence performance of community health workers (CHWs) delivering health services are
not well understood. A recent logic model proposed categories of support from both health sector and communities
that influence CHW performance and program outcomes. This logic model has been used to review a growth
monitoring program delivered by CHWs in Honduras, known as Atención Integral a la Niñez en la Comunidad (AIN-C).
Methods: A retrospective review of AIN-C was conducted through a document desk review and supplemented with
in-depth interviews. Documents were systematically coded using the categories from the logic model, and gaps were
addressed through interviews. Authors reviewed coded data for each category to analyze program details and outcomes
as well as identify potential issues and gaps in the logic model.
Results: Categories from the logic model were inconsistently represented, with more information available for health
sector than community. Context and input activities were not well documented. Information on health sector systems-
level activities was available for governance but limited for other categories, while not much was found for community
systems-level activities. Most available information focused on program-level activities with substantial data on technical
support. Output, outcome, and impact data were drawn from various resources and suggest mixed results of AIN-C on
indicators of interest.
Conclusions: Assessing CHW performance through a desk review left gaps that could not be addressed about the
relationship of activities and performance. There were critical characteristics of program design that made it contextually
appropriate; however, it was difficult to identify clear links between AIN-C and malnutrition indicators. Regarding the
logic model, several categories were too broad (e.g., technical support, context) and some aspects of AIN-C did not fit
neatly in logic model categories (e.g., political commitment, equity, flexibility in implementation). The CHW performance
logic model has potential as a tool for program planning and evaluation but would benefit from additional supporting
tools and materials to facilitate and operationalize its use.
Keywords: Community health workers, Performance, Community-based program, Malnutrition, Honduras
Background
Given the existing and increasing efforts to deliver health
services through community health workers (CHWs),
critical questions remain about their performance and how
to improve it to reach the greatest health gains possible.
Many studies on CHW performance have focused on
aspects of service delivery or health outcomes [1–3], but not
on the factors that specifically influence performance. A re-
cent evidence summit on CHW performance hosted by the
US government concluded that while it was plausible that
support, both from health systems and communities, would
positively influence CHW performance, the relationship
between these is not well understood because research on
potential support activities (individually or in combination)
have not been frequently or adequately investigated [4].
Community and health system support activities with
potential to influence performance identified at the
Summit, such as local health committees, community
* Correspondence: drodri17@jhu.edu
1
Department of International Health, Johns Hopkins Bloomberg School of
Public Health, 615 N. Wolfe Street, Rm. E-8612, Baltimore, MD 21205, USA
Full list of author information is available at the end of the article
© 2016 Rodríguez and Peterson. Open Access This article is distributed under the terms of the Creative Commons Attribution
4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Rodríguez and Peterson Human Resources for Health (2016) 14:19
DOI 10.1186/s12960-016-0115-x

participation, drug availability, and support from govern-
ment entities, reflect those under study elsewhere [5–7].
In fact, two studies are specifically focusing on
performance-related interventions. Kallander et al. have
developed a protocol to test two intervention packages
to improve CHW performance and retention in Uganda
and Mozambique: (i) an mHealth package with commu-
nication, motivational messages, and phone-based data
and supervision activities, and (ii) a community engage-
ment package with village health clubs conducting activ-
ities to improve CHW status, standing, and demand for
services [7]. Vareilles and colleagues are conducting a
realist evaluation to identify factors influencing perform-
ance of community health volunteers in an immunization
program in Uganda [8].
Building on the work from the Summit [4], Naimoli and
colleagues developed a generic logic model for CHW per-
formance that incorporates multiple dimensions (Fig. 1)
[9]. At the center of the model, three measures of CHW
performance results are highlighted: outputs of CHW-
level change; client, community, and health systems out-
comes attributable to CHWs; and population-level health
impacts attributable to CHWs. These performance mea-
sures are surrounded and driven by program-level activ-
ities by actors in the health sector and the community,
which are in turn affected by systems-level activities from
both health and community systems. The overall pro-
cesses are underpinned by inputs and contextual factors.
The objective of this study was to conduct a retro-
spective review of a CHW program using the generic
logic model for CHW performance and identify the
factors contributing to the program’s success in improv-
ing health outcomes. The program under review is a
national community-based health and nutrition program
focused on growth monitoring in Honduras known as
Atención Integral a la Niñez en la Comunidad or
Integrated Child Health Program in the Community
(AIN-C), which is described in detail below.
AIN-C program in Honduras
In the late 1980s/early 1990s, the Ministry of Health
(MOH) in Honduras suspected that persistent malnutri-
tion was a key factor in static mortality rates and devel-
oped the Integrated Care of the Child (AIN) program to
detect growth faltering in health facilities. A review con-
ducted in 1994 suggested that services should be taken
beyond the facility level [10, 11]. A community-based
approach (AIN-C) was piloted and determined to be the
best way to reach rural families. From 1995 through
2005, USAID’s BASICS program supported the develop-
ment and expansion of the AIN-C program. Figure 2
outlines the evolution of support provided to AIN-C
from the BASICS program as well as the post-BASICS
period.
AIN-C was targeted at children under five and was sup-
ported by volunteer CHWs known as monitoras who
conducted community-based growth monitoring program
(GMP) to detect faltering early and promote feeding
strategies. Monitoras were selected by the community.
The program had an unusual design in that it employed
Fig. 1 CHW performance logic model. Source [9], reprinted with permission of authors
Rodríguez and Peterson Human Resources for Health (2016) 14:19 Page 2 of 11

teams of monitoras to share responsibilities in each com-
munity at any given time [10]. The number of monitoras
on a team varied between two and five, but three was the
average [12], and teams were required to have at least one
female member and one literate member in order for the
team to perform effectively (INT 07.30.14). These arrange-
ments were intended to manage CHW turnover.
The monitoras provided monthly monitoring for chil-
dren under two and followed up with households that
missed monthly monitoring appointments [12, 13]. They
also provided counseling on child nutrition, care of com-
mon illnesses, and referrals to health center nurses for
children under five with counseling cards developed
after extensive formative research [10, 12, 13]. The pro-
gram was designed to be flexible in implementation and
monitoras were allowed to hold growth monitoring ses-
sions in the format most suitable to their community
(e.g., one session monthly, four weekly sessions per
month). The program was also designed to encourage
active community participation and evidence-based
decisions with monitora teams engaging communities in
discussion about key issues that had implications for the
nutrition and health status of local children [13].
Monitoring sessions involved weighing each child,
tracking their growth on a simple ledger, and counseling
caregivers as needed. The data collected were to be used
(i) by the monitoras to trigger dialogue with caregivers
and inform individualized counseling with counseling
cards, (ii) by the community to measure progress and
identify impediments to growth, and (iii) by health
system actors to measure outcomes and improve the
program [11].
Monitoras were supervised directly by the health cen-
ter nurse auxiliary during growth monitoring sessions,
and as monitoras built knowledge and skills, they were
given increased independence and responsibility. Health
sector nurse supervisors also visited but more infre-
quently. Further, monthly meetings were held at the
health center with other community volunteers to re-
view progress, receive training, and restock medicines
and supplies [14].
AIN-C was designed to roll out to 60 communities per
year, with all communities covered nationwide within
6 years [12]. Six disadvantaged departments were
targeted at first: Comayagua, Copán, Intibucá, La Paz,
Lempira, and Ocotepeque.
A midterm evaluation by BASICS in 2000 suggested
that AIN-C communities were more likely to know
about the program, participate in growth monitoring,
and attend weighing sessions consistently than control
communities [15]. High rates of participation and at-
tendance continued through the final evaluation; how-
ever, awareness that inadequate weight gain was a sign
of poor growth was not significantly higher [11]. A cost
analysis estimated that AIN-C had a long-term annual,
recurrent cost per child under five of $2.73, and the
average direct cost per child of an AIN-C session was
approximately 11 % of the direct cost of a single MOH
facility-based consultation [12, 14].
Methods
This study was conducted as a retrospective documentary
review, which was complemented by in-depth interviews
with knowledgeable respondents. A total of 30 documents
Fig. 2 AIN-C program history. Source [10, 14]; INT 07.30.14; INT 08.13.14; INT 08.15.14

were obtained, including World Bank and USAID reports,
MOH documents and policies, national health surveys,
and presentations (see Additional file 1). Documents were
obtained through USAID contacts, web searches, and
documents shared directly by interviewees.
Documents were coded systematically using a code-
book based on the CHW performance logic model. A
code for each logic model component was defined and
applied across documents. An additional code of “Other”
was used to code any sections of text that did not fit the
predetermined categories. At the end of coding, study
authors reviewed these portions of text together to de-
termine whether an existing code could be used. The
documents were coded using qualitative software NVivo
(QSR International).
Four in-depth interviews were conducted with indi-
viduals intimately familiar with the development and
implementation of the AIN-C program, including
former MOH officials, USAID personnel, and contrac-
tors. The first interviewee was identified by contacts
at USAID, and she then suggested additional inter-
viewees to contact. One interview was conducted face-
to-face and three were telephone-based. Interviews
lasted between 30 and 60 min. The telephone inter-
views were recorded, but not transcribed, and exten-
sive notes were taken for all interviews. The
interviews were unstructured and were used to fill
gaps in knowledge from the document review and
confirm preliminary analyses. Of the total 30 docu-
ments reviewed, seven were provided by interviewees.
Analysis was iterative. Coded output was reviewed by
code and across logic model categories to identify factors
contributing to the success of the AIN-C program as
well as critical areas that presented challenges. Gaps in
timeline or understanding as well as issues of clarity
were explored with respondents, and analyses were sub-
sequently refined.
No ethics review was sought for this project as it was
a desk review from existing documentation, with inter-
viewees reflecting on prior work experiences.
Results
Most of the available evidence focused on the period
during which BASICS supported the AIN-C program,
with minimal publicly available information about the
current status of AIN-C. Current policy documents from
the MOH [16–18] mention AIN-C as a program requir-
ing continued commitment but there were few specifics
about the current program, how it is being implemented
or how it is being financially supported. Thus, the results
below focus on the BASICS-supported period of AIN-C
(1995–2005) organized by the categories of the CHW
performance logic model.
Context
Prior to AIN-C, there had been limited improvements in
human development in Honduras. Gains had been made
in education and health with little progress in malnutri-
tion [16]. In an effort to improve equity, it was decided
that AIN-C would target communities with the greatest
need: (i) those with higher prevalence of acute respira-
tory infection (ARI) and diarrhea in children under five,
(ii) those where chronic malnutrition was a persistent
problem, and (iii) rural, disadvantaged, primarily indi-
genous populations [10, 15].
Other broader contextual factors fed into the develop-
ment of the program. First, Honduras had a history of
volunteerism, especially for health, which provided a
backbone for establishing a volunteer cadre of health
workers [12]. Second, despite taking place after AIN-C
started, the widespread destruction from Hurricane
Mitch in 1998 shifted how and where AIN-C was rolled
out and expanded. Lastly, decentralization to municipal-
ities took place during the rollout of AIN-C, though it is
unclear what role this process played during the pro-
gram’s implementation.
Inputs
The right to health and food had been established in the
constitution of Honduras, providing a legal expectation
for supporting health and well-being [16]. Policy docu-
ments also cite other global-level commitments Honduras
made to address hunger and its consequences, such as the
World Food Summit and the Millennium Development
Goals, as further basis for addressing malnutrition [16].
Unfortunately, the document review did not yield any in-
formation about financing, facilities, materials, equipment,
or policies/guidelines for the implementation of AIN-C.
System-level activities
Health system
Under governance, the MOH was committed to increas-
ing coverage and quality of care, and empowering and
incentivizing communities towards social control of
nutrition programs [10, 16]. AIN-C was seen by policy-
makers and implementers at the time as a public
demonstration of this commitment. Also, AIN-C was
integrated with poverty reduction policies and the over-
all approach called for intersectoral coordination [16].
There was evidence of leadership from high-level gov-
ernment offices (e.g., Presidency, MOH), and high-level
policies were put in place that supported AIN-C goals
and integration with national child health programs [16];
however, there was more heated discussion about adding
curative care components to the package of services
delivered by monitoras ([12] and INT 08.15.14). Lastly,
MOH decrees regarding AIN-C supported standardized

implementation even when implemented by NGOs [11],
but it was not clear who would monitor compliance.
In terms of service delivery, there were commitments
and plans for extending the health system’s reach—again
demonstrated by AIN-C. Also, CHWs appeared to be
well integrated into the health system through their
monthly visits to the health center for monitoring,
supervision, and resupply.
Although there was considerable information on pro-
gram costs, there was no information on program finan-
cing including funding sources and flows and timeliness.
Likewise, there was no clear information on health work-
force or any investment in MOH staff tasked with super-
vising monitoras and the program.
Regarding information activities, data at the individual
level were collected in a simple, useful manner for moni-
toras, but it is not clear whether and how data were used
by the MOH, either at the health facility level or higher up
([11]; INT 08.15.14). Monitoras were supposed to resupply
with medical products and supplies at the health center
during monthly supervision visits, but there was no infor-
mation available on how well the supply chain was
working.
Community
System-level activities from the community were not well
represented in the documents, especially governance/lead-
ership and social belonging/cohesion. There was limited in-
formation about active resource mobilization to address
community issues that influence child growth, but it was
anecdotal. Documents note that some community meet-
ings facilitated by monitora teams to discuss growth moni-
toring data resulted in actions such as addressing
contaminated water sources or trash sites, providing child
care during busy times, improving indoor air pollution,
and facilitating health center outreach [13].
Program-level activities
Technical support
Health system Most of the documentary sources were
focused on program design and implementation. They
highlighted several important features:
Gradual rollout of AIN-C
Formative research informed the program [10]:
o Earlier experiences in Honduras with AIN
o Evidence from other contexts (e.g., World Bank
review of GMPs) [19]
o In-country formative research (e.g., messages for
counseling cards) [20]
o Pilot testing
There was an emphasis on strengthening initial
training protocols, but investments in sustained
training were unclear
Supervision was operationalized to be standard
throughout MOH facilities, but NGO implementers
provided more supportive supervision
There was limited information about the ongoing
monitoring and evaluation of the program. Few, if any,
health centers used growth monitoring data to support
monitoras or the community in their decision-making.
Community
There was no substantive information about technical
support activities from the beneficiary communities.
Social support
Health system AIN-C was designed to focus on the
community’s capacity and responsibility to ensure that
its children are fed and growing [10]. Various relation-
ships necessary to support the program were facilitated
by health sector actors, such as media advertising the
program [10], and through integration of the monitoras
and AIN-C into the health system, including defining
roles and responsibilities [14]. However, there was no in-
formation about linkages between monitoras and other
networks that could have supported them and AIN-C.
Community In order to ensure community involvement
and support, communities were advised that they needed
to show their commitment by agreeing to join AIN-C
and vest themselves in the program. This process was fa-
cilitated by engaging in conversations with communities
and elders [15]. Further, they were tasked with selecting
monitoras, and communities were given ownership over
AIN-C materials (INT 07.30.14). It is not clear how well
the quarterly community meetings were implemented,
though, as noted above, there was anecdotal evidence of
participatory decision-making leading to community
actions [13].
Incentives
Health system Program planners did not want to ex-
ceed the limits of the inherent volunteerism of moni-
toras, so health system incentives were planned for and
operationalized [10]. Incentives from the health sector
included identification cards, diplomas, carrying bags,
letters of recognition/thanks from the Regional Health
Office, yearly party/dinner, Children’s Day piñata parties,
and preferential access to care at MOH facilities [12].

Community There was some evidence that monitoras
garnered stature and respect in the community due to
their role, but this was not widely documented [14]. An-
ecdotal evidence suggests that monitoras became com-
munity leaders and advocates, especially after Hurricane
Mitch when they helped mobilize communities to iden-
tify their needs and advocate to have them met (INT
07.30.14; INT 08.15.14; INT 09.10.14).
Performance results
Outputs for CHW performance
An implementation review was conducted by BASICS
after five full years of implementation [21]. Results of
monitora performance indicated that:
17 % were making regular classifications errors
20 % were not counseling children with inadequate
growth
60 % provided counseling with quality errors. Of
these, 50 % were not using counseling cards
correctly or at all
Challenges for counseling centered on correctly identi-
fying specific problems linked to growth faltering and
providing tailored advice. Monitoras were giving several,
general recommendations to improve feeding practices
instead of one or two targeted messages. Table 1 cap-
tures results from the implementation review regarding
other issues affecting monitora performance.
Outcomes for client, community and health system change
A midterm evaluation of the AIN-C program was con-
ducted in 2000 through a household survey that com-
pared AIN-C communities with control communities
served by the same health facilities [15]. A final evalu-
ation was conducted in 2005 but was unable to use the
same sample as earlier surveys due to issues of contam-
ination of control communities and reduced implemen-
tation intensity [11]. Instead, the final evaluation took an
individual-level approach to understand the impact of
AIN-C by comparing children who participated in the
program with those that did not, regardless of their
community [11]. The evaluations explored topics around
knowledge of and participation in a GMP, and know-
ledge, attitudes, and practices (KAP) at the household
level. Table 2 shows results for the midterm evaluation
comparing control and AIN-C communities, and results
from the final evaluation comparing children in AIN-C
and those not in a GMP program, as necessitated by the
revised sample.
At midterm, caregivers in AIN-C communities showed
improvements in their knowledge and practice despite
poorer living conditions overall [15]. Participation in a
GMP program was very high in communities targeted
for AIN-C, with caregivers attending 70 % of weighing
sessions regularly. In terms of KAP related to growth
and feeding, AIN-C communities showed improvements
over the control communities in many areas; however,
very few caregivers in either group recognized inad-
equate weight gain as poor growth. Results for caregiv-
ing practices during illness were more mixed with
limited gains in both groups. The final evaluation sug-
gests stable rates of awareness and participation in the
GMP program [11]. There were considerable improve-
ments of KAP around growth and feeding and for care-
seeking for diarrhea and ARI for both the AIN-C and
No GMP groups. In the end, participation intensity was
related to improvements in malnutrition: with every 1 %
increase in participation, weight-for-age 0.005 z-score
improved [11].
Impact
National rates of malnutrition were already declining
prior to the initiation of AIN-C, but the trend becomes
more marked after its introduction (Fig. 3). However, de-
creasing national rates of malnutrition mask significant
differences at the subnational level. While AIN-C de-
partments have shown declines in malnutrition since
2001, their rates are still substantially higher than the
national average (Fig. 4). These changes took place dur-
ing a period of overall declines in population growth,
fertility and child mortality, and improvements in life ex-
pectancy [22]. Management of malnutrition through
GMPs is particularly challenging so it is difficult to
ascribe success or failure directly to AIN-C but it ap-
pears that the program may have contributed to declines
in malnutrition rates even though its direct effect cannot
be measured.
Discussion
This article describes a GMP delivered by CHWs in
Honduras through the categories of the CHW perform-
ance logic model described by Naimoli et al. [9]. Results
on the success of the program itself were mixed. An
assessment of CHW performance indicated that many
monitoras were providing effective counseling, but issues
around classification errors, missed opportunities for
counseling, and challenges in counseling quality were
identified. In terms of health outcomes, evaluations of
AIN-C found improvements in knowledge and caregiv-
ing practices and, most importantly, in malnutrition for
children exposed to GMP with regular participation.
However, due to the limitations of data available during
this review, we were unable to draw direct linkages be-
tween components of the AIN-C program and positive
health outcomes as intended. Below, we explore charac-
teristics of AIN-C design that allowed the program to be
contextually appropriate, reflect on the logic model

itself, and raise considerations for future applications of
the model.
In Table 3, we outline a number of characteristics of
how AIN-C was designed to make the program better
suited to the local community and health systems con-
text, which can provide lessons to others designing simi-
lar programs. These have been broken down into three
overarching categories: the content of the intervention,
the delivery mechanism, and the mechanisms in place to
support delivery.
First, in terms of content, evidence was used by pro-
gram planners to design AIN-C to avoid earlier pitfalls.
The intervention was designed to focus on limited infor-
mation per visit, with regular follow-up of participants.
Second, there are several design characteristics related
to the delivery of AIN-C worth noting. AIN-C was
targeted at the communities who were the worst-off in
order to realize the most gains and address equity.
Monitoras were established as a team to allow CHWs to
share the workload and reduce the likelihood of program
collapse due to turnover. In order to avoid over-relying
on the inherent volunteerism of community members,
health sector incentives for monitoras were established.
Further, flexibility in implementation and plans for
information sharing allowed AIN-C delivery to be re-
sponsive to community needs.
Lastly, characteristics around supportive mechanisms
covered both community and broader systems supports.
On the community side, AIN-C communities had to
agree to three main responsibilities: agreeing to become
an AIN-C site, selecting monitoras, and holding commu-
nity meetings. CHWs were linked with the health system
through training, supervision, monitoring, and health re-
ferrals. Further, the strong government commitment to
the overall efforts to address equity and improve health
and well-being suggest high-level commitment to AIN-
C’s goals.
Reflections on the CHW performance logic model
First, we reflect on the ease of use of the CHW perform-
ance logic model and potential improvements to con-
sider based on our experience and then address the
potential uses suggested by the logic model’s authors.
Most monitoring and evaluation (ME) efforts asses-
sing CHW programs are not designed to assess CHW
performance at the core of their activities or as a driver
for the program’s theory of change. The CHW perform-
ance logic model is useful in helping reorient ME
approaches to focus more clearly on the intersection
between CHW performance and program- and system-
level activities. Consequently, our most important learn-
ing from using this logic model to evaluate AIN-C is
that it is critical to start any evaluation with evidence on
CHW performance as a way to provide a more focused
approach to reviewing the program. While we were able
to use the logic model to identify critical components of
the program’s design, we were unable to identify factors
contributing to programmatic success as we had origin-
ally intended.
Table 1 Implementation review findings on monitora performance
Performance
category
Findings
Retention • Average length of service for monitoras was 2.5 years.
• 25 % of the original cohort was still working after 5 years.
• Monitoras moved in and out of the program, which was facilitated by the team approach.
Motivation • Active participation of beneficiary families was critical.
• One third of monitoras noted the lack of family support.
Training • Each community had at least one monitora who had participated in the original training.
• However, 60 % of monitoras were learning by doing.
• Monitoras in MOH communities received training on AIN-C and case management, while monitoras in NGO communities
received additional training modules.
Supervision • The content and quality of supervision varied.
• Supervision was mainly focused on monthly health center meetings, but in NGO communities monitoras received additional
supervision.
Supplies • No stock-outs of basic materials were noted.
• 90 % of scales used for weighing were accurate.
Data use • 85 % of the child lists tracking children in the community were good.
• Quality of progress bars tracking attendance and growth faltering depended on the quality of the child lists.
• There was little use of bar charts by MOH for decision-making.
Community action • Implementation was not uniform.
• Community action depended on support from outside the community, with communities receiving support from health
center promoters doing better.
• Determined that about 20 % of causes for growth faltering that needed attention were issues outside the family.
Source [21]

Also, we identified several issues regarding the current
composition of the logic model to consider for future it-
erations. First, the model as represented makes it appear
that each component is equally weighted, when in reality
this may be context-specific, which the authors acknow-
ledge. Second, the technical support category is very
broad and encompasses activities for multiple program-
matic stages (e.g., design, implementation, evaluation),
both for the health sector and community, resulting in a
complex category that is unwieldy. Smaller categories
would help clarify the potential roles of different actors
to identify and address gaps, and strengthen the pro-
gram. Similarly, the inputs and context categories are
broad and include many complex components, such as
policies, funding, organizations, which have likely impli-
cations for program success. In fact, a recent review
proposed an approach stressing the critical pathways
through which contextual factors influence CHW
performance [23]. Third, key aspects of the AIN-C
program beyond CHW performance were hard to locate
in the logic model, such as government and political
commitment, cost of the program, flexibility in
implementation as a design feature, and commitments
to equity, which highlight the importance of factors ex-
ternal to CHWs in supporting both their performance
and ultimate outcomes.
Potential uses of the CHW performance logic model
The logic model authors suggested four potential uses
for the CHW performance logic model, each with their
own considerations, which we address in turn. For plan-
ning, the logic model may be too comprehensive with
many components to focus on, which could overwhelm
policymakers. It would be helpful to identify which are
the key categories to focus on at the outset of planning a
program, or provide guidance for a facilitated planning
process with the logic model as its basis. For practical
purposes, the logic model could be used to (i) assess the
current programmatic landscape and the potential
contributions of a new intervention delivered by CHWs,
(ii) explore how program- and system-level activities
may support or hinder the CHW program and meeting
its goals, and (iii) for careful reflection on community
contributions.
Table 2 AIN-C midterm and final evaluation results
Baseline (1998) Midterm (2000) Final (2005)
Control
(%)
AIN-C
(%)
Control
(%)
AIN-C
(%)
No GMP
(%)
AIN-C
(%)
Child growth monitoring and promotion program awareness/participation
Caregivers know about the GMP program in their community 7 27 15 96a
– 100
Caregivers participate in the GMP program in their community 21 30 23 92a
– –
Enrollment in GMP program within first month of life – – 27 28 – 24
Caregiver has a growth card for child with at least two weight measurements 64 59 68 91a
– 93
Attend weighing session 3 or more time in past 3 months 38 30 44 70a
– 67
Caregiver received counseling for child with at least one instance of growth
faltering on their growth card
– – 57 81a
– 81
Caregiver recognition of counseling cards – – 31 64a
45 73
KAP around growth and feeding
Exclusive breastfeeding of children under 6 months of age 15 21 13 39a
40 56b
Caregivers has their children 4 months of age or older take iron supplements 4 2 4 47a
30 66b
Caregiver aware that weight gain is sign of good growth 36 38 30 50a
33 51b
Caregiver aware that child being underweight is sign of poor growth 43 47 37 45a
41 48
KAP around illness
Child is fully immunized by the age of 12 months 65 62 66 76a
71 77
Gave oral rehydration therapy to child with diarrhea 36 37 42 57a
38 62b
Gave child fluids and continued feeding during a bout of diarrhea 17 21 16 33a
70 82b
Child experienced in episode of diarrhea in past 2 weeks taken to monitora or
health care provider
– – 25 34 41 47
Child who experience episode of ARI in past 2 weeks taken to monitora, pneumonia
volunteer or health care provider
– – 44 36 – –
Source [11, 15]
a
Significant difference between AIN-C and control communities at midterm evaluation
b
Significant difference between AIN-C and No GMP individuals at final evaluation

For consensus building, the logic model can be used
through guided discussions aiming for a coordinated ap-
proach; the caveat above about the number and complexity
of categories would also be a concern here. For program
implementation, the model could be used to inform discus-
sions around prioritizing investments and problem solving.
However, it could be difficult to tease out specific areas to
address without targeted and regular monitoring data. As
for evaluation, the model identifies all the potential categor-
ies necessary to develop a comprehensive evaluation design.
In fact, we believe that Honduras would benefit greatly
from an on-the-ground evaluation of the AIN-C program
using the CHW performance logic model to identify and
address program challenges.
Lastly, we stress an important consideration for these
strategies: communities. A clear role for communities
Fig. 3 Malnutrition prevalence in Honduras, 1987–2012. Source [24]
Fig. 4 Height-for-age below two standard deviations in AIN-C departments, 2001–2011/12. Source [25–27]

would need to be identified a priori when using the logic
model for these strategies to ensure that programs are
planned, built, implemented, and evaluated in a repre-
sentative, transparent, and responsive manner.
Limitations
There are several limitations to this review. First, there
was limited documented information about AIN-C avail-
able in the public domain. Most of the available evidence
focused on the period of BASICS and USAID support,
and minimal publicly available information on the
current status of AIN-C. We attempted to address this
limitation by conducting additional literature review
searches through databases of published research as well
as general web searches, but few additional documents
were identified. Second, for interviewees, there had been
a lag of about 10 years between their participation in
AIN-C and their interview for this project, which may
have limited their recall of details on the program.
Lastly, in terms of the application of the logic model
itself, the core documents about AIN-C contain little
discussion about the community component of the pro-
gram making it difficult to ascertain its contributions
despite the intentions in the design and original rollout.
We are unsure whether the lack of a documented com-
munity role in AIN-C represents poor community in-
volvement or uneven documentation.
Conclusions
This retrospective desk review explored a GMP in
Honduras delivered by CHWs through the lens of a
CHW performance logic model. It identifies lessons to
be learned from the program’s design as well as from the
potential of the logic model itself, which provides a com-
prehensive basis for understanding, planning, and
evaluating CHW programs into the future.
Additional file
Additional file 1: Supplementary file. (PDF 98 kb)
Competing interests
The authors received salary support from the Health Finance and Governance
Project, funded by the United States Agency for International Development, to
conduct this work. The authors declare they have no other competing interests.
Authors’ contributions
DR designed the study. DR and LP conducted the document review abstraction.
DR conducted the supplementary interviews with LP’s support. DR and LP
conducted the analysis, and drafted, edited and approved the final manuscript.
Acknowledgements
We would like to thank Marcia Griffiths, Arturo Gutierrez, Vicky Alvarado, and
Laura Molina for their time and valuable insights into the workings of the
AIN-C program under review, including access to critical documents. Joseph
Naimoli and Diana Frymus from USAID (co-authors on the logic model that
was used to guide this review) provided helpful feedback about the logic
model itself and the differentiations between categories included in the
model. USAID linked the authors to respondents who were knowledgeable
in the AIN-C program, but did not participate in data collection, analysis, or
writing of the manuscript.
Funding
This study was conducted under the United States Agency for International
Development through the Health Finance and Governance Project (Cooperative
Agreement Number OAA-A-12-00080).
Author details
1
Department of International Health, Johns Hopkins Bloomberg School of
Public Health, 615 N. Wolfe Street, Rm. E-8612, Baltimore, MD 21205, USA.
2
Abt Associates, Bethesda, MD, USA.
Received: 4 September 2015 Accepted: 25 April 2016
References
1. Kelly JM, Osamba B, Garg RM, Hamel MJ, Lewis JJ, Rowe SY, Rowe AK,
Deming MS. Community health worker performance in the management of
multiple childhood illnesses: Siaya District, Kenya, 1997–2001. Am J Public
Health. 2001;91:1617–24.
2. Lema IA, Sando D, Magesa L, Machumi L, Mungure E, Sando MM,
Geldsetzer P, Foster D, Kajoka D, Naburi H, et al. Community health workers
to improve antenatal care and PMTCT uptake in Dar es salaam, Tanzania: a
Table 3 Critical characteristics of AIN-C emerging from program design
Characteristic of AIN-C Design categorya
Learning and formative research from earlier experiences informed program design Content
Limit the education messaging per AIN-C visit Content
Regularity of follow-up with program participants Content
Targeting of worse-off communities Delivery
CHWs working as a team sharing the workload Delivery
Culture of volunteerism + operationalizing incentives Delivery
Flexibility in implementation at the community level Delivery
Information sharing up to the health system and down to the community Delivery
Standardized plans for training, supervision and monitoring of CHWs Support
Linkages between CHWs and health system: referrals, other services Support
Community participation for site selection, CHW selection and community meetings Support
Strong government and political commitment to the program Support
a
Design categories: content of the intervention, delivery mechanism, support structures

quantitative performance evaluation. Jaids-J Acquired Immune Deficiency
Syndromes. 2014;67:S195–201.
3. Roberton T, Applegate J, Lefevre AE, Mosha I, Cooper CM, Silverman M,
et al. Initial experiences and innovations in supervising community health
workers for maternal, newborn, and child health in Morogoro region,
Tanzania. Human Resources Health. 2015;13.
4. Naimoli JF, Frymus DE, Quain EE, Roseman EL, Roth R, Boezwinkle J.
Community and formal health system support for enhanced community
health worker performance—a U.S. Government Evidence Summit.
Washington: USAID; 2012.
5. Bagonza J, Kibira SPS, Rutebemberwa E. Performance of community health
workers managing malaria, pneumonia and diarrhoea under the
community case management programme in central Uganda: a cross
sectional study. Malar J. 2014;13.
6. Druetz T, Kadio K, Haddad S, Kouanda S, Ridde V. Do community health
workers perceive mechanisms associated with the success of community
case management of malaria? A qualitative study from Burkina Faso. Soc Sci
Med. 2015;124:232–40.
7. Kallander K, Strachan D, Soremekun S, Hill Z, Lingam R, Tibenderana J, et al.
Evaluating the effect of innovative motivation and supervision approaches on
community health worker performance and retention in Uganda and
Mozambique: study protocol for a randomised controlled trial. Trials. 2015;16.
8. Vareilles G, Pommier J, Kane S, Pictet G, Marchal B. Understanding the
motivation and performance of community health volunteers involved in
the delivery of health programmes in Kampala, Uganda: a realist evaluation
protocol. Bmj Open. 2015;5.
9. Naimoli JF, Frymus DE, Wuliji T, Franco LM, Newsome MH. A Community
health worker “logic model”: towards a theory of enhanced performance in
low- and middle-income countries. Hum Resour Health. 2014;12:56.
10. Griffiths M, McGuire JS. A new dimension for health reform—the integrated
community child health program in Honduras. In: LaForgia GM, editor.
Health Systems Innovations in Central America: Lessons and Impact of New
Approaches. Washington: The World Bank; 2005.
11. Schaetzel T, Griffiths M, Miller Del Rosso J, Plowman B. Evaluation of the
AIN-C Program in Honduras. Arlington: Basic Support for Institutionalizing
Child Survival Project (BASICS II) for USAID; 2008.
12. Fiedler J. A cost analysis of the Honduras community-based integrated child
care program (Atención Integral a la Niñez-Comunitaria, AIN-C). In: Health,
Nutrition and Population Discussion Paper. Washington: World Bank; 2003.
13. Griffiths M, Del Rosso J. Growth monitoring and the promotion of health
young child growth: evidence of effectiveness and potential to prevent
malnutrition. 2007. The Manoff Group.
14. Fiedler JL, Villalobos CA, De Mattos AC. An activity-based cost analysis of
the Honduras community-based, integrated child care (AIN-C) programme.
Health Policy Plan. 2008;23:408–27.
15. Van Roekel K, Plowman B, Griffiths M, Vivas de Alvarado V, Matute J,
Calderón M. BASICS II midterm evaluation of the AIN program in Honduras,
2000. Arlington: Basic Support for Institutionalizing Child Survival Project
(BASICS II) for USAID; 2002.
16. de Salud S. Política Nacional de Nutrición. Honduras: República de
Honduras; 2005.
17. Secretaría de Estado en el Despacho de Salud. Plan Nacional de Salud 2021.
Tegucigalpa: República de Honduras; 2005.
18. de Salud S. Política Nacional de Salud Materno Infantil. Honduras: República
de Honduras; 2005.
19. Griffiths M, Dickin K, Favin M. Promoting the growth of children: what
works—rationale and guidance for programs. In: World Bank Nutrition Toolkit,
vol. 4. Washington: Human Development Dept. - The World Bank; 1996.
20. Marini A, Bassett L, Bortman M, Flores R, Griffiths M, Salazar M. Promocion
del crecimiento para prevenir la desnutrición crónica: Estrategias con Base
Comunitaria en Centro América. Washington: World Bank; 2009.
21. Griffiths M. Implementación: El Vínculo Entra Teoría y los Resultados? 2004.
USAID-BASICS II.
22. Policy H, Unit S. Health systems profile, Honduras: monitoring and
analyzing health systems change/reform. Washington: Pan American
Health Organization; 2009.
23. Kok MC, Kane SS, Tulloch O, Ormel H, Theobald S, Dieleman M, Taegtmeyer M,
Broerse JE, de Koning KA. How does context influence performance of
community health workers in low- and middle-income countries? Evidence
from the literature. Health Res Policy Syst. 2015;13:13.
24. World Development Indicators Database. The World Bank.
25. Corrales G, Pavón S, Enamorado R, Stupp P, Jefferds M, Grummer-Strawn L,
McCracken S, Ballentine J, Goodwin M, Monteith R. Honduras Encuesta
Nacional de Epidemiología y Salud Familiar/Encuesta Nacional de Salud
Masculina 2001 - Informe Final. Honduras: Secretaría de Salud; 2002.
26. Secretaría de Salud [Honduras], Instituto Nacional de Estadística, Macro
International. Encuesta Nacional de Salud y Demografía 2005–2006.
Tegucigalpa: SS, INE y Macro International; 2006.
27. Secretaría de Salud [Honduras], Instituto Nacional de Estadística, ICF International.
Encuesta Nacional de Salud y Demografía 2011–2012. Tegucigalpa: SS, INE e ICF
International; 2013.
• We accept pre-submission inquiries
• Our selector tool helps you to ﬁnd the most relevant journal
• We provide round the clock customer support
• Convenient online submission
• Thorough peer review
• Inclusion in PubMed and all major indexing services
• Maximum visibility for your research
Submit your manuscript at
www.biomedcentral.com/submit
Submit your next manuscript to BioMed Central
and we will help you at every step:

A retrospective review of the Honduras AIN-C program guided by a community health worker performance logic model

More Related Content

What's hot (20)

Viewers also liked (14)

Similar to A retrospective review of the Honduras AIN-C program guided by a community health worker performance logic model (20)

More from HFG Project (20)

Recently uploaded (20)

A retrospective review of the Honduras AIN-C program guided by a community health worker performance logic model