The process of developing a nutrition sensitive agriculture intervention a multi site experience

Food Sec.
DOI 10.1007/s12571-016-0625-3

ORIGINAL PAPER

The process of developing a nutrition-sensitive agriculture
intervention: a multi-site experience
Peter R. Berti 1 & Rachelle E. Desrochers 1 & Hoi Pham Van 2 & An Lê Văn 3 &
Tung Duc Ngo 3 & Ky Hoang The 4 & Nga Le Thi 4 & Prasit Wangpakapattanawong 5,6

Received: 7 September 2015 / Accepted: 10 October 2016
# Springer Science+Business Media Dordrecht and International Society for Plant Pathology 2016

Abstract Nutrition-sensitive agriculture (NSA) interventions are of increasing interest to those working
in global health and nutrition. However NSA is a
broad concept, and there are numerous candidate
NSA interventions that could be implemented in any
given setting. While most agriculture interventions
can be made Bnutrition-sensitive^, there are few
guidelines for helping to decide what agriculture component should be tried in an NSA intervention. Based
on previous models, we developed a framework with
explicit questions about community factors (agricultural production, diets, power and gender), project

Electronic supplementary material The online version of this article
(doi:10.1007/s12571-016-0625-3) contains supplementary material,
which is available to authorized users.
* Prasit Wangpakapattanawong

1

HealthBridge, 1004-One Nicholas Street, Ottawa, Canada

2

Center for Agricultural Research and Ecological Studies, Vietnam
National University of Agriculture, Hanoi, Vietnam

3

Hue University of Agriculture and Forestry, 102 Phung Hung,
Hue, Vietnam

4

HealthBridge, Suite 202 & 203, E4 Building, Diplomatic
Compound, No. 6 Dang Van Ngu Str., Dong Da Dist,
Hanoi, Vietnam

5

Department of Biology, Faculty of Science, Chiang Mai University,
Chiang Mai 50200, Thailand

6

World Agroforestry Centre (ICRAF) c/o Knowledge Support Center
for the Greater Mekong Sub-region (KSC-GMS), Faculty of Social
Sciences, Chiang Mai University, P.O. Box 267, CMU Post Office,
Chiang Mai 50202, Thailand

factors (team capacity, budget, timelines) and external

factors that helped our team of agriculture scientists,
nutritionists and local officials identify NSA interventions that may be feasibly implemented with a reasonable chance of having positive agricultural and nutritional impacts. We applied this framework to two settings in upland Vietnam, and one setting in upland
Thailand. From an initial list of nineteen interventions
that have been tried elsewhere, or may reasonably be
expected to be appropriate for NSA, five or six candidate interventions were chosen per site. Based on
the criteria, three to four interventions were selected
per site and are being implemented. Poultry rearing
and home gardening were selected in each site. They
and the other selected interventions, hold promise for
capitalizing on underused agricultural potential to improve diets, while working with (or improving)
existing gender relationships and power structures.
The process for identifying NSA interventions was
thorough and identified reasonable candidates, but it
was very time consuming. Further efforts should focus on streamlining the process, so that promising and
appropriate NSA interventions can be identified
quickly and reliably.
Keywords Nutrition-sensitive agriculture . Vietnam .
Thailand . Food security

Introduction
The potential and real contributions of agriculture to
the health and nutrition of low and middle income
country farmers is of increasing interest to development donors, practitioners and researchers through


Berti P.R. et al.

what has been termed Bnutrition-sensitive agriculture^
(NSA). Instead of considering agriculture solely as an
economic activity, NSA is concerned with the broad

relationship between food production and the health
of the farmer and the farm family. Their relationship
may be expressed in terms of pathways that lead from
agriculture to improved nutrition (Herforth and Harris
2014; Headey et al. 2011; Kadiyala et al. 2014).
These include: a household’s own crop and animal
production to consumption of its own produced food;
increased household income (through increased production or beneficial agricultural policies) allowing
increased purchasing of food and/or health care; adjustment of women’s workload to allow for improved
child care and/or maternal nutritional status; and improved women’s status and ability to control allocation of resources for food, health and care.
These relationships are theoretical and the amount of
supporting evidence for each pathway is variable but
generally limited. There were eight reviews of the impact of agriculture interventions on nutrition outcomes
carried out between 2001 and 2013, summarized in
Webb and Kennedy (2014), but these reviews included
only as few as 13 and at most 52 separate studies, and
collectively they drew on about 100 separate studies.
This is a small number of studies from which to draw
globally applicable lessons. The studies cover different
agricultural foci in various climatic, topographic and
geopolitical environments, with various ethnic groups
and covering different age groups. All the reviews conclude that more research is needed to better understand
the relationship between agricultural interventions and
nutrition and to be able to design NSA interventions
with a high probability of having a positive health outcome. Fortunately, much more research is ongoing now
– a recent review of the research landscape indicated
that there were 151 ongoing NSA studies (Turner
et al. 2013), and probably many more that were not
captured in that review. With all this interest, it is
worthwhile considering how to select a specific NSA

intervention to implement in a given situation.
Guidelines from the FAO direct the user on aspects of
situation appraisal to plan NSA, to define program objectives, to select target groups, to choose interventions
and implementation modalities (FAO 2015).
The focus here is on the development of a framework
specifically for choosing which NSA(s) to trial. In our
case, this meant choosing an agriculture intervention
that the agriculture scientists (HPV, ALV, PW, plus other local team members) thought may be successful from
a production point of view, and the nutritionists (PRB,
NLT, plus other local team members) thought could be
made Bnutrition-sensitive^ and have a positive impact

on nutrition outcomes. The authors represent teams from
agriculture departments in two Vietnamese universities,
and a biology department in a Thai university, as well
as scientists from a health-focused international NGO.
Funding from the International Development Research
Centre (IDRC), Canada from March 2013 to February
2016 supported carrying out research on NSA interventions in three research sites (one for each of the university teams), but the specifics of the interventions were
not pre-defined. A preliminary step for the research
teams was to decide which intervention, or interventions, should be trialled at each site. The Vietnamese
and Thai research teams did not have prior experience
with explicit NSA interventions and some care was
needed to select appropriate interventions to implement
at these sites.
The purpose of the paper is to describe our methods
of determining suitable NSAs to test, and suggest ways
forward for other researchers and programme designers
in a similar situation. Based on previous models, we
developed a framework with explicit questions about

community factors (agricultural production, diets, local
power structure and gender), project factors (team capacity, budget, timelines) and external factors that
helped our team of agriculture scientists, nutritionists
and local officials identify NSA interventions that may
feasibly be implemented with a reasonable chance of
having positive agricultural and nutritional impacts.
Note that we did not carry out a comprehensive needs
assessment to identify the single best strategy for poverty
alleviation or health improvement – perhaps a water and
sanitation intervention, or a maternal health services intervention would have a greater impact. Rather, our research
mandate was to identify and test NSA interventions and
monitor the nutrition impact of the intervention, accepting that
while there are risks in NSA interventions (Dury et al. 2015),
they can have positive health benefits when well done
(Berti et al. 2004; Webb and Kennedy 2014).
Study setting
Malnutrition (General Statistical Office of Vietnam
2011; National Statistical Office of Thailand 2013) and
food security remain serious problems in both Vietnam
(as determined using household food insecurity access
score (Humphries et al. 2015) or whether households
produce enough staple crop (FAO 2004) and Thailand
as determined by access to adequate resources
(Isvilanonda and Bunyasiri 2009), particularly amongst
ethnic minorities living in remote, upland areas
(Government of Vietnam Socialist Republic 2005;
Limnirankul et al. 2015). The national prevalence of
underweight and stunting is 11.7 % and 22.7 %,

The process of developing a nutrition-sensitive agriculture

Map 1 The three study sites, Mae Chaem, Yen Chau, and A Luoi are shown in orange on the right hand panel

respectively, in Vietnam and 9.2 % and 16.3 %, respectively, in Thailand. Rates in both countries are higher in
the poverty-stricken highland-minority communities
(General Statistical Office of Vietnam 2011; National
Statistical Office of Thailand 2013).
The current study was carried out in three locations:
Mae Chaem district of Chiang Mai province in northern
Thailand (with Lua and Karen ethnic groups), district of
Hue province in central Vietnam (with Cotu, Taoi, Paco,
Van Kieu and Kinh ethnic groups), and Yen Chau district of Son La province in north Vietnam (with Kinh,
Thai, Kho Mu, and H’mong ethnic groups) (see Map 1).
Specific villages were selected based on accessibility by
the research team, perceived potential for change, and
representativeness of major ethnic groups. The three
sites have broadly similar climates, with temperatures
hitting lows of 15 to 20 °C in December and highs of
25 to 35 °C in May through July, and the rainfalls
ranging from close to 0 mm in December to as much
as 250 to 300 mm in August in Yen Chau and Mae
C h a e m , o r 1 0 0 0 m m i n A L u o i i n O c t o b e r.
Agriculture practiced in these communities is predominantly subsistence agriculture (or semi-subsistence in the
case of Yen Chau district in northern Vietnam with
60 % of farmers producing their own rice and nearly
all producing their own vegetables).
There is a need in all three locations to identify longterm, sustainable solutions to increase local food availability in these vulnerable areas. With existing high
rates of malnutrition, their food security is threatened
further by transitions taking place in the local farming

systems. Traditionally, farmers in highland areas (more
than 600 m above sea level) have practiced Bshifting
agriculture^, also known as swidden or slash-and-burn
agriculture. The burning of the trees by farmers provides

a nutrient-rich layer of ash that allows for a season or
two of high levels of crop production. Farmers would
then move to new areas to repeat the cycle, while the
burnt area was left to regenerate (Rambo and Cuc 1996;
Wangpakapattanawonga et al. 2010). There is now intense
pressure to transition away from shifting agriculture in
both Vietnam (where it is now illegal) and northern
Thailand as alternative methods of farming are now
being sought because there is insufficient land for the
number of people farming to leave land fallow long
enough to regenerate (Sikor and Nguyen 2011;
Thomas et al. 2004). There is tension between farmers’
needs for high levels of production in the short term,
which makes increasing levels of agriculture inputs
(especially synthetic pesticides and fertilizers) attractive
to the farmers, and the longer term degradation of the
environment caused by this industrialization of agriculture
(Dung et al. 2008; Tuan et al. 2014).

Methods
Development of framework to guide selecting NSA
We developed a framework for assessing the Community,
Project, and External factors listed in the introduction. In order to develop the framework, we drew on existing descriptions of the pathways between agriculture and nutrition
(Headey et al. 2011; Herforth and Harris 2014), and drew on
our experience in previous community development projects,

nutrition research, agriculture research, and integrated
nutrition-agriculture research projects. This led to the framework in Box 1, which consists of a series of questions for
assessing the Community, Project, and External factors.


Berti P.R. et al.

Box 1 Factors taken into consideration when deciding on the
nature of nutrition-sensitive agriculture interventions

In choosing the NSA interventions, guided by the framework in Box 1, the user can then select from a list of potential
interventions that have been shown in other settings to be
successful at improving nutrition outcomes, or that could be
reasoned to potentially improve nutrition. A list of interventions the authors considered is shown in Table 1.

Application of the framework to guide selecting NSA
Given budget and timing limitations, the site PIs (HPV, ALV,
PW) estimated they could test, at most, four interventions per
site. In other words, they had sufficient time and resources to
introduce four distinct NSAs to participating communities,
provide adequate training and requisite materials for the families to start the intervention, and allow it to continue for at
least one growing season. The application of the framework,

therefore, helped to reduce the candidate interventions from
the 19 listed in Table 1 to only four.
To collect the information required to answer the questions
in Box 1, we used qualitative and quantitative methods to
conduct a situational assessment characterizing the nutritional
situation, local farming practices and the relationships between food production, availability and consumption, as well
as the conditioning factors (e.g. gender dynamics, market infrastructure, natural resources, including local wild foods) that

promote (or impede) healthy diets. Questions regarding
Community Factors were answered through formative research involving household surveys, in-depth interviews and
focus group discussions targeting different ethnic groups at
the study sites. External Factors were assessed through key
informant interviews with district and sub-district/commune
officials. The project teams answered the questions about
Project Factors.


The process of developing a nutrition-sensitive agriculture
Table 1 Agriculture interventions we considered for testing in
integrated NSA interventions
Animal-source food production Horticulture
and consumption
Large animal husbandry –
for meat and milk
Small animal husbandry
– for meat
Poultry rearing – for
meat and eggs
Fish farming
Insect farming
Field crops
Rice (System of Rice
Intensification,
known as SRI)
Legumes and intercropping
Infrastructure
Improved irrigation system
Fish pond improvement

Fruit trees (various types)

teleconferencing, to all the authors. The authors questioned
the site leaders and asked them to defend their decisions,
and when all the questions were satisfactorily addressed, the
decisions on the NSA selections were endorsed.
Research ethics review

Home gardening of various vegetables
Home gardening of pumpkins
School gardens of various vegetables
Perennial vegetables
Market development
Local markets for fruit, vegetables,
meat and eggs, and preserved foods
Export markets
Behaviour
Time management changes to
reduce women’s workload
Nutrition promotion and education aimed at breastfeeding practices,
complementary feeding, young child
feeding, diet of pregnant and
breastfeeding women, or diet
of whole family

Sloping land protection

Data were collected during October 2013 and January
2014. Specifically:

(a) In household surveys: parents or caregivers of children under 5 years old were randomly selected to
be interviewed using a standardized questionnaire;
(b) Individuals from surveyed households were purposively sampled to take part in in-depth interviews
to more fully characterize agricultural production
and opportunities at the household level;
(c) Focus group discussions were held with men and women
who had children under 5 years;
(d) Key informant interviews were conducted with
commune/sub-district leaders, agriculture staff,
healthcare staff and school nurses and with district agriculture and health officials.
The focus of this paper is not the data gathering process and
quantitative results, but rather how the results were interpreted
and used, with respect to Box 1. Therefore, details of the
methods and the survey results are in the Supplementary
Materials section.
The leaders of each of the three research sites used the
results of the data collection to answer the questions in Box
1 and to guide their preliminary selection of the NSAs to be
trialled. Their results and preliminary NSA selections were
presented to local partners, and then, through

The research was approved by the Ethical Review Board for
Biomedical Research, Hanoi School of Public Health, and
Human Experimentation Committee Research Institute for
Health Sciences, Chiang Mai University. Permission was also
granted by the community via the local community leaders.
Before conducting the interviews, the purpose of the data collection was explained to the respondents, the confidentiality of
their responses was assured, and their informed and voluntary
consent was sought. Agreement to participate was granted
orally (in Vietnam) or in writing (in Thailand).

Results
Development of framework to guide selecting NSA
Through a series of team meetings and discussions, the authors
developed and refined the questions, leading to the framework
shown in Box 1, including, as planned, questions related to
community factors, project factors and external factors. In the
development of the framework we recognized that the process
would work best if we split the process into two steps. (1)
Characterized the site in terms of agricultural production, and
food and nutrition to identify basic needs and potentials, and
considered project factors that would present limits or provide
opportunities at each site. Then the observations and interpretations would suggest a number of candidate interventions
drawing from the options in Table 1. (2) Assess the suitability
of each of the candidate interventions according to the various
Community, Project, and External factors laid out in Box 1.
Application of the framework to guide selecting NSA
The first step of characterizing agricultural production, food
and nutrition, and considering project factors suggested a
number of possible interventions for each site (Box 2a, b
and c). Five or six NSA interventions were considered in
detail for each site, with 11 different interventions considered
across the three sites. For A Luoi, the site team chose the
candidate interventions Poultry, Fishponds, Home gardens,
Beans and inter-cropping, and Infant and child feeding. For
Yen Chau, they selected the candidate interventions Poultry,
Pumpkins, Home gardens, Rice, School gardens, and Sloping
land crop diversification. For Mae Chaem, the site team chose
the candidate interventions Poultry, Food preservation, Home
gardening, Infant and child feeding, and Improving water

Berti P.R. et al.

systems. In all cases, the candidate interventions were selected
because the local teams saw that they could meet the

Box 2a Characteristics of A Luoi, in relation to Box 1

production and nutrition needs within the time and resource
constraints of the project.


The process of developing a nutrition-sensitive agriculture

Box 2b Characteristics of Yen Chau, in relation to Box 1


Berti P.R. et al.

Box 2c Characteristics of Mae Chaem, in relation to Box 1


The process of developing a nutrition-sensitive agriculture

Box 3a Evaluation of candidate interventions in A Luoi district


Berti P.R. et al.

Box 3b Evaluation of candidate interventions in Yen Chau


The process of developing a nutrition-sensitive agriculture

Box 3c Evaluation of candidate interventions in Mae Chaem


Berti P.R. et al.

The site teams then assessed the suitability of each of the
candidate interventions according to various social, project, and
external factors from Box 1. There was marked heterogeneity
within and between sites regarding current cropping and animal
husbandry practices, decision-making about farming and feeding, food security, and dietary practices. All three sites had the
same time constraints to carrying out the interventions (about
18 months, due to funding deadlines), which limited potential
interventions to annual crops or small animals, and precluded
working on major infrastructure or legislative changes or market
development. In some cases the implications of the answers were
clear cut and easily applicable. For example, in A Luoi, poultry
were regarded favourably because women would be involved in
the raising of the poultry and be able to feed eggs to their children. In other cases, the implications were less clear. For example, increased income through agricultural production and market
access for farmers are two of the pathways through which farming households can improve diet and health (Herforth and Harris
2014). However, the site teams (the scientists and local officials)
felt that increasing emphasis on the markets would move farmers
from food crops to non-food crops, which provide variable and
uncertain returns, and thus threaten local food security. Local
officials report that growing non-food crops for market (rubber,
acacia, cassava, and maize for cattle) have increased cash incomes but not improved household diets. Market development

was therefore eliminated from consideration at the current time,
but if production were to increase and households became more
food secure, improving market access may be an appropriate
next step to helping farmers move out of poverty.
The selection of three or four of the candidate interventions, and rejection of other interventions, varied by site, based
on contextual factors. While all the interventions were thought
by the site teams and the international teams to have the potential to be successful, some of them were perceived to be
more promising candidates, and given the self-imposed maximum of four interventions to test per site, less promising
interventions were discarded:
For A Luoi, fish ponds would likely result in increased
work load for women, and were rejected largely for this
reason. Similarly, the bean intervention would require a
great deal of extra work and, if women were involved, it
would take them away from childcare. Thus in A Luoi,
the team chose small scale poultry rearing and home
gardening, as well as training on infant and child feeding, which would all be expected to lead to improved
diets without compromising child care.
For Yen Chau, there were no clear reasons to reject any of
the candidates. Having to narrow the candidates down, it
was reasoned that some pumpkin cultivation could be incorporated within home gardening, and the school was
thought to be a less promising site for gardening than the

home. Rice intensification was selected, anticipating that it
would be popular with farmers and local officials, and
knowing that while men would do most of the work, women usually control the sale of the surplus and could be
anticipated to use the proceeds for the family. Increasing
production on sloping land was also selected, as it was
expected to be carried out by the men while providing
increased nutritious foods from underused lands. Poultry
rearing would require some work by women, but would

have minimal impact on childcare since it would be close
to home, while producing nutrient-rich food.
For Mae Chaem, there was some consideration of working on improving water systems. While desirable, it
would probably stretch the timelines and budget of the
project. Food preservation may be a good intervention,
but it was rejected as it was deemed to have less potential for positive impact on health and food security and it
was decided that it would be better to focus on food
production (poultry and home gardening) and infant
and child feeding practices.
These assessments were made by the local teams and, after
presentation to the entire international team through teleconferences, the decisions were endorsed by all the team members.
In summary, four of the selected interventions involved
annual vegetables or legumes (Home gardens in all three sites,
and Improved bean production). All three sites included poultry rearing. Two of the three sites also decided to include
training on infant and child feeding, although this is perhaps
not strictly a separate NSA intervention. The two other selected interventions included one conventional agriculture intervention involving staple crops (Rice intensification, focusing
on utilization of young seedlings, lowering planting density
and fewer plants), and one unconventional intervention, focusing on diversification into perennial vegetable and fruit
crops on underused sloping lands. Most of the interventions
fit within existing gender roles, and no need was felt to try and
change these roles. The exception was childcare and feeding,
where the sites would work to increase male involvement in
order to reduce maternal workloads. In all cases, in all the
research sites, following selection of the intervention from
the team, the suggested interventions were presented to local
officials and participating households, and endorsement of the
selected NSA interventions was obtained with no further
modifications.

Discussion

In this paper, we describe the process of using a framework to
select NSA interventions to implement. The framework presented here may be most useful to others as a supplement to


The process of developing a nutrition-sensitive agriculture

existing FAO guidelines (FAO 2015), which provide overall
guidance on the entire NSA program cycle, whereas our
framework focuses almost exclusively on choosing which agricultural intervention should be trialed. Other frameworks
use a Bpathways^ approach, identifying the three pathways
from which agriculture can contribute to improved nutrition
outcomes (Food Production, Agricultural Income, and
Women’s Empowerment pathways (Herforth and Harris
2014)) and then seek to move participants along one or more
of the pathways. Following our framework, we examined
community needs and potential and then considered various
possible barriers and enablers for realizing the potential. One
novel barrier/enabler that we considered, albeit briefly, was
Bpower^ - considering the nature of local power structures
and whether they would be supportive of the intervention –
a factor that is often overlooked in development planning
(Green 2012).
There is no gold standard for assessing whether the
framework Bworked^. Perhaps at the end of the project,
if the selected NSAs have been shown to be successful
(feasible, favoured by the farmers, and with positive
dietary impacts), then we may conclude that the framework did not steer us in a wrong direction. However,
there were two key weaknesses to be considered.
1. Some of the intervention selections were made following in-depth, multi-layered discussions among
the research team, local officials and community

members. While the discussions were in part
prompted by working through the framework, the
discussions are not easily summarized, they would
be difficult to replicate, and we are not able to provide guidelines on how to carry out these discussion. For example, there is nothing in Box 3b that
would lead an outside observer to choose home gardens over pumpkin cultivation, yet this was the result of detailed discussions. Pumpkins are commonly
grown, often intercropped with maize. They are
mainly used as animal feed and are not included
in the families’ diets. In Yen Chau fewer than
20 % of children ate Vitamin A-rich foods in the
previous 24 h and so increased consumption of
pumpkins would provide a beneficial increase in vitamin A intake. The Yen Chau research team included an agronomist, sociologist, and nutritionist and
they discussed training local households in pumpkin
processing (e.g., powder, jam, syrup, etc.) with local
agricultural and nutrition officials as a means to increase pumpkin consumption. However, this would
be new to the farmers and there would be numerous
constraints such as the time required for pumpkin
processing, additional ingredients for pumpkin processing that were not available locally, lack of

infrastructure for processing and storing pumpkins.
Considering all of this, the team decided not to
work on pumpkin production and processing. Other
possible interventions were never even considered,
as they were eliminated for what are obvious reasons for locals. An example of this is insect farming, which would not be accepted by the farmers in
Vietnam. Similar nuanced discussions were held
around every candidate intervention. Such supplementary knowledge, not captured in the framework,
is an important part of the overall process.
Another aspect of the nuanced discussions involved
in the decision-making and not adequately captured is
explicit weighting to account for the varying potential
for nutritional impact of different interventions. For example, there is increasing evidence of the positive nutritional impact of animal-source food interventions

(e.g., (Hoddinott et al. 2015; L. Iannotti et al. 2013;
L. L. Iannotti et al. 2014), and the biology behind the
impact is self-evident, as animal-source foods are
nutrient-dense with small amounts providing large proportions of requirements (Murphy and Allen 2003). On
the other hand, cereal-based interventions have much
less evidence supporting their effectiveness (Arimond
et al. 2011). This weighting is included in the decision-making, but not captured in the framework.
2. It took many months and hundreds of person-hours
of work to use the framework. Given the relatively
short time available to carry out the project, the
NSAs had to be selected before the evaluation presented here was fully completed. Frameworks with a
lighter burden would be more easily implemented.
The NSAs selected through this evaluation may
not be different from those that would be selected
with a much lighter and less formal approach. We
thus see a continuum of methods for choosing
NSAs.
a. On the intense end of the continuum, programmes
would collect and analyse at least as much data as
we collected and summarized (see Supplementary
Materials), if not more. The process could begin even
further upstream with a full community health assessment and livelihoods analysis, if the programmers
were not pre-committed to working on NSAs and
had a year or more for planning and assessment.
b. At the lighter end of the continuum, a one-hour meeting of a small team of agronomists, nutritionists and
health workers familiar with the setting could select a
few candidate interventions. A community participatory workshop would narrow the list down further,


Berti P.R. et al.

filtering out culturally or otherwise inappropriate interventions. For example, poultry rearing and home
gardening are existing, appreciated practices with increasing evidence supporting their effectiveness in
addressing malnutrition. In our current research sites,
we could have quickly settled on poultry and home
gardening with meetings lasting only a few hours.
The very best interventions may not necessarily have
been identified but we could be reasonably confident
that they would, at worst, be useful.
Whether practitioners tend towards the heavy or light end
of the continuum will depend on their familiarity with local
needs, possibilities and culture, and resources available for
their assessment.
A third weakness, not specific to the framework, but
to NSAs in general, is that from the outset, our intention was to identify and test NSA interventions, without
first evaluating the need for improved food production
and diet in the study communities relative to other possible areas of intervention. While we had no doubt that
production and diet could be improved, there were and
are certainly other needs in the communities, and they
may be of higher priority to the community members.
We acknowledge this overall weakness, while also recognizing this is generally how development research
takes place – agriculture scientists conduct agriculture
research, nutrition scientists conduct nutrition research,
health services scientists conduct health services research, and so on, without first fully appreciating community needs and priorities.
Despite these weaknesses, there is value in laying out
the specific strengths and weaknesses of candidate interventions, and being intentional in NSA selection. Some of
the factors, such as potential impact on breastfeeding, are
not often considered by agronomists. Seasonality and gender roles in food production are not often considered by
health workers. Power (within households, within communities, and at a higher level) is not regularly considered in
development programs of any kind. Explicitly enumerating

the various strengths and weaknesses of candidate interventions, as is done through the application of Box 1,
helps the agriculture and health fields communicate, understand one another, and appreciate the importance of
integration for effective nutrition-sensitive agriculture.
NSA interventions are of increasing interest to those working in global health and nutrition and the guidelines developed and tested here can help implementers to decide
what agriculture component may be feasibly implemented
with a reasonable chance of having positive agricultural
and nutritional impacts.

Acknowledgments The research was carried out with the aid of
a grant from the International Development Research Centre,
Ottawa, Canada. We acknowledge the contributions of: Lisa
Macdonald (HealthBridge) in managing the project: the Thai
team, including Palika Champrasert, Surachet Jina Keaw,
Tanawit Wongsur, Natjan Chairat, and Anantika Ratnamhin of
ICRAF Thailand Office, and Sakda Pruenglampoo, and Posri
Leelapat of Research Institute for Health Sciences, Chiang Mai
University; the collaboration of researchers of the Center for
Agriculture, Forestry Research and Development of Hue
University of Agriculture and Forestry.
The paper benefitted greatly from reviews from the Senior Editor and
two anonymous reviews.
Compliance with ethical standards
Conflict of interest statement All authors declare that there are no
conflicts of interest.

References
Arimond, M., Hawkes, C., Ruel, M., Sifri, Z., Berti, P., LeRoy, J., et al.
(2011). Agricultural interventions and nutrition: lessons from the
past and new evidence. In B. Thompson & L. Amoroso (Eds.),
Combating micronutrient deficiencies: food-based approaches (pp.

41–75). Rome: Food and Agriculture Organization/CABI
International.
Berti, P. R., Krasevec, J., & FitzGerald, S. (2004). A review of
the effectiveness of agriculture interventions in improving
nutrition outcomes. Public Health Nutrition, 7(5), 599–609.
doi:10.1079/PHN2003595 S1368980004000722.
Dung, N. V., Vien, T. D., Lam, N. T., Tuong, T. M., & Cadisch,
G. (2008). Analysis of the sustainability within the composite
swidden agroecosystem in northern Vietnam. Agriculture,
Ecosystems and Environment, 128, 37–51.
Dury, S., Alpha, A., & Bichard, A. (2015). The negative side of the
agricultural–nutrition impact pathways: a literature review. World
Food Policy, 2(1), 78–100.
FAO (2004). Food insecurity and vulnerability in Viet Nam: profiles of
four vulnerable groups. Agriculture Development Economics
Division (ESA).
FAO (2015). Designing nutrition-sensitive agriculture investments.
Checklist and guidance for programme formulation. Rome:
Food and Agriculture Organization of the United Nations.
General Statistical Office of Vietnam (2011). Viet Nam multiple indicator
cluster survey 2011, final report. Viet Nam: Ha Noi.
Government of Vietnam Socialist Republic (2005). Vietnam
achieving the Millenium development goals. Socialist
Republic of Vietnam.
Green, D. (2012). From poverty to power: how active citizens and effective states can change the world. U.K.: Practical Action Publishing
and Oxfam International.
Headey, D., Chiu, A., & Kadiyala, S. (2011). Agriculture’s role in the
Indian Enigma: Help or hindrance to the undernutrition crisis?
IFPRI discussion paper 01085. Washington, DC: International
Food Policy Research Institute (IFPRI).

Herforth, A., & Harris, J. (2014). Understanding and applying
primary pathways and principles. Brief #1. Improving nutrition through agriculture technical brief Series. . Arlington,


The process of developing a nutrition-sensitive agriculture
VA: USAID/Strengthening partnerships, results, and innovations in nutrition globally (SPRING) Project.
Hoddinott, J., Headey, D., & Dereje, M. (2015). Cows, missing
milk markets, and nutrition in rural Ethiopia. The Journal of
Deve lopme nt St udies , 5 1( 8), 95 8– 9 75 . d oi :1 0. 1 0 80
/00220388.2015.1018903.
Humphries, D. L., Dearden, K. A., Crookston, B. T., Fernald, L. C., Stein,
A. D., Woldehanna, T., et al. (2015). Cross-sectional and longitudinal
associations between household food security and child anthropometry at ages 5 and 8 Years in Ethiopia, India, Peru, and Vietnam.
Journal of Nutrition, 145(8), 1924–1933. doi:10.3945/jn.115.210229.
Iannotti, L., Muehlhoff, E., & Mcmahon, D. (2013). Review of
milk and dairy programmes affecting nutrition. Journal of
Development Effectiveness, 5(1), 82–115. doi:10.1080
/19439342.2012.758165.
Iannotti, L. L., Lutter, C. K., Bunn, D. A., & Stewart, C. P. (2014). Eggs:
the uncracked potential for improving maternal and young child
nutrition among the world’s poor. Nutrition Reviews, 72(6), 355–
368. doi:10.1111/nure.12107.
Isvilanonda, S., & Bunyasiri, I. (2009). Food security in Thailand.
Status, rural poor vulnerability, and some policy options.
Department of agricultural and resource economics,
Kasetsart University, Bangkok.
Kadiyala, S., Harris, J., Headey, D., Yosef, S., & Gillespie, S. (2014).
Agriculture and nutrition in India: mapping evidence to pathways.
Annals of the New York Academy of Sciences, 1331, 43–56.
doi:10.1111/nyas.12477.

Limnirankul, B., Promburom, P., & Thongngam, K. (2015). Community
participation in developing and assessing household food security in
the highlands of northern Thailand. Agriculture and Agricultural
Science Procedia, 5, 52–59.
Mackintosh, U. A., Marsh, D. R., & Schroeder, D. G. (2002). Sustained
positive deviant child care practices and their effects on child growth
in Viet Nam. Food and Nutrition Bulletin, 23(4 Suppl), 18–27.
Murphy, S. P., & Allen, L. H. (2003). Nutritional importance of animal
source foods. Journal of Nutrition, 133(11 Suppl 2), 3932S–3935S.
National Statistical Office of Thailand (2013). Multiple indicator cluster
survey: MICS4, Final Report. Bangkok.
Rambo, A. T., & Cuc, L. T. (1996). Development trends in
Vietnam’s Northern Mountain region. Hanoi: Centre for
Natural Resources and Environmental Studies, Vietam
National University.
Sikor, T., & Nguyen, Q. T. (Eds.). (2011). Realizing Forest rights in
Vietnam: addressing issues in community Forest management.
Bangkok: Center for People and Forest (RECOFTC).
Thomas, D. E., Preechapanya, P., & Saipothong, P. (2004). Landscape
agroforestry in northern Thailand: impacts of changing land use in
an upper tributary watershed of montane mainland southeast asia.
Chiang Mai, Thailand. ASB.
Tuan, V. D., Hilger, T., MacDonald, L., Clemens, G., Shiraishi, E., Vien,
T. D., et al. (2014). Mitigation potential of soil conservation in maize
cropping on steep slopes. Field Crops Research, 156, 91–102.
Turner, R., Hawkes, C., Jeff, W., Ferguson, E., Haseen, F., Homans, H.,
et al. (2013). Agriculture for improved nutrition: the current research
landscape. Food and Nutrition Bulletin, 34(4), 369–377.
Wangpakapattanawonga, P., Kavinchan, N., Vaidhayakarn, C.,
Schmidt-Vogt, D., & Elliott, S. (2010). Fallow to forest: applying indigenous and scientific knowledge of swidden cultivation to tropical forest restoration. Forest Ecology and

Management, 260, 1399–1406.
Webb, P., & Kennedy, E. (2014). Impacts of agriculture on nutrition:
nature of the evidence and research gaps. Food and Nutrition
Bulletin, 35(1), 126–132.

Peter R. Berti Peter joined
HealthBridge in 1998. As
Nutrition Advisor he is responsible
for ensuring the technical quality
of all of HealthBridge’s work in
food and nutrition research and
programming. Much of his work
has been in the Americas
(Ecuador, Bolivia and Brazil)
where he has conducted research
and managed programmes in food
system research, nutritional assessment and food fortification. His
current work emphasizes
nutrition-sensitive agriculture: the
integration of food and nutrition interventions with agriculture interventions. He brings to his work a command of data analysis, which enables
novel insight into program evaluation. This includes analyses showing the
impact of an integrated nutrition and agriculture intervention in Malawi, the
adequacy evaluation of a large-scale World Vision intervention, and the
widely cited review of nutrition impacts of agriculture programs.

Rachelle E. Desrochers joined
HealthBridge in 2011 as a specialist in the emerging field of
GeoHealth. During her graduate
work in macroecology, Rachelle
became interested in the connection between human modifications

of the environment and human
health. Her role principally consists
of scientific and technical support
for the development, implementation and analysis of research designed to inform policy and program planning. She leads projects
focused on mapping malaria and
arbovirus (e.g. Dengue Fever) risk; has experience in many field surveys
employing direct data capture, having provided technical support for several
large-scale household surveys; and has collaborated on operational research
for the malaria program in Togo. She also provides support for research
within the context of HealthBridge’s Livable Cities program in Asia.

Pham Van Hoi is a lecturer on
agroecology and environmental
sciences at Vietnam National
University of Agriculture. He
completed his PhD on
Environmental Sociology at
Wa g e n i n g e n U R , t h e
Netherlands in 2010. He has been
co-PI and researcher in a number
of international projects on sustainable agricultural and rural sociology. His publications focus on
environmental governance practices in the agricultural sector in
Vietnam.


Berti P.R. et al.
Lê Văn An Prof. An finished a
BSc. degree in animal production
from Agricultural University
No.2, Ha Bac, Vietnam, in 1983

and a MSc. degree in livestock
production system from
the Swedish Un iversity of
Agriculture Sciences in 1999. He
finished his PhD. degree in animal science at the same
University in 2004 and is now
the Rector of Hue University of
Agriculture and Forestry.

Agric Ngo Tung Duc Mr. Ngo
Tung Duc finished his Ph.D degree in natural resource management from Kyoto University,
Japan in 2011, MSc. degree in agricultural systems from Chiang
Mai University, Thailand in
2005, and BSc. degree in forestry
from Hue University of
A g r i c u l t u r e a n d F o r e s t r y,
Vietnam in 1998. He has vast experience in natural resource management and livelihood of local
people in upland areas, as well as
social forestry. He is now the
Head of the Post-graduate Training Department, Hue University of
Agriculture and Forestry.

Hoang The Ky is working at
HealthBridge Vietnam (HBV) on
the Nutrition and Food Security
project (or From Field to Fork
project) in the highlands of
Vietnam and Thailand. He is responsible for assisting the
Research Coordinator with implementation of the project. In particular, he is responsible for coordinating project training and workshops, leading nutrition training,
and providing nutrition technical

support to project partners. He coordinates and follows up on the
progress of project implementation. Besides, he is working on a nutrition
initiative through cooperative research between HealthBridge and
Bioversity International under the CGIAR Research Program on Humid
Tropics to assess the potential of diverse local foods and improve diet
quality and diversity, especially for women of reproductive age and for
children between 23 and 12 months in Vietnam. Ky graduated from
Hanoi School of Public Health (HSPH) with a Masters degree in
June 2012. From July 2008 - May 2013, he worked for World Vision
International in Vietnam in technical support for health and nutrition. In
this period, he successfully conducted models of nutrition (such as: PD/
HEARTH; nutrition clubs, green vegetable garden…) for minority groups

(H’mong, Thai, Muong …) in Northern provinces (Hoa Binh, Dien
Bien…). From August 2010 to August 2011, he worked for CHILILAB
of HSPH, a research centre for public health and population in the Hanoi
office as a researcher and field coordinator.

Nga Le Thi Dr. Le Thi Nga, a
medical doctor with more than
20 years’ experience in public
health and research, is Deputy
Country Director at
HealthBridge Vietnam. Before
joining HBV, Dr. Le worked at
various organizations, including
PAT H / U S , P a t h f i n d e r
International, Vietnam-Australia
Malaria Control Project, Save the
Children/US, and the National

Institute of Nutrition. Her experience is in operational research,
TB prevention, reproductive
health, maternal and child health, immunizations, nutrition, malaria,
HIV/AIDS, public-private partnerships and health systems strengthening.
Dr. Le’s expertise is in operational research and epidemiology; program
design, public-private partnership, management; training and capacity
building; information, education and communication (IEC), and monitoring and evaluation (M&E). Dr. Le was granted the Master Degree of
International Research Bioethics by Monash University, Australia;
Master Degree of Public Health by the University of Queensland,
Australia; and Medical Doctor by the Hanoi Medical University.

Dr. Wangpakapattanawong’s
research interest is forest ecology.
He graduated with a Ph.D. in
Forest Sciences from
the University of British
Columbia (UBC) in 2001. While
his B.Sc. in Biology is from
C h i a n g M a i U n i v e r s i t y,
Thailand, where he has been
working since finishing his UBC
degree, and his M.Sc. in Botany is
from Iowa State University, USA.
At CMU, he teaches ecology and
biodiversity for undergraduate
and graduate students in Biology,
Environmental Science, and Biodiversity and Ethnobiology. He began his
research career with research on tropical forest restoration (http://www.
forru.org). Having been working for more than 12 years at CMU as an
assistant professor, his research interests have broadened to include

several topics that relate to forest ecology, for example, carbon
sequestration in forest restoration, ecology of figs, and ethnoecology. In
addition, he also works as a part-time visiting scientist/country representative of the Thailand Office of World Agroforestry Center (ICRAF)
overseeing research projects on land tenure, land-use changes, and
agricultural-system management and human nutrition.