e ˝˘ˆ - enn · acf, and plenary discussion during the acf research conference, paris, 2016. t he...

菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The second Lancet series on undernutri-tion was a key milestone in the globalcall for more robust evidence on nutritionsensitive and specific interventions

(Lancet, 2013). Academics, non-governmentalorganisations, national stakeholders and donorshave been increasingly involved in operationalresearch to improve the effectiveness of a setof multi-sectoral interventions for better pre-vention and/or treatment of undernutrition.Donors and policy-makers regularly reaffirmedtheir needs for more robust results to orienttheir policies. This has led to a significant increasein research concerned with humanitarian nu-trition programming and policies in crisessettings (Blanchet, 2015 ). While channels topublish or communicate results exist, they offerlimited space for sharing learning on processimplementation and research uptake by relevantstakeholders. There is no specific space dedicatedto the presentation of scientific results on nu-trition in humanitarian settings.

To help fill this gap, Action Contre la Faim(ACF) organised in Paris on November 9th 2016the first annual international conference entitled“Research for nutrition - operational challengesand research uptake in prevention and treatmentof undernutrition”. ACFs longstanding experiencein programming and operational research inhumanitarian settings, active research portfolioon malnutrition prevention and treatment, andopenness to share challenges as well as successes,meant we were well placed to convene thisgathering. The conference had two main ob-jectives: to present a selection of the latest re-search related to the identification of effectivenutrition specific and sensitive interventions incrises contexts; and to provide a space for dis-cussion and debate about nutrition researchmethodological design and uptake challenges.Almost 200 individuals participated, with 12international universities, 16 international non-governmental organisations (INGOS) and sevendonors represented .

Two panel debates explored operational re-search challenges and research uptake consid-erations; proceedings are summarised in thisissue of Field Exchange. In both sessions, pan-ellists and plenary participants explored commonexperiences and themes. Research shared inpresentations spoke to key evidence gaps theinternational community is working on:• Impact of cash transfers on undernutrition:

In recent years, a set of robust studies were launched to assess the effects of cash transfers on nutrition. Preliminary results of two randomised controlled trials assessingthe effects of cash transfers on wasting in Burkina Faso and Pakistan were presented.

• Anthropometric assessment of undernutrition:A significant area of research interest is to identify nutritional vulnerability in children,including but not limited to anthropometricindicators. One study explored identificationof wasting, and outcomes amongst infants under 6 months of age in Bangladesh; a second investigated the association of mortality risk with different anthropometricmeasures.

• Impact of water, sanitation and hygiene (WASH) activities on undernutrition: A recent movement, “Wash In Nut’,” aims to orientate WASH programmes to improve nutrition treatment and prevention outcomes. Evidence of the effects of such strategies on nutrition are still missing. The results of two fascinating randomised con-trolled trials in Chad were presented.

• Health service delivery models for acute malnutrition: There has been recent drive and a number of initiatives to strengthen malnutrition treatment services within health systems, at national and local levels and within communities. Two approaches being researched were presented from Burkina Faso and Mali. Of particular note, research on characterisation of MAM in Burkina Faso should allow for better future treatment strategies.

Presented research that is available is summarisedin this special section of Field Exchange 54,with headlines and key contacts provided forstudy results not yet available.

Feedback on the conference was extremelypositive, reaffirming the importance and rele-vance of having dedicated time and space forresearchers, practitioners and donors to exchange,discuss and reinforce collaboration on a spectrumof nutrition research with a strong operationalfocus. ACF is already preparing for the 2017conference with a view to this becoming an an-nual event to share, think and innovate aroundnutrition research in humanitarian settings.

For more information, contact: Myriam Aissa,email: [email protected]

To view video footage of the day, visit:http://bit.ly/2kDgLnQ

Editorial

Pierre Micheletti, Vice-President ofACF France, opens the conference

G. G

affiot

/ AC

F Bu

rkin

a Fa

so

Mau

rine

Tric

By Myriam Ait Aissa

Myriam Ait Aissa leads the Action Against Hunger Research and AnalysesDepartment, which overall aim is to produce evidence for better action. Beforejoining Action Against Hunger, she worked for the Scientific Department ofACTA International, a French applied agricultural research institute, as aresearch fellow. Her work focused on nutrition and food security research areas.

ReferencesBlanchet, K et al (2015). An evidence review of research onhealth interventions in humanitarian crises. LSHTM,Harvard School of Public Health, ODI. Commissioned byELHRA. Technical report. October 2013. HYPERLINKwww.elrha.org/wp-content/uploads/2015/01/Evidence-Review-22.10.15.pdf

Ruel, M et al (2013). Nutrition-sensitive interventions andprogrammes: how can they help to accelerate progress inimproving maternal and child nutrition? Lancet 2013; 382:536-51

Produced by ENN. Extracted from Field Exchange 54, February, 2017, www.ennonline.net/fex

菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Panel discussions

Methodological challenges for operationalresearch in the humanitarian contextBy Myriam Ait Aissa and Melchior de Roquemaurel

Myriam Ait Aissa leads theAction Against HungerResearch and AnalysesDepartment, which overallaim is to produce evidencefor better action. Before

joining Action Against Hunger, she worked forthe Scientific Department of ACTA International,a French applied agricultural research institute,as a research fellow. Her work focused onnutrition and food security research areas.

Melchior de Roquemaurelcoordinated the ActionAgainst Hunger Researchfor Nutrition Conference#R4NUT.

This article summarises contributions fromYves Martin-Prével, Institute for Research onDevelopment (IRD); Kate Golden, ConcernWorldwide; Timothy Williams from the SPRINGproject; Victoria Sibson, UCL; Myriam Aït-Aissa,ACF, and plenary discussion during the ACFresearch conference, Paris, 2016.

The ACF research conference, Paris, 2016included a panel discussion on the oper-ational challenges of research in human-itarian contexts1. This article summarises

the session. The discussion was moderated byYves Martin-Prével of the Institute for Researchon Development (IRD). Panellists were KateGolden from Concern, Timothy Williams from theSPRING project, Victoria Sibson from UCL andMyriam Aït-Aissa from Action Against Hunger.

The panellists shared some of the operationalchallenges they have experienced in a number ofrecent research studies (see Box 1) and suggestedrecommendations to improve the overall qualityand efficiency of research on nutrition in human-itarian contexts. The studies in question focusedon stunting and/or severe acute malnutritionand/or chronic undernutrition, involving experi-mental and observational designs. Discussionslargely centred on experiences with randomisedcontrolled trials (RCTs) in complex settings.

Challenges related to RCTsRCT methodology is often used for nutrition re-search in humanitarian contexts and is acknowl-edged as the ‘gold standard’ in terms of qualityof evidence. Nevertheless, implementing an RCTin a humanitarian crisis context poses manychallenges.

Humanitarian programmes operate withinstrict and often circumscribed and short-term,donor-driven timeframes, whereas scientific re-search often requires a lengthier period for im-plementation.

In some contexts, RCTs are not feasible: a well-designed observational study may be better thana poorly implemented RCT. Nevertheless, RCTsare feasible in many humanitarian contexts, evenin volatile situations. Critically, researchers mustremain mindful of the context and ‘expect the un-expected’, which typically impacts the length andcost of the research. For example, there may besecurity issues during randomisation, a pipelinebreakdown in nutritional products, or sudden dif-ficulty in population access. An acute conflict ornatural disaster may occur on top of a chronicemergency situation. From a research perspec-tive, with the right attitude and preparedness,these are manageable situations and sometimesresearch consequences may be limited, e.g. anevent may impact on the control and interven-tion groups equally. However, researchers mustoften engage in lengthy discussions with opera-tions teams concerned with the challenges andrisks of implementing research in volatile con-texts. In a multi-sector research project, these

conversations are multiplied by the number ofsectors involved. Good communication with allparties regarding the project’s purpose and ob-jectives can help minimise adverse impacts.

RCTs are important but can be a ‘risky’ invest-ment for non-governmental organisations(NGOs). From the outset, funding can limit thescope of investigation, such as limiting studyarms to a control and one intervention. Where noimpact is found in a trial, programme teams maybe disheartened by the lack of a clear ‘positive’ orimmediately actionable result. This is particularlytrue when extraordinary efforts by the pro-gramme team have been required to carry out anRCT in challenging contexts. To help address this,a mixed-methods approach, including other ele-ments such as qualitative research, monthly sur-veillance and process evaluations alongside RCTs,can help ensure practical learning takes placethat can be applied even if the headline resultsare less ‘exciting’.

Conducting experimental studies on nutritionrequires investment in robust context analysis todocument impact pathways or process evalua-tions, given that nutrition is context-specific,multi-sectorial and related to seasonality. It is im-portant to implement qualitative methodologies,capture seasonal features and take the necessarytime for full analysis of all monitored indicators,involving experts from the field, operations andacademia.

Funding challengesFunding timing and flexibility is an ongoing chal-lenge in humanitarian contexts. Programme andresearch funding sources are typically distinct,with different timeframes and donor require-ments. For example, the REFANI project researchelement was funded by DFID, while programmeactivities were funded by ECHO. Programmefunding was not fully committed until severalmonths after the research funding was securedand study preparations needed to start, includinghiring staff. This required both the research andprogramme partners to plan flexibly and tospend research funds before the trial was 100%guaranteed to happen. Single-source fundingthat covers both research activities and the inter-vention(s) would allow better coordination of ac-tivities. It would also facilitate study designs thatcould answer the most relevant questions, ratherthan just those permitted by the current pro-gramme design.

1 To view a video of this panel debate, visit: http://bit.ly/2jwmmKx

Box 1 Research experience ofpanellists

In 2015 UCL and Concern collaborated totest the effectiveness of an early seasonalrandomised controlled trial (RCT) toprevent acute malnutrition in rural Niger.This was one of three studies by the ACF-led consortium REFANI(www.actionagainsthunger.org/refani).Concern also conducted an RCT on a‘community resilience to malnutrition’integrated programme in Chad. Action Against Hunger researchdepartment has a multidisciplinary teamand currently has a portfolio of six RCTsand seven observational studies recentlycompleted/in progress.

The Strengthening Partnerships, Results,and Innovations in Nutrition Globally(SPRING) project is USAID’s flagship, multi-sector nutrition project focused onreducing stunting and anaemia amongchildren in the first 1,000 days. SPRING hasconducted operations and implementedresearch and evaluations in over 10countries to guide USAID nutritionprogrammes and contribute to the globalevidence base.


菀

Panel discussions

Bridget Fenn, ENN, contributingto plenary discussion

Mau

rine

Tric

Managing the unexpected in research re-quires donor flexibility. Delays mean extendeddeadlines are often required. Having a financialenvelope for unplanned events is key in the im-plementation phase, empowering researcherswith the flexibility and reactivity needed to re-spond quickly to change.

Challenges with humanresourcesEstablishing international and national staffteams is not straightforward. Cultural sensitivityis important but is seldom taken into account orgiven the priority it needs. Capacity developmentof national staff is often a ‘tick-box’ exercise with-out proper investment; agencies and donorsneed to take it seriously to have an impact. Also,we often fail to appreciate that capacity develop-ment is two-way; national staff have a wealth ofcontextual knowledge that is critical to the im-plementation of field research.

Experiences from the REFANI Project in Nigerand ACF research have found that recruiting na-tional staff with the necessary research skills andexperience is challenging. This has led to opera-tional/support staff not being hired as plannedand overburdening some existing staff, who be-came responsible for both research and pro-gramme activities.

A number of lessons were learned. First, it isimportant to be realistic when recruiting researchstaff where capacity is known to be limited, i.e.appoint international staff or include capacity-building for national staff if time and budget per-mit. Secondly, recruit adequate numbers of extrapersonnel to support the implementation of re-search activities. Thirdly, having a part-time data-base manager is invaluable, given the need forremote data management in many contexts. Fi-nally, paying casual-hire enumerators bonuses forundertaking all data collection rounds (e.g. base-line, mid-term and endline studies) saves timeand money, helps build capacity (e.g. through re-fresher training) and supports data quality. Staffturnover at field level is common; for REFANI, the

oversight and continuity of an HQ-based re-search coordinator was a successful element inthe research implementation.

Monitoring and indicatorsData collectionDuring the Concern research studies, many per-sonnel hired as study staff had different expecta-tions about the quality and types of datacollection required, given their experiences ofworking for NGOs rather than academic institu-tions (which have more rigorous standards fordata quality). External parameters such as deteri-orating security can constrain access for interna-tional research staff, which limits opportunitiesfor researchers to provide support and oversight.For example, during the UCL REFANI study, datacollection was undertaken with increasingly lim-ited access for UCL’s international research staffdue to deteriorating security. To compensate, theteam conducted relatively lengthy trainings andhired consultants to provide support as trainersand supervisors. Tablets were used for data col-lection which posed some practical challengesbut at the same time facilitated real-time accessto recently collected data at office level in Tahoua,Niamey and London. This proved invaluable forremote data management.

IndicatorsHaving a comprehensive theory of change re-garding nutrition-related issues is key to under-standing research findings. To be truly effective,research needs to go beyond simple quantitativefrequencies to learn which programme compo-nents contribute to improved nutrition out-comes; how and why; and whether the results aregeneralisable. Rigorous quantitative methods,complemented by qualitative research, are nec-essary to answer these questions. SPRING hassuccessfully used mixed-methods research in sev-eral countries, but finding time and resources tofully analyse data, especially qualitative data, re-mains challenging.

Project indicators, while important for ac-countability, have limitations: they may under-emphasise or fail to capture key factors which canhave direct impact and could benefit decision-making. Outcome indicators are unique to eachcountry, making cross-country comparisons diffi-cult. At country-level, however, they do allow fortailoring research and evaluation to local needs.This can help build ownership and investment,since the indicators measure what is directly rel-evant to countries.

MonitoringResearch challenges usually relate to projectmanagement, particularly regarding data collec-tion and monitoring. Quality data collection andmanagement can be lacking in the nutrition andhealth sectors. It is therefore essential to invest ingood monitoring and evaluation systems andlink this with observational research, resulting instrong data being embedded in programmes.This requires working with operation teams, bet-ter use of the field data documented routinely orthrough audit, and opening discussions around

this. Investing in a good MEAL (Monitoring, Eval-uation, Accountability, and Learning) system iskey. Action Against Hunger, for example, is work-ing on a tool (NEAP: Nutritional Evaluation As-sessment; http://bit.ly/2ktDtuW) to improve theassessment of nutritional outcomes in pro-grammes at field level.

Ethical dilemmasControl groups present an ethical dilemma, par-ticularly in resource-constrained settings. Theidea of targeting an intervention based on ran-dom selection rather than on need presents op-erational agencies with a real quandary. InConcern’s experience, ways were found to lever-age a control or comparison group while main-taining what was considered an acceptable levelof accountability to the communities with whichthey work.

In Chad, the same intervention was providedto the control group, albeit three years later. InNiger, a comparison group – not a full controlgroup – was used, whereby both study arms re-ceived the same total amount of cash but over sixmonths versus four months. In Somalia, the con-trol group (an internally displaced persons (IDP)camp close to the IDP camp receiving the cash in-tervention) was not randomly assigned. It wasidentified after it had not been prioritised to re-ceive cash following an independent targetingprocess. The study team had also devised an al-ternative study design and analysis plan in theevent that the control camp did become a targetfor cash; as it turned out, this did not happen dur-ing the three-month study period.

Research data sharing is also a challenge; afundamental question is whether data should beaccessible to all. Open data is transforming re-search methods and data treatment.

Multi-sector approaches andpartnerships and internationalpartnershipsThe implementation of longitudinal studies in-creasingly requires multidisciplinary approachesand the creation of international (and national)spaces to enable the necessary connections andpartnerships. Formal structures where NGOs andacademics convene are rare. Creating a formalforum for partnerships to develop can be criti-cally important. Even where financial and timeconstraints limit this type of collaboration, wemust try to capitalise more on sharing past expe-riences and greater investment in multi-sectorapproaches.

ConclusionDetailing the operational challenges of conduct-ing research in humanitarian contexts is impor-tant. It is also important to invest in longitudinaland multidisciplinary studies to help understandthe causes of undernutrition and the means tomanage it. These studies should be comple-mented by in-depth observational studies. Suchambitious projects should be managed by con-sortiums of NGOs and academics, supported bydonors willing to invest and innovate.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Panel discussions

How to ensure quality research uptake By Stephanie Stern

Stephanie Stern leads the Action Against Hunger LAB project which aims to reinforce the impact anduptake of knowledge. Before joining Action Against Hunger, she worked for the Strategy & AnalysisDepartment of Save the Children International and was a research fellow at IRIS, the French think tankon international relations and strategic affairs. Her work focused on the transformation of thehumanitarian system.

The ACF research conference, Paris, 2016, included a plenary session and panel discussion on how to ensure researchuptake for nutrition research in emergencies. This article summarises this session, reflecting contributions from PatrickKolsleren, Gent University (moderator); Abdoulaye Ka, Undernutrition Unit of the Senegalese Prime Minister’s office; ZviaShwirtz, Uptake and Communication Manager for the REFANI project; and Mahaman Tidjani Alou, Abdou MoumouniUniversity of Niamey, and plenary session panel discussion.

Defining research uptakeThe discussion began with panellists sharing per-spectives on what constitutes research uptake.Zvia Shwirtz asserted how important it is to define‘research uptake’ and what it entails before anyresearch uptake strategy (RUS) or activities areput in place. The definition agreed upon by allpartners for the REFANI project was: “The processwhereby research findings are communicatedand utilised by a target audience.” The definition,and associated strategy and activities, was basedon the premise that continuous stakeholder en-gagement and dissemination of evidence will ul-timately lead to research uptake. Tidjani Alousuggested that sharing research findings thathave social implications is a means to create ap-petite that can galvanise “social agitation”, as de-scribed by the sociologist and philosopher JürgenHabermas (Calhoun, 1992) around subjects ofinterest.

Two fundamental questions are: “How do weensure evidence is utilised by key stakeholders,both in policy and practice?” and: “When shouldresearch uptake activities begin?” The panel re-flected that an efficient RUS is more than just asum of activities; it is a continuous process thatshould take place throughout the entire researchproject cycle, with various challenges and op-portunities. These were examined in more detailduring panel exchanges and plenary discussion;a selection of these insights follows.

Uptake challengesAcademic and operational partners may havedifferent interests when collaborating on researchtogether. It is important to find a happy mediumbetween producing robust evidence and pub-lishing findings in journals, and sharing resultswith stakeholders as quickly as possible. Academicstypically seek robust evidence and finalised resultsbefore making any statements on findings or en-gaging in dissemination; this may conflict withoperational partners who want to apply theresults as quickly as possible and rapidly influencepolicy and practice.

Researchers tend to move on once a researchproject is completed, yet this is when the criticaldissemination and communication to key stake-holders needs to take place to ensure maximumimpact. At the same time, continuing local stake-holder engagement and local dissemination canbe difficult once a project has finished.

Keys to success and attentionImproving communicationContinuous engagement with stakeholders iscritical. Project information should be shared asit becomes available, e.g. through conferences,meetings with various stakeholders and donorsand by hosting specific events. Communicationwith stakeholders should not wait until resultsare published, but should start from the very be-ginning, when research questions are elaborated.But while traditional methods (publications, publicand technical debates, conferences, etc) have arole, these routes are arguably limited; they donot create genuine open dialogue. This reflects amissing link between researchers and practitionersthat enables coherent dialogue; practitioners and

researchers may speak to each other, but nottruly understand or hear what the other is saying.The two communities have different logics andways of thinking. A third specialism may beneeded, comprising individuals capable of un-derstanding the research results and translatingthem into practical and understandable infor-mation. This concept has been described as “socialmediation” (Nassirou Bako-Arifari et al, 2000);social mediators support the research during theentire process and once the results are issued,work with the public sector to see where theycan make things change.

A critical question, given that knowledge pro-duced can have political and social impact, is:“What is the researchers’ responsibility in media-tion?” An example was shared from Niger whereresearch on water points enabled resolution ofdecades-long conflict between two villages re-garding land issues. The researchers were involvedin mediation with local government. Such medi-ation by researchers has limitations as he/shemay have limited familiarity with the context.

Research uptake panelists (l to r): Patrick Kolsleren,Abdoulaye Ka, Mahaman Tidjani Alou and Zvia Shwirtz

Mau

rine

Tric

菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Panel discussions

Patrick Kolsteren, contributing to plenary discussions

Mau

rine

Tric

Researchers publish in a format that is not ac-cessible or is too technical for most stakeholders.Eighty-two per cent of articles published in thehumanities are not even cited once. Of those ar-ticles that are cited, only 20 per cent have actuallybeen read. Half of academic papers are neverread by anyone other than their authors, peer,reviewers, and journal editors (Lattier, 2016). Thefora where research information is exchangedand discussed are not frequented by practitionerstakeholders and decision makers. Research in-formation should be presented to stakeholdersin a format that is short and gives a clear conclusionso that the information can be quickly translatedinto a decision. Important efforts are needed toimprove communication between stakeholdersand researchers and identify new ways to com-municate coherently and succinctly.

Fulfilling users’ expectationsPractitioners and agencies need evidence. TheUN Food and Agricultural Organization (FAO), forexample, is a “knowledge organisation” that worksto create linkages between research and politicsand between research results and policy and pro-gramming. The FAO works with research instituteson subjects defined by FAO; the primary goal isnot scientific publication but to have a guidancenote with a strong operational focus to help pro-grammers and policy makers. More broadly, itwas observed that there is often a discrepancybetween expectations of the researcher and thefinal user. Scale-up is also an important issue;final users need guidance on how and whetherresults are scalable or applicable to other contexts.

Upstreaming research instrategic planningThere are important lessons to learn from up-streaming research in strategic planning at countrylevel. In Senegal in early 2000, for example, eval-uations were used to investigate the effectivenessof the national nutrition programme, which provedextremely useful to prove the impact of the ap-proach and strategy. When the government de-cided to adopt a multi-sector approach to fightundernutrition, research played a key role indefining what needed to be implemented, sup-porting strategic and operational guidance. Themulti-factor and multi-sector character of com-batting malnutrition makes the challenge of en-suring coherence between the needs of researchersand operational actors even more interesting. Atcountry level, we are typically tasked with settingup multi-actor, multi-service platforms at house-hold, community and government levels. Researchchallenges play out at all tiers.

In Senegal, research is well positioned in nu-trition policy as a transversal element to supportthe definition and effective implementation ofstrategies aligned with the needs of the country.It fuels the choices of different strategies in theplanning stage, supports implementation by pro-viding guidance on how to ensure the effectivenessof interventions and is fundamental in providinginformation to the M&E system to explain successesand failures.

Research still faces multiple challenges, includ-ing: the lack of French language publication/translation of research that hinders uptake in Fran-cophone countries; how to adapt to different na-tional and local contexts; the need to have dialoguewith decision-makers and operational staff (prac-titioners); and the critical mass of skills that mustbe created to generate evidenced actions.

Bridging the gap betweenresearchers and practitionersA more holistic approach to knowledge manage-ment is needed, involving more than sporadicannual discussions and more open, transparentand continuous dialogue. We need to broadenour horizons and set up shared knowledge plat-forms, breaking down the divides that currentlyexist between researchers, practitioners and de-cision makers. This is how the Senegalese gov-ernment is approaching its nutrition policy, gath-ering all the concerned stakeholders round thetable – including academics and scientists engagedin all the different sectors impacting nutrition –and monitoring and evaluating the impact ofthis holistic approach.

Action-driven researchThere are three dimensions to research uptake.There is the needs aspect that is the problem athand; the demand side expressed as a need forinformation to support decision making; and theoffer, i.e. what researchers produce. In an idealsituation, these three dimensions overlap, butoften in practice they do not. Researchers offerresearch results within the perspective of theiracademic freedom. When this is offer only, theresearch may be innovative but is still perceivedby stakeholders as useless, because it does nothelp them make decisions. From the demandside, stakeholders have difficulties expressing

their need for information in a format to whichresearchers can respond. Given that the researchagenda is driven by external donors, the demandside cannot be met if the need does not fit donorpriorities. Local funding for research is oftenlacking, so that local demand has little or no trac-tion. Research to answer a particular questionmight not be innovative enough or too imple-mentation-oriented, making chances of publicationslim; academics may be less interested in pursuingsuch research from the outset.

Uptake of research results can be enhancedby identifying the questions stakeholders havefrom the beginning and responding with researchto answer these questions specifically. Researchuptake should be considered as a participatoryprocess, engaging all the stakeholders at differentlevels and moments of the research cycle. Its effi-ciency lies in the combination of various elements:responding to a question which interests donors,stakeholders (including beneficiaries) and re-searchers and ensuring all the concerned actorsare adequately informed and engaged in the re-search process, the dissemination of its resultsand their application to improved policies andpractices.

To view a video of this panel debate, visithttp://bit.ly/2kAciSq

ReferencesHabermas and the Public Sphere, ed. Craig Calhoun.Cambridge, MIT Press, 1992, p.109-142.

Nassirou Bako-Arifari T, Bierschenk G, Blundo Y, Jaffré T,Alou-Apad. Karthala, 2000. Une anthropologie entrerigueur et engagement.

Daniel Lattier. Why Professors Are Writing Crap ThatNobody Reads. October 26, 2016.www.intellectualtakeout.org/blog/why-professors-are-writing-crap-nobody-reads


菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

A cluster RCT to measure the effectiveness of cash-based interventions on nutrition status in Pakistan

By Bridget Fenn

Summary of conference abstracti

Bridget Fenn is anepidemiologist with abackground innutrition. She iscurrently a consultantfor the Emergency

Nutrition Network (ENN) and is PrincipleInvestigator on the Research in FoodAssistance for Nutrition Impact (REFANI)Pakistan study involving seasonal cashtransfers (cash and vouchers) and theirimpact on childhood nutritional status inhumanitarian settings.

Video footage of the conference presentationis available at http://bit.ly/2kAe7e5

Location: Pakistan

What we know: Evidence of theimpact of cash-based interventionson nutrition outcomes is limited.

What this article adds: A four-arm(standard cash (SC), double cash(DC), fresh food voucher (FFV),control) parallel, longitudinal,cluster randomised controlled trialwas implemented over six monthsby Action Against Hunger and ENN.Primary outcomes were weight-for-height z scores (WHZ) <-2 andmean WHZ in children under fiveyears old. Preliminary results founda significant decrease in risk ofbeing wasted (WHZ <-2) with DConly and significant improvementsin mean WHZ for DC and FFV. Allthree interventions saw asignificant decrease in bothstunting and mean height for age.There was a significant decrease inmean haemoglobin (Hb)concentration for children andmothers in the FFV and for mothersonly in the SC arm. A pathwaycausal analysis is underway thatwill greatly aid interpretation, withresults available mid-2017.

BackgroundThe Research into Food Assistance for NutritionImpact (REFANI) consortium comprises two op-erational partners; Action Against Hunger (leadagency) and Concern Worldwide, and two aca-demic/research partners; ENN and University Col-lege London (UCL). REFANI is a three-yearresearch project funded by UK Aid and co-fi-nanced through funding from the EuropeanCommission (EU & ECHO). The overarching aim ofREFANI is to increase the evidence base of cash-based interventions (CBIs) on nutrition outcomesin humanitarian settings by addressing a numberof evidence gaps. The use of CBIs among human-itarian agencies to prevent wasting in children isincreasing, but questions remain on how best toincorporate CBIs into emergency programmes tomaximise their success in terms of improved nu-trition outcomes.

The REFANI Pakistan study is a collaborationbetween Action Against Hunger and ENN, set inDadu district, Sindh province. Dadu district islargely agrarian, dependent on crop production,livestock keeping and agriculture labour. The ma-jority of the population are highly vulnerable toshocks, especially the poorest households, andthere is a lack of alternative income sources, fur-ther constrained by lack of opportunities. Dadudistrict experiences frequent flooding, droughtsand high temperatures (above 45°C).

MethodsThis study involved a four-arm, parallel, longitudi-nal, cluster randomised controlled trial (cCRT)(registered trial number ISRCTN107615320). Theprotocol has been published1. Three CBIs were im-plemented: two unconditional cash transfers (a‘standard cash’ (SC) amount of 1,500 Pakistan Ru-pees (PKR) and a ‘double’ cash (DC) amount of3,000 PKR) and one fresh food voucher (FFV) witha value of 1,500 PKR, which could be exchangedfor specified fresh foods (fruits, vegetables andmeat). A fourth arm acted as the control groupand received no additional intervention beyondthe basic activities implemented by ActionAgainst Hunger that were provided to all groups. The SC was set to equal the amount disbursed byPakistan’s national safety net programme, the Be-nazir Income Support Programme (BISP). The cashcomponents were disbursed on a monthly basiseither by mobile banks that travelled to a centrallocation for some of the participating villages orthrough central banks that served a number of vil-lages. The FFVs were disbursed to participating

households at village level. All three interventionswere delivered with verbal messages that childrenshould benefit from the transfers.

The interventions were implemented over sixconsecutive months (July to December 2015) andtargeted to mothers from poor/very poor house-holds with a child 6-48 aged months at baseline.The implementation and the use of the CBIs weremonitored both quantitatively and qualitativelythrough monthly questionnaires or quarterlyfocus group discussions and key informant inter-views.

The main research question assessed the ef-fectiveness of different CBIs at reducing the riskof undernutrition during the lean season. The pri-mary outcomes were weight-for-height z scores(WHZ) <-2 and mean WHZ in children under fiveyears old. The study also encompassed a mixed-methods process evaluation to help interpret theresults and a costs and cost-effectiveness analysis(results not presented here).

ResultsThe results presented here are a summary of theshort-term impact of CBIs on nutrition outcomes.The full analysis of both short and medium-termterm impacts is forthcoming. The group with thehigher amount of cash (DC) saw a significant de-crease in risk of being wasted (WHZ <-2) com-pared to the control group. There were nosignificant differences in risk of being wasted foreither SC or FFV arms. Both the DC and FFV armssaw significant improvements in mean WHZcompared to the control arm. All three interven-tions saw a significant decrease in both stunting(height-for-age z-score (HAZ) <-2 and <-3) andmean HAZ compared to the control group. In theFFV arm, there was a significant decrease in meanhaemoglobin (Hb) concentration for children andmothers and for mothers only in the SC arm.

Lessons learnedThe results have identified a number of questionsthat still need to be answered and for now re-quire careful interpretation. In terms of risk ofbeing wasted, we need a better understandingof why children in the DC arm were significantlyless wasted. This will be attempted through apathway analysis whereby different pathways inthe causal framework will be quantified. It was

1 http://bmcpublichealth.biomedcentral.com/articles/10.1186 /s12889-015-2380-3

i Presented at the ACF research conference, November 9th, 2016.


菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Risk factors for severe acute malnutrition in infants<6 months old in semi-urban Bangladesh:a prospective cohort study to inform future assessment/treatment tools

By M Munirul Islam, Yasir Arafat,Nicki Connell, Golam Mothabbir,Marie McGrath, James Berkley,Tahmeed Ahmed, and Marko Kerac


M Munirul Islam and TahmeedAhmed both work at the Nutritionand Clinical Services Division,International Centre for DiarrhoealDisease Research, Bangladesh(icddr,b). Yasir Arafat and GolamMothabbir work in the HealthNutrition and HIV/AIDS Sector, Savethe Children, Bangladesh. NickiConnell works in the Department ofGlobal Health, Save the ChildrenUSA. Marie McGrath works with theEmergency Nutrition Network, UK.James Berkley is based with theKEMRI/Wellcome Trust ResearchProgramme, Kenya. Marko Kerac isbased at the Department ofPopulation Health, London Schoolof Hygiene & Tropical Medicine,

Video footage of the conferencepresentation is available at:http://bit.ly/2kA6B33

Location: Bangladesh

What we know: The burden of acute malnutrition in infants < 6 months varies bycountry. Community-based case management for uncomplicated cases is lacking.

What this article adds: A recent study investigated the prevalence of acutemalnutrition in infants<6m in semi-urban Bangladesh (two seasons) and undertooka prospective cohort study to describe current outcomes of identified cases at sixmonths (180 days) of age. Prevalence of acute malnutrition was low post-harvest butincreased pre-harvest; from 0.4% to 5.9% for severe acute malnutrition (SAM) and2.8% to 10.1% for global acute malnutrition. At age six months, 24% of identifiedSAM cases (by eight weeks of age) and referred for available treatment (inpatient),remained severely malnourished. A range of infant and maternal risk factors forinfant SAM were identified, involving breastfeeding status, the nutrition and mentalhealth of the mother, infectious disease and water/sanitation/hygiene. A package ofcare is warranted in this age group.

Current WHO guidelines on severe acute mal-nutrition (SAM) management recommendoutpatient management of uncomplicatedacute malnutrition in infants under six

months of age (infants <6m), in line with the now-es-tablished treatment approach for older children(WHO, 2013). However, there is a lack of practicalguidance on how to identify those infants <6m at risk

and how to manage them. Current WHO case defini-tion for SAM in infants <6m is weight-for-length ofless than -3 Z-score (WLZ); visible severe wasting;and/or bilateral pitting oedema.

To inform the development of assessment toolsand treatment approaches for SAM in infants <6m, a

not possible to establish the minimum level ofcash required to have a significant effect; we canonly say that this threshold falls somewhere be-tween the amounts allocated in the SC and theDC interventions.

Regarding mean WHZ, it appears that childrenin the FFV arm were getting fatter but not taller,especially if interpreted with the lack of improve-ment in being wasted. As well as this, the Hb lev-els of children and mothers in the FFV arm weresignificantly lower compared to the controlgroup. We had hypothesised that the FFV wouldimpact growth and micronutrient status throughincreasing dietary diversity. However, while allthree arms showed a significant improvement inmother and child dietary diversity, this improve-ment was lowest in the FFV arm (highest in theDC arm). These results suggest that somethingunplanned was occurring in the FFV arm. It is pos-

sible that the vouchers themselves were too re-stricted, being dependent on what vendorsstocked (such as chicken being the only availablemeat). There were also anecdotal reports regard-ing vendors overcharging for food items re-deemed against the vouchers as a way to covertheir own administration fees in recovering thevoucher costs. In this respect, the actual transferamount given may have been lower. Another question that needs to be addressedwas the lack of improvement in Hb status in lightof the improvements in ponderal growth andprevalence of being wasted as seen in the DCarm. The explanation for this will also be at-tempted through a pathway analysis.The study setting presented a number of difficul-ties affecting data collection. The baseline surveytook longer than expected, since recruitment offemale enumerators was difficult and the datacollection coincided with Ramadan and reduced

working hours. Added to this, temperaturesreached 52oC, which not only affected researchteam working ability but also had an effect on thehaemocues (this was managed).

ConclusionThe results illustrate the impacts of different CBIson nutrition status. However, the theory ofchange regarding ‘how’ CBIs may influence nutri-tion outcomes in children is complex. In the RE-FANI Pakistan study, we show that it is not astraightforward task to simply interpret impactresults as working or not; understanding thepathways and processes through which CBIs areimplemented is essential to understand how bestto implement them. Such analyses will be com-pleted mid-2017.

For more information, contact: Bridget Fenn,email: [email protected]

1 Presented at the ACF research conference, November 9th, 2016.


菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Standard SAM Normal Expanded SAM

Non-exclusive breastfeeding 34% 13% 23% p=0.01

Duration of exclusivebreastfeeding

3.9 weeks 5.7 weeks 4.5 weeks p<0.001

Dissatisfaction withbreastfeeding

22% 10% 7% p<0.001

Mothers educated beyondschool year five

56% 71% 77% p<0.023

Maternal mental health/distressscore (max score is 20)*

8.4 6.8 7.5 p<0.008

Maternal mid-upper-armcircumference

233mm 246mm 241mm p<0.012

Infant illness episodes requiringhospitalisation

21% 6% 9% p<0.001

Household income/month $89 $114 $114 p=0.007

* WHO Self Reporting Questionnaire 20. World Health Organization (WHO). A User’s Guide to the Self Reporting Questionnaire. WHO/MNH/PSF/94.8. http://apps.who.int/iris/bitstream/10665/61113/1/WHO_MNH_PSF_94.8.pdf

Table 1 Differentiating characteristics between cohorts

Agnè

s Var

rain

e-Le

ca /

ACF

Libe

ria

MUAC case definition does not exist for infants <6m;data were collected for research purposes only

study was conducted in semi-urban Bangladeshin 2015/2016 with the following objectives:1. To estimate the prevalence of infants <6m

with acute malnutrition in the community;2. To develop an assessment tool/case defini-

tion checklist for infants <6m with acute malnutrition; and

3. To describe current outcomes following infant <6m with acute malnutrition.

The study involved two prevalence surveys (indistinct seasons) and a prospective cohort studyof three infant groups (77 in each group), fol-lowed from 4-8 weeks to 180 days post-partum.The groups comprised:

Standard SAM: WLZ <-3 and/or bipedal oedema;

Normal: WLZ ≥-2 to <+2 z-scores, no oedema;

Expanded SAM: MUAC <115mm but WLZ ≥-2, no oedema.

MUAC case definition does not exist for infants<6m; data were collected for research purposesonly. Mother/caregiver interviews at enrolmentassessed potential risk factors. Infants with ‘Stan-dard SAM’ were referred to existing services fortreatment (inpatient care) according to existingprotocols. The primary outcome was nutritionalstatus at age completion of 180 days.

Preliminary resultsPrevalence surveyThe prevalence of GAM and SAM were low in thepost-harvest period but increased pre-harvest;from 0.4% to 5.9% for SAM and 2.8% to 10.1% forGAM. Severe underweight (weight for age < -3 Z-score (WAZ)) slightly increased (severe: 5% to6.1% and severe/moderate combined, from14.4% and 16.3%).

Cohort studyAt enrolment, ‘Standard SAM’ and ‘ExpandedSAM’ were younger than ‘Normal’ infants (5.1, 5.5,6.5 weeks respectively, p<0.001). A selection ofcharacteristics that differed between these

groups at enrolment is shown in Table 1. Type oftoilet was significantly different among thegroups; Standard SAM had more people with a pitlatrine vs. a flushing toilet compared to the othertwo groups. Handwashing and source of waterwere not significantly different among thegroups. Duration of breastfeeding was not signifi-cantly different among the groups.

At age six months, ~24% of ‘Standard SAM’, 1%of the “Normal”, and 5% of ‘expanded SAM’ infantshad SAM (p≤0.001). Three infants with ‘StandardSAM’ died; compared to none in the ‘Normal’group and one in the ‘Expanded SAM’ group. In the Standard and Expanded SAM groups, sig-nificantly fewer infants were still being breastfed,more infants were fed anything other than breast-milk at enrolment and endline, and more infantswere given animal milk at an earlier age, com-pared to the Normal group. Duration of exclusivebreastfeeding was also shorter in both groups,mothers breastfed less often, had higher mentalhealth/distress score and were significantly lesssatisfied when asked how breastfeeding was

going. Maternal Body Mass Index (BMI) andMUAC were lower in the Standard SAM group.More infants in the Standard SAM group had atleast one episode of illness that required hospi-talisation at enrolment; this proportion (21%) hadincreased to 40% by endline.

Challenges to implementing the research in-cluded:• Access to villages in the rainy season;

surveyors had to cross many bodies of water with equipment including motorbikes.

• Randomly selected villages were geographi-cally spread out, so travel time was high.

• Measuring anthropometry in infants <6m is difficult, especially length.

• The number of questions in the questionnairewas a challenge as many were necessary to ensure comprehensiveness.

Discussion and conclusionsA range of maternal, infant and environmentalrisk factors are associated with SAM among in-fants <6m. Successful future treatments shouldfocus on a package of care rather than single in-terventions that include breastfeeding support;the nutrition, physical and mental health of moth-ers; infectious disease management; andwater/sanitation/hygiene conditions. Over onequarter of the infants identified with SAM at theoutset remained severely malnourished at sixmonths of age; this suggests inadequate provi-sion and/or access to treatment of SAM and arisky environment. It is necessary to distinguishthese vulnerable infants from those who had re-covered by six months; indicators in addition toanthropometry are probably necessary.

For more information, contact: Nicki Connell,email: [email protected]

ReferencesWHO. Guideline: Updates on the management of severeacute malnutrition in infants and children. Geneva: WorldHealth Organization; 2013.


菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Water, livestock, and malnutrition findings froman impact assessment of Community Resilienceto Acute Malnutrition programme in Chad

By Anastasia Marshak, Helen Young and Anne Radday


Anastasia Marshak is a Researcher at the FeinsteinInternational Center, Friedman School of NutritionScience and Policy, Tufts University. Helen Young isResearch Director at the Feinstein International Centre,Friedman School of Nutrition Science and Policy, TuftsUniversity. Anne Radday is Research ProgrammeManager at the Feinstein International Center,Friedman School of Nutrition Science and Policy, TuftsUniversity.

Video footage of the conference presentation isavailable at: http://bit.ly/2jYbWC5

Location: Chad

What we know: Evidence and mechanismsof impact are limited regardingprogrammes that combine nutrition withother interventions, such as health, water,sanitation and hygiene (WASH) and foodsecurity.

What this article adds: A randomisedcontrol trial impact evaluation wasconducted of the Community Resilience toAcute Malnutrition (CRAM) project in Chad,a three-year, integrated, multi-sectorprogramme. The findings indicatesignificant programme impact, particularlyin relation to acute malnutrition, althoughprevalence of wasting (and stunting)remained high.

At endline, children in CRAM settlementswere significantly better off, including lowerprevalence of wasting, stunting and illness.WASH promotion activities also showed asignificant positive impact of CRAM andwere correlated to child nutrition outcomes.Differences in child outcomes betweensettled and former pastoralist communitiesmay be explained by seasonal factors andlivestock management practices affectingwater contamination by cattle. Thissuggests access to clean water is notsufficient; hygiene of the supply chain alsoappears critical.

The Dar Sila region of eastern Chadexperiences highly variable rain-fall, seasonal food insecurity andhigh prevalence of acute malnutri-

tion. In 2012, Concern Worldwide put inplace an integrated programme that com-bines nutrition, health, water, sanitationand hygiene (WASH) and food, incomeand markets (FIM) provision in Dar Silacalled Community Resilience to AcuteMalnutrition (CRAM). The programme wasdesigned to reduce child acute malnutri-tion in the face of seasonal shocks.

Concern collaborated with the Fein-stein International Center, FriedmanSchool of Nutrition Science and Policy atTufts University to carry out a randomisedcontrol trial impact evaluation to betterunderstand the level of programme im-pact and the mechanisms behind it. Threesurveys took place in November and De-cember of 2012, 2014 and 2015 in 69 set-tlements encompassed by the Concernprogramme area. This article summarisesthe impact of the CRAM programme andhighlights household and communitycharacteristics correlated with acute mal-nutrition in Dar Sila, Chad.

MethodThe study covered 1,400 households,spread evenly between 69 settlements.The survey collected information onhousehold demographics, socio-eco-nomic characteristics, food insecurity, ac-cess to natural resources, and childnutrition and morbidity. In addition, a

qualitative investigation was carried out in2013, 2015, and 2016 using focus groupsand key informant interviews.

All the data were adjusted for the sam-pling design and included populationweights. To establish programme impact,logit and ordinary least squares (OLS)analysis regression models were used forbinary and continuous outcome variablesrespectively. To take advantage of thepanel nature of the dataset, a random andfixed effects model was run using theweight-for-height z-score (WHZ) of thechild in the household with the lowestscore as the outcome variable, simply re-ferred to as ‘nutritional status’.

ResultsAt the endline, children in the CRAM set-tlements performed significantly betterthan the non-intervention group on a hostof key nutrition and health indicators.They had lower prevalence of wasting;higher WHZ; lower prevalence of chronicmalnutrition (stunting); higher height-for-age z-scores (HAZ); and lower prevalenceof illness (Table 1). Even when controllingfor child, household and settlement char-acteristics using a random and fixed ef-

Control Treatment Significance

WHZ -1.13(-1.29 to -0.98)

-0.85(-1.02 to -0.66)

p < 0.05

Wasting (WHZ<-2) 0.21 (0.17 to 0.24)

0.15 (0.10 to 0.19)

p < 0.05

HAZ -1.27(-1.44 to -1.09)

-1.07(-1.23 to -0.92)

p < 0.1

Stunting (HAZ<-2) 0.37 (0.33 to 0.40)

0.30(0.26-0.35)

p < 0.05

Child sick in the past twoweeks

0.37(0.31 to 0.43

0.28 (0.23 to 0.34)

p < 0.05

Table 1Nutrition and health indicators at endline (mean with confidenceintervals in parentheses)

i Presented at the ACF research conference, Novem-ber 9th, 2016. Full report: Anastasia Marshak, HelenYoung and Anne Radday. Water, Livestock, and Malnutrition. Findings from an Impact Assessment of Community Resilience to Acute Malnutrition Program-ming in the Dar Sila Region of Eastern Chad, 2012-2015. Feinstein International Center, December, 2016. http://fic.tufts.edu/publications/


菀菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Control Treatment Significance

Borehole utilisation 46%(33 to 60%)

79%(66 to 91%)

p < 0.01

Transport container cleaned once a weekwith soap, closed, and “looks” clean

12%(9 to 16%)

21%(15 to 26%)

p < 0.01

Know the two times for handwashing 57%(50 to 64%)

67%(60 to 74%)

p < 0.05

Table 2 WASH indicators at endline (% with confidence intervals in parentheses)

Ana

stas

ia M

arsh

ak

fects regression model, both being in the inter-vention group (p<0.1) and moving from not re-ceiving CRAM to receiving CRAM (p<0.1) wassignificantly correlated to better household childnutritional status (i.e. minimum household WHZ).

The WASH promotion activities also showed asignificant positive impact of CRAM and werecorrelated to child nutrition outcomes, specifi-cally in relation to the following variables: greaterutilisation of boreholes; greater reports of regu-larly washing the transport and storage containerwith soap; and greater knowledge around thetwo main times for handwashing (Table 2). How-ever, the proportion of respondent’s correctlypracticing handwashing significantly decreased(p < 0.05) in the intervention settlements only,suggesting increased knowledge did not trans-late into changes in behaviour. A link was ob-served between water consumption and CRAMin relation to exclusive breastfeeding. Respon-dents in the intervention settlement were signifi-cantly more likely to exclusively breastfeed at theendline, primarily driven by a reduction in givingwater to children under the age of six months(78% of mothers reported giving water in thenon-intervention settlements versus 54% in theintervention settlements (p < .05)). In the regres-sion analysis, utilisation of a borehole (as op-posed to an open water source) was significantly(p<0.01) and negatively correlated with childacute malnutrition. However, the same relation-ship was not observed in the non-interventionsettlements, indicating a potential role for hy-giene practices along the water chain in reducingcontamination of water.

A possible source of water contamination isthe concentration of cattle in a village. In the re-gression analysis, while individual livestock own-ership correlated with better child nutritionoutcomes (p<0.1), as the concentration of cattlein a village increases, so do rates of acute malnu-trition (p<0.05). However, children who lived informer pastoralist settlements had consistentlybetter nutritional status (p<0.05), even thoughhouseholds in these settlements were signifi-cantly more likely to own more cattle (p<0.01).

Differences in livestock management and sea-sonal location of livestock, uncovered in the qual-itative research, may be driving this observeddifference in child nutrition outcomes. House-holds in the former pastoralist damre communi-ties, specialised in cattle production, reportedmigrating farther with their herds during the dryseason to areas with permanent rivers. Villagehouseholds brought their cattle back to the vil-lage at the height of the dry season and utilisednearby water sources, including the borehole.

DiscussionThese findings indicate significant programmeimpact, particularly in relation to acute malnutri-tion. One cautionary note is that, for the durationof the programme, the prevalence of global acutemalnutrition remained around 15% or above,while stunting prevalence was between 30 and45%. There is no statistical evidence that CRAMreduced the rate of malnutrition in the interven-tion settlements; rather, unlike in the non-inter-vention settlements, malnutrition rates did notincrease over time. These continuing high ratesof acute malnutrition, and the increases seen inthe non-intervention settlements, are causes forconcern and indicative of the extreme vulnerabil-ity of these communities as they emerge frommore than a decade of protracted crises. There is,however, greater resilience in CRAM settlementsas a result of the programme.

The data also offer clues regarding the mech-anisms related to impact and how impact couldbe increased. Using a borehole without propertraining on the water chain does not in itself sig-

nificantly decrease rates of malnutrition. Routinewater testing found that, while contaminationlevels (coliforms) of borehole water at the pointof collection were low to non-existent, they in-creased at certain points along the water chain(from borehole to transport container to storagecontainer). This finding suggests that the positiveimpact of CRAM on malnutrition may be via theWASH activities that are focused on reducing therisk of contamination of potable water furtheralong the water chain. These activities promotegood hygiene in relation to water containers.

A possible source of water contamination isthe concentration of cattle in a village. The asso-ciation between village cattle density and childnutritional status is a possible explanation of con-tamination. Previous literature had identified sev-eral pathogens associated with diarrhoea anddeath among infants. One of those pathogens –cryptosporidium parvum – is water-borne, ispassed from cattle to humans, and has been toshown to cause rather than simply exacerbatemalnutrition.

These findings indicate that CRAM had a sig-nificant programme impact on both acute andchronic malnutrition. The evaluation also ex-plored why the CRAM programme might haveworked and how it could be improved. Access toclean water appears to play an important al-though not sufficient role; hygiene practicesalong the water chain might also be critical inpreventing contamination along the water chain.

For more information, contact: Anastasia [email protected]


菀菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Inflammation and moderateacute malnutrition in children: a cross-sectionalstudy in Burkina Faso

By Bernardette Cichon


Bernardette Cichon is undertaking her PhD at theUniversity of Copenhagen. She has an MSc in PublicHealth Nutrition from the London School of Hygieneand Tropical Medicine and more than eight yearsexperience working with nutrition in humanitariancontexts.

The treatFOOD project1 is a collaboration betweenALIMA, MSF-Denmark and the University of . It wasfunded by DANIDA, MSF-Denmark, MSF-Norway,WFP, USAID, ECHO and Arvid Nilsson’s Foundation.

Video footage of the conference presentation isavailable at: http://bit.ly/2jYbWC5

Location: Burkina Faso

What we know: The role of morbidity inmoderate acute malnutrition (MAM) isnot well understood.

What this article adds: An observationalstudy in Burkina Faso, using baselinedata from a randomised controlled trial,described morbidity and inflammation inchildren with MAM. Almost 90% ofchildren with MAM in this setting had aninfection and/or inflammation. Maternalhistory reported a 38% infection rate inthe previous two weeks; 71.8% were illon the day of visit. Most prevalentdiagnosed illnesses were malaria (40.2%),lower respiratory tract infections (23.2%),and upper respiratory tract infections(14.6%); fever was common. A total of10.7% and 46.5% of asymptomaticchildren had elevated acute phaseproteins (CRP and AGP, respectively),suggesting sub-clinical infection. Thiswas largely unexplained by maternalreports and clinical examination. Moreemphasis on identification and treatmentof infections as part of MAM treatmentand investigation into how this affectsnutritional status and recovery is needed.

BackgroundMorbidity plays an important role in the de-velopment of and recovery from malnutri-tion. Morbidity in children with moderateacute malnutrition (MAM) has not been de-scribed in detail and it is unclear how mor-bidity compares to serum levels of acutephase proteins (APPs) which indicate sys-temic inflammation that can impede re-sponse to therapeutic nutritional inter-ventions. The objective of this study was todescribe morbidity and inflammation inchildren with MAM and to assess to what ex-tent maternally reported and clinically diag-nosed morbidity explains the variation inAPPs.

MethodsThe data for this observational study werebaseline data collected as part the treat-FOOD trial, a randomised controlled trialtesting effectiveness of food supplementsfor treatment of MAM, carried out in the Pas-soré Province, Northern Region, BurkinaFaso. Children aged 6-23 months with MAM,resident in the catchment area and whoseparents/guardians consented to participate,were included. Recruitment took place fromSeptember 2013 until August 2014.

Socio-demographic, anthropometricand morbidity data were collected bytrained staff. Morbidity data collection in-cluded a patient history based on 14-daymaternal recall of symptoms and a physicalexamination carried out by study nurses. Ve-nous blood (2.5 ml) collected from the armwas used for diagnosis of malaria, using arapid diagnostic test (RDT), and to measureserum concentrations of C-reactive protein(CRP) and α1-acid glycoprotein (AGP). Feverwas defined as an axillary temperature >37.5 °C. Upper and lower respiratory tract in-fections were diagnosed by experiencednurses based on an adapted version of theIntegrated Management of Childhood Ill-nesses (IMCI) guidelines. Diarrhoea was de-fined as three or more loose, watery stoolsper day based on information provided bythe mother. The thresholds used for defining

elevated APP levels were CRP >10mg/l andAGP >1g/l. Multivariate ANCOVA modelswere used to explore the associations be-tween morbidity and CRP as well as AGP.These models were also used to determineto what extent morbidity explains variationin APPs.

ResultsA total of 1,609 children were enrolled in thestudy. Over half (54.6%) of participants werefemale. Prevalence of stunting (height-for-age <-2 z score) was 37.7%. The mean (SD)age was 12.3 (4.8) months.

Mothers reported illnesses in the two-week period prior to admission in 38% ofchildren. Furthermore, 71.8% of childrenwere ill on the day of the visit according tothe physical examination by the studynurse. The most prevalent illnesses diag-nosed by the nurse were malaria based onpositive RDT (40.2%), lower respiratory tractinfections (23.2%) and upper respiratorytract infections (14.6%). Fever was alsocommon (17.7%). Almost a quarter (24.2%)and two thirds (66.4%) of children hadserum CRP >10 mg/l and serum AGP >1 g/l,respectively.

Positive malaria RDTs were more com-mon among children admitted based onmid-upper-arm circumference (MUAC) onlythan children admitted based on weight-for-height z score (WHZ) only, after adjustmentfor age and sex (38% vs 26%, p<0.001). Morechildren had lower respiratory tract infectionif they were admitted based on WHZ onlycompared to MUAC only, after adjustmentfor age and sex (29% vs 21%, p=0.006). Therewere no associations between other symp-

1 See www.treatfood.org

A Treatfood project research site,Province du Passoré, Burkina Faso

i Presented at the ACF research conference, November 9, 2016. Published research: B Cichon, F Fabiansen, CW Yaméogo, MJH Rytter, C Ritz, A Briend, VB Chris-tensen, KF Michaelsen, R Oummani, S Filteau, P Ashorn,S Sheperd and H Friis. Children with moderate acute malnutrition have inflammation not explained by maternal reports of illness and clinical symptoms: a cross-sectional study in Burkina Faso. BMC Nutr. 2016;2:10.1186/s40795-016-0096-0.


菀菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Research

Upcoming research shared at ACF research conference

At the ACF research conference, November 6th, 2016, experiences were shared from a number of studieswhere final results will be made available in 2017. A snapshot of what to expect, video footage of the con-ference presentations and contacts for the studies, are included below.

The PROMIS project: integrating the prevention of child undernutrition into community-based management of acutemalnutrition programmes in Senegal, Mali and Burkina Faso

Innovative Approaches for the Prevention ofUndernutrition (PROMIS) is a three-year (2014-16) project funded by Global Affairs Canada

that seeks to prevent and improve the treatmentcoverage of acute in children in Burkina Faso,Mali, and Senegal. The intervention is imple-mented Helen Keller International and evaluatedby the International Food Policy Research Insti-tute (IFPRI) using mixed study designs. ThePROMIS project looks to improve performanceand beneficiary coverage of current community-based management of acute malnutrition(CMAM) programmes by integrating a package ofpreventive measures into child acute malnutri-tion (AM) screening offered by different deliveryplatforms. Impacts on child AM prevalence andincidence are hypothesised. In Mali and BurkinaFaso, the programme’s impact is assessed usinga cluster randomised controlled design, while inSenegal, a smaller study assesses the pro-gramme’s feasibility in a peri-urban setting.

In Mali, the delivery platform consists ofmonthly community health volunteers-led villagegatherings of caregivers with children 6-23months of age to screen children for AM and todeliver the enhanced preventive package(strengthened Behaviour Change Communica-tion (BCC) on nutrition and health, and a smallquantity of lipid-based nutrient supplement (SQ-LNS)). The comparison group receives monthly vil-lage-based group BCC and screening for child AM.In Burkina Faso, well-baby consultations (WBC) inhealth centres is the primary platform to offermonthly screening for AM among infants startingat birth. Caregivers of infants from 0-6 months al-located to the intervention group that participatein WBC receive strengthened BCC on nutritionand health after regular child AM screening. Fromthe age of six months onwards, the provision ofpreventive SQ-LNS is added. The comparisongroup receives unspecific BCC as prescribed bythe national policy. In Senegal, community healthworkers trained by local NGOs organise group

BCC, screen children for AM and distribute SQ-LNSto caregivers of children 6-23 months of age.

In Mali and Burkina Faso, the programme’s im-pact is assessed by two study designs. A baseline-endline comparison study assesses theprogramme’s impact on the prevalence of acutemalnutrition, whereas a longitudinal study withmonthly follow-up measurements during 18months evaluates the preventive impact on theincidence of child AM. A mixed methods processevaluation assesses the programme’s impactpathways and feasibility. Finally, a cost-effective-ness study will provide insight into the economicdimension of this integrated programme. Resultswill be available at the end of 2017 and through-out 2018.

Video footage of the conference presentation isavailable at: http://bit.ly/2k8Zodl

For more information, contact: Lieven Huybregts,email: [email protected]

toms, illnesses and APP levels and admission cat-egories. A total of 10.7% (n=36) and 46.5%(n=157) of asymptomatic children had a CRP >10mg/l and AGP >1 g/l, respectively. Only 19% ofchildren had normal CRP and 12% had normalAGP in the absence of symptoms.

History of fever as well as nurse-documentedfever, malaria, respiratory tract infections and skininfections were associated with higher levels ofboth APPs. History of cough and diarrhoea at theinclusion visit was associated with higher AGPonly. Overall, morbidity data only explained asmall amount of the variation in APP levels (ad-justed R2 below 0.2 in all tested models).

Lessons learnedThis cross-sectional study has shown that almost90% of children with MAM in this setting had an

infection and/or inflammation. MAM treatmentprotocols usually only provide for supplementaryfood and routine medication such as deworming,vitamin A and iron and folic acid supplements.These results indicate a possible need for moreemphasis on identification and treatment of in-fections as part of MAM treatment.

Furthermore, elevated APP levels in childrenwithout identified symptoms are not uncommonand morbidity data explained only a small pro-portion of the variation, as demonstrated by theadjusted R2 which was <0.2 in all models, bothindicating a presence of sub-clinical inflamma-tion. It is unclear what causes this sub-clinical in-flammation and whether it affects nutritionalstatus and response to treatment. Possible expla-nations for the sub-clinical inflammation cited bythe authors include missed infections; the factthat APPs can rise during the incubation phase of

a disease before clinical symptoms become ap-parent or remain elevated during convalescence;and the presence of other conditions such as en-vironmental enteric dysfunction (EED); recentvaccinations; cooking with biomass fuels; and ex-posure to toxins that may elicit an acute-phaseresponse.

ConclusionMorbidity among children with MAM in this set-ting is common but maternal reports and clinicalexamination explained only a small proportionof the variation in APPs, indicating a presence ofsub-clinical inflammation. Further research isneeded into the causes of this sub-clinical inflam-mation, as it could affect nutritional status andsuccess of MAM treatment.

For more information, contact: Bernadette Chichon, email: [email protected]


菀菀

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Upcoming research shared at ACF research conference

The MAM’Out research project aims to testthe effectiveness and cost-effectiveness ofa seasonal and multiannual cash transfer

programme, likely to influence multiple under-lying causes of undernutrition. The two-armcluster randomised controlled trial was imple-mented in 2015 and 2016 in Burkina Faso. Thestudy was informed by a Nutrition Causal Analy-sis that identified a number of perceived causesof undernutrition in the population: women’s fi-nancial insecurity, inadequate birth spacing andpoor access to potable water. A formative re-

search was conducted to assess the relevance ofa cash based intervention and provided detailedoperational guidance on the target population,the type of cash transfer, the seasonality, the de-livery mechanism and the amount to transfer.Participating households in the interventiongroup were offered seasonal unconditional cashtransfers (UCTs) from July to November, overtwo years (2013 and 2014). A monthly allowanceof 10,000 XOF (≈US$17) was given by mobilephone to mothers, identified as primary recipi-ents of the transfers. Mothers were told that the

cash transfer was to support their child’s devel-opment and to prevent undernutrition. Traineddata collectors performed home visits on aquarterly basis to collect quantitative data suchas child anthropometrics and morbidity, socioeconomic and demographic indicators. Resultswill be published in 2017.

Video footage of the conference presentation isavailable at http://bit.ly/2kAaAMG

For more information, contact: Freddy Houngbé,email: [email protected]

Effects of multi-annual seasonal unconditional cashtransfers on young children’s nutritional status andmorbidity in Burkina Faso: the MAM’Out trial

Delivering SAM treatment through communityhealth workers in Mali

Community-based management of severeacute malnutrition (SAM) has increased ac-cess to treatment, but coverage of cases re-

mains inadequate. Experience from otherplatforms show that modifications to the servicedelivery model, such as the delivery of malaria,pneumonia and diarrhoea health services byCommunity Health Workers (CHWs) at communitylevel, can lead to over 90% coverage of affectedcases. More evidence is needed to develop a sim-ilar model using CHW for the treatment of SAM,to allow Ministry of Health to adapt treatmentmodels to deliver higher coverage and perform-

ance. A clinical, prospective, multi-centre cohortstudy was conducted between February 2015 andFebruary 2016 by ACF in Kita in Southwest Mali,to investigate if SAM treatment delivered throughCHWs is as effective as treatment delivered athealth facilities. Secondary objectives were to as-sess the coverage, quality of care and cost-effec-tiveness of the intervention compared to routineoutpatient care. One cohort (consisting of fourhealth centres) followed a traditional outpatientmodel of treating SAM (control group) and thesecond cohort (consisting of three health centres)used CHWs to treat uncomplicated SAM cases in

the community (intervention group). The alloca-tion of treatment between the two groups wasrandomised. Clinical outcomes (cure, death anddefaulter rates), cost effectiveness, treatment cov-erage and quality of care were examined in boththe control and intervention groups. Results willbe made available in 2017.

Video footage of the conference presentation isavailable at: http://bit.ly/2jboaKj

For more information, contact: Pilar Charle, email: [email protected]

Effectiveness of adding a household WASH-package toa routine outpatient programme for severe acute malnutrition in Chad – the Ouadi’nut study

Action Contre la Faim is currently imple-menting a cluster randomised controlledtrial in Mao and Mondo health districts,

Kanem region in Chad, in partnership with theInstitute of Tropical Medicine in Antwerp, Bel-gium, and the Sahel Association of Applied Re-search for Sustainable Development (ASRADD)in Chad. The study is investigating the effective-ness of adding a household water, sanitation andhygiene (WASH) package to a routine outpatientprogramme for severe acute malnutrition. Theaim is to protect children against new episodes

of diarrhoea and other WASH-related infections,and to contribute to nutritional recovery. Thehousehold WASH-package includes: 1) House-hold water treatment and hygiene kit (watercontainer, water disinfection consumables, soap,cup, simple hygiene promotion leaflet with im-ages); 2) Weekly hygiene promotion sessions athealth centre level with others/caretakers of chil-dren admitted to the programme. Primary eval-uation outcomes are recovery and relapseproportions. Secondary outcomes include time-to-recovery, weight gain, longitudinal preva-

lence of morbidity (diarrhoea, vomiting, cough,and fever), and adherence to the householdWASH-package, hygiene and care practices ofthe mothers/caretakers.

The trial is registered at clinicaltrials.gov underthe identifier: NCT02486523. The final resultsand recommendations will be published in2017.

For more information, contact: Mathias Alt-mann, [email protected]


e ˝˘ˆ - enn · acf, and plenary discussion during the acf research conference, paris, 2016. t he...

Documents