Evidence on the impact of community health workers in the prevention, identification, and management of undernutrition amongst children under the age of five in conflict-affected or fragile settings: a systematic literature review

Background Malnutrition, specifically undernutrition, is a significant global challenge that contributes to nearly half of deaths in children under the age of five. The burden of undernutrition is disproportionately borne by conflict-affected, fragile settings (CAFS); children living in a conflict zone being more than twice as likely to suffer from malnourishment. Community health worker (CHW) models have been employed in CAFS to improve healthcare coverage and identify and treat illnesses. However, there lacks systematic evidence on the impact of CHW models in preventing, identifying, and managing child undernutrition in CAFS. We conducted this review to systematically evaluate evidence of CHW models in preventing, identifying, and managing undernutrition in children under the age of five in CAFS. Methodology This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting standards. The search strategy was developed using the Population-Intervention-Comparisons-Outcomes-Setting framework as a guide. Searches were performed using Ovid online database search platform, searching the databases of Ovid MEDLINE(R), COCHRANE, Embase Classic, Embase, Econlit, Global Health, SCOPUS, and Social Policy and Practice. Peer-reviewed publications were eligible for inclusion if they evaluated an intervention using a CHW model that aims to prevent, identify, or manage some form of undernutrition in children under five in a CAFS. Results We identified 25 studies—spanning 10 countries—that were included in the systematic review. CHW models were implemented alongside a variety of interventions, including behaviour change communication, supplementary foods, nutrition counselling, and integrated community health programmes. Key barriers in implementing successful CHW models include disruption of programmes due to active conflict, states of emergency, militancy, or political unrest; weak links between the community-based interventions and public health system; weak health system capacity that impeded referral and follow-ups; and cost of care and care-seeking. Key facilitators include CHWs’ connection to the community, close proximity of programmes to the community, supervision, and investment in high quality training and tools. Conclusions The findings suggest that CHW models may be effective, cost-effective, acceptable, feasible, and scalable in the prevention, identification, and management child undernutrition in CAFS. The study findings also confirmed a need for greater evidence in the field. These findings may inform policymaking, programme implementation, and design to strengthen best practices for CHW models addressing child undernutrition in CAFS.


Background
Malnutrition, specifically undernutrition, is a significant global challenge that contributes to nearly half of deaths in children under the age of five [1].Malnutrition is described as 'deficiencies, excesses, or imbalances in a person's intake of energy and/or nutrients' [1] and includes the condition undernutrition.Undernutrition is subdivided into four types: acute malnutrition, 1  stunting, 2 underweight, 3 and micronutrient deficiencies.Acute malnutrition, is characterised by wasting-defined as a low mid-upper arm circumference (MUAC) and/ or low weight-for-height z-score (WHZ)-and bilateral pitting nutritional oedema that often indicate a child has suffered recent and severe weight loss due to lacking sufficient food or having a disease that causes weight loss [2].In 2022, approximately 148.1 million children under five were estimated to be suffering from stunting-defined as low height-for-age z-score (HAZ) and indicating chronic or recurring undernutrition [3]-and 45 million children were estimated to be suffering from wasting [3].These conditions are often linked to a child experiencing adverse socioeconomic conditions, poor maternal health and nutrition, persistent illness, and/or improper infant and young child feeding (IYCF) and care.
The burden of undernutrition in children under five is disproportionately borne by conflict-affected, fragile, and low and middle-income countries (LMICs) [4,5].Conflict and violence cripple healthcare, economic, agricultural, labour, and supply chain systems, which in turn exacerbate food insecurity, hunger, and malnutrition [5].The destruction and deterioration of housing and water, sanitation, and hygiene (WASH) infrastructure that are often associated with conflict and violence may cause mass displacement and overcrowding, leading to increased spread and risk of infectious diseases that in turn increase the risk of undernutrition [6].Evidence indicates that conflict is the single greatest driver of hunger, with 60% of individuals suffering from hunger residing in areas experiencing war and violent conflict [7].Children living in a conflict zone are more than twice as likely to suffer from malnourishment and four out of every five children whose growth has been stunted due to malnutrition today live in countries affected by conflict [8].
While food insecurity is a major cause of malnutrition, particularly in LMICs, food assistance has been reported as a short-term, unsustainable solution [9].Instead, programmes that are integrated, intersectoral, sustainable, and community-based have been encouraged [10,11].Systematic reviews that investigated the efficacy of interventions in child malnutrition found that effective, evidence-based approaches to child stunting in LMICs were characterised by political commitment, multi-sectoral collaboration, community engagement, and communitybased service delivery platforms [12].
Community health worker models have been employed in conflict-affected and fragile settings (CAFS) as an effective tool to expand and improve healthcare coverage, identify and treat illnesses, and ultimately, save lives and enable people to thrive [13].The term community health worker (CHW) has been defined and described using a variety of methods, but generally refers to community-based health workers, often lay people from a community, that work within their community to monitor community health, assess community health needs, deliver health services, promote healthy behaviours, and receive training on different elements of community health and mobilisation [14].CHW models are particularly useful in resourceconstrained settings, extending weaker health systems so they can reach 'last-mile' communities and progress 1 Acute malnutrition is subdivided into two types: severe acute malnutrition (SAM) and moderate acute malnutrition (MAM).MAM, also known as moderate wasting, is defined by a WHZ ≥ -3 and < -2 standard deviations (SD) of the WHO child growth standards median, a MUAC ≥ 115 mm and < 125 mm, or both.SAM, also known as severe wasting, is defined by a WFH z-score < -3 SD of the WHO child growth standards median, a MUAC < 115 mm, the presence of a bilateral pitting oedema, or a combination of the three.Within the term SAM, the terms marasmus (referring to children very thin for their height) and kwashiorkor (referring to the presence of oedema in malnourished children) have been absorbed.Global acute malnutrition (GAM) refers to the combination of MAM and SAM and is used to measure the nutritional status of a population and as a severity indicator in emergency settings [3]. 2 Stunting is defined as chronic malnutrition, with a HAZ < −2 [3]. 3 Underweight is defined by a low weight-for-age z-score (WAZ) < −2 SD of the WHO child growth standards median.An underweight child may be stunted, wasted, or both [2].
towards universal health coverage and health equity.These communities are difficult to access due to a multitude of reasons, including geographical isolation, distrust, poverty, marginalisation, and conflict, and disproportionality suffer from poorer health outcomes, including in child health and nutrition [13].
In places where access to health facilities is limited and disrupted-as is the case in low-resource and conflict-affected and fragile settings-CHWs and community-based interventions can be critical in detecting, referring, and managing cases of child malnutrition [15,16].The use of CHWs and community-based models in the identification, prevention, and management of child malnutrition in emergency settings has been advocated for by humanitarian actors and academic researchers alike [13,15].The United Nations, with a joint statement released by the World Health Organization (WHO), World Food Programme (WFP), and United Nations International Children's Emergency Fund (UNICEF), emphasised the need to prioritise and integrate community-based management of SAM in emergency settings [17,18].
There exist systematic reviews on delivering nutrition interventions to women and children in conflict settings [19] and the effectiveness of interventions in managing acute malnutrition in children under five in LMICs [20].There is also a rapid review synthesis of lessons learned from community-based management of acute malnutrition programmes in CAFS [21].However, to our knowledge, there lacks systematic evidence that comprehensively examines the impact of CHW models in preventing, identifying, and managing child undernutrition in CAFS.We conducted this review to explore the effectiveness, cost-effectiveness, ability, acceptability, and feasibility of CHW models in identifying, managing, and preventing undernutrition in children under the age of five in CAFS.These findings may inform policymaking, programme implementation, and programme design to strengthen best practices for CHW models addressing child undernutrition in CAFS.

Study design
The review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting standards [22].The search strategy was developed using the Population-Intervention-Comparisons-Outcomes-Setting (PICOS) framework [23] as a guide.We performed our search using Ovid online database search platform, specifically searching the databases of Ovid MEDLINE(R), COCHRANE, Embase Classic, Embase, Econlit, Global Health, SCOPUS, and Social Policy and Practice.Additional relevant articles were identified through reference-checking articles included from full text review stage onwards.Search terms were related to children being identified with or treated for malnutrition or undernutrition by CHWs in CAFS.The specific search terms and strategy used underwent multiple reviews, test searches, screenings, and iterations to ensure that the final search results retrieved as many relevant citations as possible.See Table 1 for final search terms and search strategy.The review was managed and conducted using Covidence review software [24].No ethical approval was required for this systematic review; paperwork was submitted to the London School of Hygiene and Tropical Medicine MSc Research Ethics Committee to document the study and confirm that no ethical approval was needed.

Eligibility criteria
We aimed to gain a comprehensive understanding of effectiveness, cost-effectiveness, ability, acceptability, and feasibility of CHW models in the prevention, identification, and treatment of child undernutrition in CAFS.Therefore, we used a mixed-methods systematic literature review approach to identify and include quantitative, qualitative, and mixed-methods study designs.Peer-reviewed studies published between 2006 and October 2022 are included in our review.Grey literature, studies that were not peer-reviewed, and studies not published in English were excluded.Our rationale for the selected time frame of studies is twofold.First, we reasoned that the time frame needed to allow for focus on modern conflict-affected and fragile settings-when there has been a rise in active violent conflict that is more severe and protracted-and enable practical application of the systematic review to contemporary CAFS.Second, we required a standardisation method for conflictaffected and fragile settings.And, the selected time frame is when the World Bank Group's (WBG) annual Fragile and Conflict-affected Situations (FCS) lists-which commenced publication in 2006 to classify study settings as non-CAFS or CAFS-became available.
The study population inclusion criteria were children under the age of five years old with or at risk of any type of undernutrition including, severe acute malnutrition (SAM), moderate acute malnutrition (MAM), global acute malnutrition (GAM), stunting, micronutrient deficiencies, and underweight.The intervention criteria were CHW models that were used to prevent, identify, or manage undernutrition, SAM, MAM, GAM, stunting, micronutrient deficiencies, and/or underweight in children under the age of five in CAFS.The outcome criteria included clinical outcomes (cure rate, defaulter rate, death rate, recovery rate, relapse rate), anthropometric outcomes (WAZ, HAZ, LAZ, WHZ, MUAC, bilateral pitting nutritional oedema), dietary diversity, minimum adequate diet, treatment coverage, IYCF indicators (including breastfeeding, exclusive breastfeeding, complementary feeding, and supplementary feeding), behaviour change, knowledge absorption, acceptability, feasibility, quality of care, costs, and cost-effectiveness.The settings inclusion criteria was community-based and CAFS-with CAFS defined using the WBG's annual FCS list.The WBG classified fragile countries as such if they exhibited high levels of institutional and social fragility, which was determined through indicators measuring the quality of policy and institutions, and manifestations of fragility.Conflictaffected countries were determined by a threshold number of conflict-related deaths relative to the country's population [25][26][27][28][29].When applying the WBG FCS list to the inclusion and exclusion criteria for this systematic review, a study that had occurred in a country within a year of that country being classified as a fragile or conflictaffected situation was included in the literature as a CAFS.A study that had occurred in a setting that had been previously classified as a CAFS was included as a CAFS.See Table 2 for a full overview of the inclusion and exclusion criteria for this study.

Data extraction and analysis
All retrieved papers were de-duplicated on the Ovid online search platform; unique records were imported into the Covidence platform.Any title lacking an abstract was hand-searched by title and author on Google, downloaded, and then imported into Covidence.To minimise As this was a systematic review of qualitative, quantitative, and mixed-methods study designs, methods for extraction and synthesis aimed to capture quantitative and qualitative data points.Information were extracted on country, study population, sample size and determination (if applicable), methods, intervention, comparator (if applicable), outcomes, findings, and facilitators or barriers to CHW effectiveness.A meta-analysis was not deemed to be appropriate for this study as the quantitative studies were few and not sufficiently homogenous.Instead, a narrative synthesis of the data was undertaken.
To appropriately assess the quality of each study according to their specific methodological design, the Critical Appraisal Skills Programme (CASP) tools were chosen.CASP checklists were used to assess the quality of qualitative studies, cohort studies, systematic reviews, randomised control trials, and economic evaluations [30].One of the co-authors reviewed, assessed, and graded each study that had been included in the data extraction step, coding each study as high, moderate, low, or very low quality.No studies were excluded based on quality grade.

Limitations
To the author's knowledge, this systematic review is the first to focus specifically on CHW models being used to prevent, identify, or manage child undernutrition in CAFS.However, there are limitations to this systematic review and results shall be understood within the context of these limitations.The search results suggest limitations in the search strategy.Despite multiple test searches, reviews, and consultation with an academic librarian, the initial search string yielded a high volume of irrelevant studies, and the modified search string and syntax yielded a small number of better matched studies.We note that this limitation may be due to the exclusion of terminology, including stunting, underweight, and various terms related to micronutrient deficiencies, and hope the information provided here may inform future studies.We initially set the search time frame to be between 1992 and 2022; but the classification for CAFS were unsystematic and convoluted prior to 2006, when the WBG issued the FCS list.As such, we adjusted our time frame to capture studies between 2006 and 2022, ensuring that our results are not only comparable, but also reflect evidence on modern CAFS that policymakers and organisations can apply in current and emerging CAFS.This particular limitation may be due to the lack of a commonly used, standardised definition of CAFS.We used the WBG annual FCS lists that included only countries and the territory of the West Bank and Gaza to classify study settings.This type of country-level classification excludes countries and regions that border and are impacted by conflict; it also excludes conflicts such as those in the Kashmir region.Nevertheless, this is the most comprehensive list that we could identify and use to systematically classify different settings.
The diversity of study designs, outcomes investigated, and interventions researched led to significant heterogeneity, limiting the synthesis of results.We hope the inclusion of a variety of angles to holistically and effectively address child malnutrition in CAFs provided a range of modalities that may be useful in informing research, programmes, and policies.To ensure we included only results generated from robust methodology, we restricted the search to include only peer-reviewed publications.This decision inevitably excluded potentially relevant studies, particularly grey literature, conducted by humanitarian health and nutrition organisations.Programme reports from these organisations can be of particular importance as there is a dearth of published research in CAFS due to the challenges these settings present to research.Systematic reviews may also be prone to publication bias with interventions yielding null results being ignored; we caution against viewing the captured interventions as the only existing interventions.Instead, the included interventions shall be viewed as interventions with some evidence to suggest their ability, effectiveness, cost-effectiveness, acceptability, and feasibility.

Search results
The PRISMA diagram provided in Fig. 1 describes the search, screening, and review process for this study.A total of 3252 results were returned from the online academic database search platform Ovid.Upon import of these results into Covidence and deduplication, a total of 3044 citations remained for screening.An initial title and abstract screening conducted independently by two reviewers, resulted in 2975 citations being excluded.For the remaining 68 studies, a full text review was then undertaken, resulting in 38 further studies being excluded and 30 studies remaining for extraction.Of those excluded at the full text review stage, 16 studies were excluded due to being the wrong intervention, with many not being CHW interventions.Ten studies were excluded due to being the wrong study design, with many being grey literature.Eight of the studies were systematic, literature, or scoping reviews and, following the reference checking of these publications and identification of individual publications within each of them to include in this systematic review, they were excluded.An additional 24 citations were identified for full text review through reference-checking articles included in the full text review stage.Of those, 17 were ultimately included.Prior to extraction, the included publications were reviewed a second time to ensure they aligned with the updated inclusion criteria and WBG FCS lists.From this, an additional 22 studies were excluded, 10 of which were excluded because they were determined to not be in a CAFS.This process resulted in a final number of 25 studies meeting the inclusion criteria and being included in the review.

Characteristics of included literature
We described the details of the included studies in this section.See Appendix

Study setting characteristics
The study settings of the 25 included studies spanned three regions and ten countries.The majority (84%) of publications included in the review researched interventions in the region of sub-Saharan Africa (SSA) and totalled 21 publications (n = 21).The next most common research setting region was Middle East North Africa (MENA), representing 12% of included studies and 3 publications (n = 3).Lastly, one publication's research took place in Haiti in the Latin America region, representing 4% of included studies.Within sub-Saharan Africa, the subregion of West Africa had eight studies, the subregion of Central Africa had four studies, the subregion of East Africa had six studies, and the subregion of Southern Africa had three studies.

Study design characteristics
The 25 studies eligible for inclusion were of quantitative, qualitative, or mixed-methods study designs.Five of the studies were randomised control trials (RCTs) investigating the effectiveness of various community-based malnutrition interventions involving a CHW model.Stewart et al. [31] conducted a multi-arm, cluster-RCT investigating the effectiveness of lipid-based nutrient (LNS) supplementation distributed by community nutrition workers (CNWs) on child anaemia and micronutrient status in Madagascar.Kurdi et al. [32] used data from a cluster-RCT on the Yemen Cash for Nutrition programme, which provided impoverished women that were pregnant or had children under two years old with monthly cash transfers conditional on their showing up to monthly community health volunteer (CHV)-led nutrition sensitisations.They used data on secondary outcomes of breastfeeding and water treatment to ascertain the impact of CHV-led nutritional training and counselling on IYCF practices and knowledge amongst participating women.Ayalew et al. [33] conducted an RCT studying the effect of complementary feeding behaviour change communication (BCC) through community-level actors in Ethiopia.A cluster-RCT by Maust et al. [34] measured the effectiveness of an integrated nutrition protocol-which included community-based, mother peer-counselling care groups with nutrition messaging delivered both onsite and at home visits-in managing SAM and MAM in Sierra Leone.A cluster-RCT by Lelijveld et al. [35] looked at the effectiveness of treating high-risk MAM with clinic-delivered therapeutic food and community elder delivered nutrition counselling versus just community elder delivered nutrition counselling in Sierra Leone.
Seven publications were cross-sectional comparison studies.Addo et al. [36] studied the combined effects of community-based nutrition education for mothers and pregnant women, community-based outreach counselling by CHWs on IYCF and small-quantity lipid-based nutrition supplementation (SQ-LNS), and the provision of SQ-LNS by CHWs on anaemia and growth in children in the Democratic Republic of the Congo (DRC).Locks et al. [37] studied the same programme, but instead looked at the impact of the integrated IYCF and SQ-LNS delivered by CHWs on outcomes of IYCF behaviour change and child dietary diversity in the DRC.Bisimwa et al. [38] researched the effectiveness of nutritional monitoring of children by community volunteers (CVs) in the DRC.Kim et al. [39] investigated the impact of community-based behaviour change interventions led by different cadres of CHWs on feeding practices and child stunting in Ethiopia.Mayhew et al. [40] evaluated the impact of a community-based growth monitoring and promotion (cGMP) programme using community health volunteers (CHVs) in Afghanistan on improving nutrition in underweight children.Worku et al. [41] examined the impact of child dietary diversity by the Sustainable Undernutrition Reduction programme in Ethiopia (SURE), a BCC project that used CHW cadres to enhance community-based nutrition (CBN) to address complementary feeding, improve household dietary diversity through IYCF, and sensitise on nutrition-specific agriculture.Getachew et al. [42] measured the accuracy of health extension workers (HEWs) in diagnosing child illnesses, including acute malnutrition.
Five of the study designs included in this systematic review were cohort studies.A retrospective, dual cohort study by Rajabi et al. [43] examined the relative effectiveness of supplementary feeding plus counselling by community respected elders compared with solely nutrition counselling by community respected elders in reducing risk of SAM and death in children with moderate wasting in Sierra Leone.A cohort study and impact evaluation were conducted by Balaluka et al. [44] in the DRC to determine the impact of CHWs promoting optimal breastfeeding practices through doorto-door visits and community meetings and organising monthly community child weighing sessions on exclusive breastfeeding (EBF) practices.An observational, clinical prospective multi-centre cohort study by Alvarez Morán et al. [45] researched the quality of care for uncomplicated SAM by CHWs in Mali.Conducted by the same primary author and in the same country was a multi-centre, randomised intervention study looking at the effectiveness of CHW models in identifying and treating SAM compared to a facility-based model [46].Also in Mali, was a prospective, nonrandomised community intervention study by Charle-Cuéllar et al. [47] examining the effectiveness of different levels of supervision for CHWs when treating SAM.
Three of the publications included in the systematic review researched the acceptability and feasibility of interventions.Paul et al. [48] studied the impact of community-based complementary feeding messages on child feeding practices and used the Trials of Improved Practices (TIPs) methodology to understand the acceptability and feasibility of using local child nutrient supplements in Zimbabwe.Desai et al. [49] investigated the independent and combined effects of WASH practices and IYCF on child stunting and anaemia, the understanding of WASH and IYCF messages and tools piloted, and-using the TIPs methodology-the feasibility of delivering the intervention by village health workers (VHWs) in Zimbabwe.A study by Van Boetzelaer et al. [50] looked at the performance, feasibility, and community acceptability of community-based distributor (CBDs) with low literacy in treating SAM in South Sudan.
There were four studies that described using mixedmethods in their research, including the previously described Desai et al. [49] study, which conducted an effectiveness evaluation, TIPs acceptability and feasibility evaluation, and knowledge assessment of the VHWdelivered WASH and IYCF education interventions.Van Boetzelaer et al. [50] also used mixed methods, studying the performance of CBDs, child treatment outcomes, and community acceptability.Ayoya et al. [51] analysed both qualitative and quantitative data on community-based child nutrition interventions in Haiti.Renzaho et al. [52] used mixed methodologies to assess the impact and scalability of integrated, community-based programmes for the management of severe wasting in South Sudan.
Economic evaluations were conducted by two of the citations included in the systematic review, with both studies' research setting being Mali.Rogers et al. [53] evaluated the costs and cost-effectiveness of the treatment of uncomplicated SAM by CHWs in comparison to an outpatient facility in Mali.Isanaka et al. [54] conducted an incremental, cost-effectiveness analysis of community-based screening and treatment of SAM set within a cluster-RCT.
Lastly, one case study in Yemen by Tappis et al. [55] investigated the delivery of reproductive, maternal, newborn, child, and adolescent health and nutrition (RMN-CAH + N) services.

Ability of CHW models in treating child SAM in CAFS
Three studies assessed the ability and performance of CHWs in the treatment of SAM in CAFS.A cohort study by Alvarez Moran et al. [45] in Mali examined CHWs' performance and ability to correctly treat uncomplicated SAM in children under five years old, with uncomplicated SAM defined by national protocol as MUAC < 11.5 cm, WHZ < − 3, and/or nutritional oedema.The study reported that in 75% of cases CHWs correctly administered medical treatment for SAM, and in 100% of cases CHWs correctly administered RUTF for management of acute malnutrition or risk of acute malnutrition.Van Boetzelaer et al. [50] found that through conducting performance checklists, CBDs with low literacy in South Sudan were highly accurate in their adherence to a simplified SAM treatment protocol for uncomplicated SAM.Accuracy was determined through case management observations of CBDs by supervisory staff who then recorded a performance score.The performance score was made up of specific tasks for SAM treatment, including welcoming the caregiver, assessing for danger signs including bilateral pitting oedema, taking the child's MUAC measurement, conducting an appetite test, determining the weekly RUTF dosage based on the child's weight, administering medication (Amoxicillin for week 1, Albendazole for week 2), filling out the patient register, counselling the caregiver, child progress monitoring, and child discharge.During the study, 141 performance assessments were conducted, resulting in a mean performance score for CBDs of 89.9% (95% CI 86.4-96.0%)[50].Furthermore, an investigation by Charle-Cuéllar et al. [47] into the effectiveness and performance of CHWs in treating uncomplicated SAM, using different models of supervision: high supervision, light supervision, and no supervision, found that the proportion of children cured was 81.4% in the high supervision group, 86.2% in the light supervision group, and 66.9% in the no supervision (control) group.

Acceptability and feasibility of CHW models in preventing, identifying, and treating child undernutrition in CAFS
Four studies examined the acceptability and feasibility of CHW models in preventing, identifying, and managing child malnutrition in CAFS.A qualitative acceptability and feasibility analysis by Paul et al. [48] found that the VHW-delivered nutrition intervention involving BCC for complementary feeding methods, use of locally available complementary foods, and use of LNS, was widely accepted by the intervention community, with mothers eagerly absorbing and spreading the new knowledge they had received from the VHWs.The results of the study suggest that community-based approaches to creating and disseminating context-specific feeding messages were feasible and effective for the study setting [48].Desai et al. [49] came to a similar conclusion for their study on the feasibility of VHWs in effectively delivering EBF, IYCF, and WASH promotion: that it is feasible for VHWs to deliver context-specific IYCF messages.Evidence also suggested that using low-literate CBDs to treat uncomplicated SAM in South Sudan was acceptable and feasible and that a simplified treatment protocol and tools adapted for low literacy were key [50].A cGMP programme in Afghanistan delivered by CHWs was accepted by communities due to comprehensive community consultation; intentional, culturally acceptable, and accessible programme design; and high levels of participation amongst targeted children early in the programme [40].

Effectiveness of CHWs in identifying child undernutrition
Four publications evaluated the effectiveness of CHW models in identifying child undernutrition in CAFS.A cross-sectional cohort study evaluated CVs' effectiveness in monitoring the growth of children under five and identifying those highly susceptible to acute malnutrition in South Kivu, DRC-a context experiencing endemic malnutrition and armed conflict [38].The monthly community weighing sessions conducted by CVs resulted in high rates of coverage (calculated as percentage of children weighed per village), even during periods of conflict, with the median percentage of children weighed per village ranging between 80 and 90% for those aged 12-59 months old and between 80 and 100% for those less than 12 months old.The median percentage of children aged 12-59 months old per village ranked as highly susceptible to malnutrition by the CVs decreased from 4.2% in 2004 to 2.8% in 2005.Another cross-sectional study in Ethiopia evaluated the diagnostic accuracy of HEWs when diagnosing common childhood illnesses, including malnutrition, and found that the ability of HEWs to correctly diagnose malnutrition was particularly poor, with only four out of ten children with malnutrition being correctly identified [42].HEWs diagnosed MAM, uncomplicated SAM, and complicated SAM according to the WHO growth standards, using WFH/L, oedema, MUAC, presence of medical complications, the ability to finish an RUTF for children older than six months and the existence of a breastfeeding issue for children six months or younger.Overall, diagnostic sensitivity and specificity for child acute malnutrition by HEWs was 39% and 99%, respectively [42].A cohort study in Mali looked at CHWs' performance and ability to correctly screen for child uncomplicated SAM (defined by national protocol as MUAC < 11.5 cm, WHZ < − 3, and/or nutritional oedema) and that, when screening, CHWs correctly assessed 97.6% of children for existence of cough, diarrhoea, fever, and vomiting; 95.2% of children for presence of major danger signs; 96.8% of children's MUAC; 78.4% of children's oedema; and 100% of children's height [45].Additionally, 77.8% of children who required an appetite test were tested correctly.A mixed methods study on the precis of nutrition in children in Haiti noted that CHWs were a key link to the health system, playing a central role in screening and follow up for malnutrition and services delivered outside the community [51].

Effectiveness of child undernutrition behaviour change communication (BCC) interventions involving CHWs
Nine of the included studies investigated the effectiveness of community-based, BCC interventions in addressing child undernutrition and improving child growth and nutrition outcomes.Addo et al. [36] evaluated the impact of an enhanced IYCF programme-which included improved IYCF training for CHWs and outreach counselling conducted by CHWs-on anaemia and the anthropometric status of children in Katanga, DRC.The study results revealed that the intervention was associated with a 11.0% (95% CI − 18.1, − 3.8; P < 0.01) adjusted relative reduction in anaemia prevalence.However, there was no intervention effect on anthropometry, iron deficiency, or vitamin A deficiency.A cross-sectional comparison study by Mayhew et al. [40] researched how a cGMP programme carried out by illiterate CHWs impacted the nutritional status of Afghan children and found that children that had participated in the programme had a statistically significant higher mean WFA z-score of − 0.9 (95% CI − 1.0, − 0.8), compared to − 1.2 (95% CI − .3,− 1.1) for those that had not participated.Kim et al. [39] investigated the effectiveness of an intensive, community-based BCC intervention delivered by CHWs and community leaders on complementary feeding practices and child stunting in Ethiopia and found that while minimum dietary diversity and acceptable diet did increase significantly in the intensive group, they remained low at the endline (24.9% and 18.2%, respectively).Additionally, in children 6-23.9 months old, the study observed significant differential declines in stunting prevalence, with a decrease from 36.35 to 22.8% in the intensive group.Another BCC programme leveraged HEWs to improve dietary diversity amongst children aged 6-23 months; Worku et al. [41] found that children living in districts covered by the programme were 2.5 times more likely to have sufficient dietary diversity than those living in uncovered districts.A cluster-RCT by Ayalew et al. [33] investigated the effect of a community-based, complementary feeding BCC delivered by CHWs on infant growth and morbidity in rural communities in Ethiopia and found that, compared to the control group, infants in the intervention group had significantly higher weight gain (Mean Difference (MD) 0.46 kg; 95% CI 0.36-0.56)and length gain (MD 0.96 cm; 95% CI 0.56-1.36)and reduced rates of infant stunting (by 7.5% points; Risk Ratio (RR) = 0.68; 95% CI 0.47-0.98)and underweight (by 8.2% points; RR = 0.55; 95% CI 0.35-0.87).
A mixed methods pilot study by Desai et al. in Zimbabwe researched the effectiveness of VHWs' promotion of IYCF messaging on mothers' child feeding practices and subsequent changes in child dietary diversity.It reported that maternal knowledge about infant feeding increased after receiving the VHW modules on IYCF [49].Another Zimbabwe-based study by Paul et al. [48] on VHW-disseminated feeding messages found that there was a significant increase in intakes of energy, protein, vitamin A, calcium, zinc, iron, and folate from complementary foods after counselling (p < 0.05).
An RCT of the Yemen Cash for Nutrition programme-a programme that provided cash incentives to impoverished women that were pregnant or had children under two years old to attend monthly nutrition sessions by community health volunteers (CHVs)-found that programme exposure increased the probability of breastfeeding initiation within the first hour after delivery by 15.6% points (p < 0.05), the probability of exclusive breastfeeding (EBF) during the first six months by 14.4% points, the probability of households treating water consumed by adults by 16.7% points (p < 0.01), and the probability of treating water consumed by children under the age of two by 10.3% points (p < 0.10) [32].
The effectiveness of CHWs in promoting EBF was also explored in an area of the DRC with endemic malnutrition [44].The study found that the median duration of EBF from birth was six months in the intervention group compared to four months in the comparison (the comparison group had no CHWs or community-based programme) group (p < 0.001).The intervention area also had a higher proportion of infants receiving EBF at six months old (57.7%), versus 2.7% in the comparison area (p < 0.001) [44].

Effectiveness of CHW models in relation to supplementary feeding
Five of the included studies investigated the effectiveness of interventions combining CHW nutrition sensitisation and supplementary feeding in addressing child undernutrition and improving child growth and nutrition outcomes.A study by Rajabi et al. in Sierra Leone investigated the effectiveness of facility-based supplementary feeding plus nutrition counselling by community respected elders compared with solely nutrition counselling by community respected elders in reducing the risk of SAM and death in children with moderate wasting.It found that children in the supplementary feeding plus nutrition counselling cohort had a lower risk of developing SAM or dying over the 24-week period of follow up than children in the solely nutrition counselling cohort.The supplementary feeding cohort also had greater rates of weight gain and greater rates of MUAC gain when compared to the solely counselling cohort [43].A cluster-RCT investigated the effectiveness of two treatments for high-risk MAM, with high-risk MAM defined as a child having one or more of the following criteria: MUAC < 11.9 cm, and/or WAZ < − 3.5, and/or the mother not being the primary caregiver, and/or child being less than two years old [35].The intervention consisted of ready-to-use therapeutic food (RUTF) provision, plus antibiotics, plus nutrition counselling; the control consisted of solely nutrition counselling [35].The nutrition counselling for both arms was led by trained community elders who sensitised on IYCF and gave cooking demonstrations.The study results reported that children in the intervention arm had greater short-term recovery (48% compared to 39% at week 12; risk difference (RD): 0.08; 95% CI 0.03, 0.13), weight gain, and MUAC gain compared to the control arm.They also had lower risk of developing SAM (18% compared to 24%; RD − 0.07; 95% CI − 0.11, − 0.04) and dying (1.8% compared to 3.1%; RD − 0.02; 95% CI − 0.03, − 0.00) compared to the control arm.However, it was found that by week 24 the risk of SAM was similar in both groups [35].
The effectiveness of community-based supplementary feeding in conjunction with intensive community-based nutrition counselling, compared to just community-based nutrition counselling was further explored in a five-armed, cluster-RCT by Stewart et al. [31] in Madagascar.The control arm (T0), consisted of status quo treatment based on standard Madagascan growth monitoring and nutrition education protocol, this included monthly growth monitoring and key messages on maternal nutrition, early initiation of breastfeeding, EBF for the first six months, continued breastfeeding through two years old, and age-appropriate complementary feeding and hygiene behaviours.Community nutrition worker (CNWs) demonstrated cooking with local ingredients that were complementary foods.The government also distributed vitamin A biannually for children under 5 years old and pregnant women were given iron-folic acid supplements during ANC visits.There were four intervention arms: T1 consisted of T0 plus home visits for intensive nutrition counselling (defined as intensive due to additional CNWs being deployed for home visits); T2 consisted of T1 plus LNS for children aged 6-18 months that was distributed by CNWs; T3 consisted of T2 plus LNS for pregnant and lactating women; T4 consisted of T1 plus early childhood stimulation and parenting messages distributed by CNWs.It discovered that children in the study arms providing LNS had an approximately 40% lower prevalence of anaemia, as well as an approximately 25% lower prevalence of iron deficiency than children in the control study arm [31].Locks et al. analysed the impact of an enhanced IYCF programme compared to the standard IYCF programme on breastfeeding, handwashing, and micronutrient deficiencies measured through child dietary diversity in DRC.The intervention was an enhanced IYCF programme consisting of both community-and facility-based counselling for mothers on handwashing, SQ-LNS, and IYCF practices; monthly SQ-LNS distributions for children 6-12 months; and additional investments in the CHW platform including bike provision, enhanced training, updated CHW guidebook with images on IYCF, logbooks, improved/standardised supervision of CHWs, and standardised roles, responsibilities, and expectations for CHWs.The control was the national IYCF programme, with only facility-based counselling and no additional investments in the CHW platform [37].The study found that the intervention group was associated with improvements in breastfeeding and handwashing behaviours, but not with significant change in dietary diversity [37].In both groups, the minimum dietary diversity and minimum acceptable diet remained below 10% [37].A study by Maust et al. exploring the recovery and coverage rates of an integrated protocol using RUTF for AM (intervention) versus standard management for AM (control).The standard management included mother peer-counselling care groups with nutrition messaging delivered on-site only, no home visits, treatment of MAM with fortified blended flour, treatment of SAM with RUTF, and WFH as the admission and discharge tool to the treatment programme.The integrated management of AM included mother peer-counselling care groups with nutrition messaging delivered both on-site and at home visits, treatment of MAM with fortified blended flour, treatment of SAM with RUTF, and MUAC used for admission and discharge, with a MUAC < 12.5 cm defining malnutrition.The study reported GAM recovery of 83% in the integrated protocol and 79% in the standard protocol.It also generated preliminary results that care group participation was associated with a greater recovery rate [34].Coverage, calculated from number of children who received treatment over number of children eligible for treatment, was 71% for the intervention group, versus 55% in the control group (P = 0.0005) [34].

Effectiveness of CHW models versus facility-based models
When comparing the effectiveness of a CHW model to the standard facility-based model in improving the coverage and treatment of SAM, a study by Alvarez Moran et al. [46], 2018 found that the CHW-enhanced intervention group had a cure ratio of 94.2%, while the health facility-based control group had an 88.6% cure ratio, giving a risk ratio of 1.07 (95% CI 1.01, 1.12).Mortality ratios between the two arms were very similar and not statistically significant [46].Defaulter ratios in the control group were double the intervention group, while the coverage rate in the intervention group (86.7%) was more than double the coverage rate in the control group (41.6%) (p < 0.0001) [46].

Costs and cost-effectiveness of CHW models in identifying and managing child undernutrition in CAFS
Two studies assessed the costs and cost-effectiveness of CHW models in identifying and managing child malnutrition in CAFS.Isanaka et al. included communitybased screening in their economic evaluation of a study on the health outcomes, costs, and cost-effectiveness of a 'do nothing' strategy, 'Treat SAM only' strategy, and four dietary supplement strategies for MAM for screening and managing AM.Community-based screening costs were found to be only 4.7% of total costs for the 'Treat SAM only' arm and 1.7% to 1.9% of total costs for the MAM treatment arms.Costs were tabulated from the healthcare provider perspective, of which communitybased screening was one component.The breakdown of costs for bi-monthly community-based screening for MAM treatment and SAM treatment were as follows.For MAM treatment, community-based screening costed 1.89 USD per MAM child identified, including 1.46 USD for personnel (77% of total activity), 0.26 USD for infrastructure and logistical support (14% of total activity), 0.17 for management and administration (9% of totally activity).For SAM treatment only, communitybased screening costed 14.51 USD per SAM child identified, including 11.19 USD for personnel, 2.00 USD for infrastructure and logistical support, and 1.32 USD for management and administration [54].Another study by Rogers et al. [53] estimated the costs and cost-effectiveness of CHW-delivered treatment for SAM in comparison with outpatient facility-based treatment for SAM.In the intervention arm, CHWs conducted nutrition sensitisations, screened for SAM, referred complicated cases, and treated uncomplicated cases in communities.In the control arm, CHWs conducted nutrition sensitisations, screened for SAM, and referred all cases to the outpatient facility.The study findings were that both provider and beneficiary costs were higher in the intervention arm than in the control arm.On the provider side, this was due to the greater role CHWs played in service delivery.On the beneficiary side, this was due to greater enrolment.However, at the individual household level, individual households in the intervention arm spent considerably less time (2.15 h in comparison to 3.92 h) and money (0.60 USD in comparison to 1.70 USD) than in the control arm.In terms of cost-effectiveness, CHWdelivered care was found to be more cost-effective than outpatient, facility-based care for treating uncomplicated SAM.The average cost per child treated by CHWs was 244 USD versus 442 USD in the outpatient facility.The average cost per child recovered was 259 USD by CHWs and 501 USD in the outpatient facility.

Key barriers to CHW models in identifying or managing child undernutrition
Four studies reported barriers such as the interruption or disruption of programmes due to active conflict, states of emergency, militancy, or political unrest [38,39,52,55].Bisimwa et al. [38] reported an interruption in health facility services due to active conflict in Lwiro District, DRC, which corresponded to a 10% decrease in median percentage of children weighed monthly at community weighing sessions.However, they also noted that mean percentage of children weighed monthly never went below 80% and attributed the high mean percentage to the CVs' and village nutrition committees' ability to conduct the community-based nutrition intervention (including a public awareness campaign on malnutrition and growth monitoring, and arranging monthly weighing sessions) without the health facility [38].In Ethiopia, a ten month state of emergency interrupted the implementation of an intensive IYCF BCC intervention using nutrition sensitisation, community mobilisation, and mass media campaign activities delivered by HEWs, HDTLs, AEWs, religious leaders, and CBOs [39].Renzaho et al. [52] indicated that community-based management of severe wasting (CMSW) programmes in South Sudan were weakened by active conflict.Tappis et al. [55] cited insecurity as a main barrier to the effective delivery of both facility-and community-based RMNCAH + N services, with CHVs only able to address emergency needs (including child malnutrition needs) when the conflict environment allowed for movement and outreach to communities.
Two articles indicated weak links between communitybased interventions and the existing public health system and overall weak health system capacity impeded referrals, health facility follow-ups, and the ability to provide care [52,55].Renzaho et al. observed a dearth of human resources for health in the form of well-trained CHWs and facility-based staff that could independently and effectively run CMSW programmes in South Sudan.The study also noted that the limited integration of CMSW programmes into existing local health infrastructure and national health budgets impeded government ownership of the CMSW programmes [52].Tappis et al. [55] described the weak public health system in Yemen, and specifically the Ministry of Public Health and Population's limited capacity to oversee and support health directorates, restricted the impact and activities of community-based RMNCAH + N service delivery.
The costs of care-seeking also hindered programme success, particularly when a programme was not homebased or proximate; namely, the time and transport costs were found to be an challenge [40,50,55].In studying the impact of cGMP programme on the nutrition status of underweight children of participant caretakers in Afghanistan, Mayhew et al. [40] found that in the control group (the non-participants' caretaker group), 52% of non-participants cited an absence of a nearby cGMP programme as the reason for not attending a cGMP programme.Ahead of Van Boetzelaer et al. [50] conducting their 2017 study in South Sudan on the performance of low-literate CBDs' ability to treat SAM, only 41% of SAM children in the intervention area were enrolled in one of the International Rescue Committee's static CMAM programmes, with caregivers citing the main barriers to accessing care as distance to health facilities, inaccessibility of health facilities during rainy season, and high opportunity costs.Tappis et al. [55] found that beneficiaries struggled to access facility-based RMN-CAH + N care in Yemen due to transportation fees, far distance to reach functioning facilities, and roadblocks and checkpoints.Each of these studies made clear that static malnutrition interventions had a higher potential for inaccessibility for beneficiaries, and made the case for the use of mobile, CHW models that meet beneficiaries where they are.
Lack of demand due to distrust, disbelief in the helpfulness of the programmes, and the politicisation of aid was reported to have caused challenges for CHWs when delivering services to communities, with Mayhew et al. finding that 36% of the non-participant group did not attend a cGMP programme because they believed the programme would not be helpful; Van Boetzelaer et al. reporting that that high food insecurity and demand for RUTF made community members suspicious when a child was deemed ineligible for the CBDs' SAM treatment; and Tappis et al. citing that politicisation of aid led to restrictions in aid flow and service delivery based on local and geopolitical power dynamics [40,50,55].In their study on the impacts of an enhanced IYCF programme-which included community-and facility-based counselling, SQ-LNS, and additional investments in the CHW platform-on IYCF practices and dietary diversity, Lock et al. [37] observed that lack of access to and purchasing power for healthy, nutrient-dense foods in communities not only hampered community members' ability to access the foods they were sensitised to feed their children, but may have led to CHWs underemphasising nutrient-dense foods in their counselling as they knew the foods were difficult to acquire.
Three publications described weather-related barriers-with the study on CVs' effectiveness in communitybased child growth monitoring by Bisimwa et al. [38] noting the rates of acute malnutrition varying seasonally, the study on CMSW programmes in South Sudan by Renzaho et al. [52] reporting that wasting prevalence fluctuated with agriculture seasonality (remaining above the 15% emergency threshold during lean season), and Van Boetzelaer et al. [50] citing the inaccessibility of static CMAM services and programmes during the rainy season as previously described in the paragraph on costs of care-seeking.Tappis et al. [55] reported that an ongoing cholera epidemic hindered the effectiveness of CHWs' work delivering RMNCAH + N services, with human resources, funding, technical expertise, and logistics being prioritized of cholera over other services.Mayhew et al. [40] reported that CHWs delivering the cGMP programme to underweight children faced challenges including low-literacy levels, lack of job aids, lack of security, and cultural norms around women (including female CHWs and female caretakers) needing to be accompanied by men in public.When considering the scalability of CHW models to address child malnutrition in CAFS, Renzaho et al. [52] found that CMSW programme scalability was hindered by a resource constraints, weak context-specificity frameworks for the CMSW programmes, and weak community mobilisation due to a lack of involvement of community members in CMSW programme leadership, while Alvarez Moran et al. [45] 2017 found that when working to scale up CHW treatment of uncomplicated SAM in Mali, balancing an increased workload was a barrier.

Key facilitators to CHW models in identifying or managing child undernutrition
Key facilitators to successfully implementing CHW models and community-based interventions for child undernutrition were described in 15 of the included publications.The factor cited most widely, by six of the included articles, was the CHWs' connection to the community, with studies reporting how it enabled influence and engagement with the community and families, social mobilisation, community trust, context-specificity, and community ownership and knowledge sharing [32,33,38,40,48,50].While investigating the effect of complementary feeding BCC disseminated through community level actors on infant morbidity and growth, Ayalew et al. [33] found that CHWs' connection to the community enabled them to influence change and feeding practices.In a study evaluating CVs' effectiveness in child growth monitoring and awareness raising for acute malnutrition, Bisimwa et al. [38], found that CVs effectively mobilised the community around the seriousness of malnutrition.Kurdi et al. studied the impact of the Yemen Cash for Nutrition program, which provided conditional cash transfers to impoverished women with young children to incentivise attendance at CHV-led nutrition sensitisations, on knowledge and practices related to breastfeeding, complementary feeding, and handwashing.They reported that the trainings being provided by women from the community (CHVs) allowed for a trusting relationship to form between the CHVs and participants, and for the programme to be run without strict oversight [32].Van Boetzelaer et al. [50], in studying the performance of lowliterate CBDs treating SAM in South Sudan, reported that, in general, caregivers expressed trust in CBDs and their ability to provide SAM treatment.A study by Paul et al. researching the impact of VHW-delivered complementary feeding messages and complementary feeding resources on infant diets, found that IYCF messages from CHWs that addressed context-specific barriers, context-specific local foods, and mothers directly were key to improving underlying diet and behaviour change.It also found that community ownership and knowledge sharing was spurred by VHW engagement [48].Mayhew et al. [40] studied the impact of the cGMP programme on the nutritional status of young, Afghan children, and found that the CHW-delivered programme was culturally acceptable in both design and implementation.
Five articles cited that the close proximity of programmes, in the form of family-friendly home visits and local community sessions, facilitated interventions through eliminating beneficiary time and transport costs and delivering services in a comfortable, familiar setting [38,40,41,44,53].Bisimwa et al. [38] noted that CVs living in the same area as participants and conducting weighing and nutrition sensitisations in the community supported effective monitoring of child growth and awareness raising on malnutrition.A study by Worku et al. [41] researched the impact of an HEW-delivered BCC nutrition education programme on child dietary diversity, and reported that household visits from HEWs were significantly associated with dietary diversity (AOR = 2.0; 95% CI 1.25, 3.21).Balaluka et al. [44] cited the proximity of CVs to breastfeeding mothers as a significant advantage in their study on CVs' ability to improve breastfeeding of children under six months old in DRC.A cost-effectiveness study by Rogers et al. [53] reported that CHWs treating uncomplicated SAM was cost-effective, and that there were lower beneficiary costs incurred in the CHW-delivered SAM arm due to lower transport costs and opportunity costs (in the form of travel time).Lastly, Mayhew et al. [40] reported that conducting the cGMP programme through home visits and in villages led to greater attendance by study participants, which was associated with improved nutrition of underweight children in the study.
Two studies investigating IYCF BCC education through local community actors in Ethiopia cited that having a programme element focused on women's participation through cooking demonstrations or food preparation was associated with behaviour change and improved child growth and dietary diversity [33,41].Ayalew et al. [33] reported that CHWs training women in local complementary food preparation contributed to infant weight and length gain, while Worku et al. [41] reported that children of women participating in CHW-led food preparation programmes had a dietary diversity AOR of 1.9 (95% CI 1.19, 2.96).
Supervision was cited as a key success factor in three of the included publications [37,45,50].Four studies reported that investment in high quality training and tools, including active-learning practices of songs, practical exercises, role playing, and pictorial tools for illiterate CHWs, was essential to the success of the CHW programming [37,40,45,50].Locks et al. [37], studying the impact of an enhanced IYCF programme (which consisted of community-and facility-based counselling on IYCF practices; monthly SQ-LNS distributions; and investments in the CHW platform through enhanced CHW training, updated CHW guidebook with IYCF images, and improved/standardised supervision) on IYCF behaviours and dietary diversity attributed training and appropriate supervision of CHWs to success in improving breastfeeding and handwashing behaviours.Assessing the quality of care for CHW treatment of uncomplicated SAM, Alvarez Moran et al. found that well-trained and supervised CHWs could effectively manage uncomplicated SAM.Van Boetzelaer et al. [50] reported that low-literate CBDs could adhere to a simplified treatment protocol for uncomplicated SAM, with success facilitated through frequent supervision and the use of low-literacy adapted tools and training including use of songs, practical exercises, and role playing.Mayhew et al. [40] cited that providing pictorial tools for illiterate CHWs contributed to their effectiveness in the cGMP programme.
Two studies reported the beneficial use of incentives, with Bisimwa et al. [38], noting local government agreeing to employ CVs when there was a paid activity and Locks et al. [37] noting the use of bikes as transportation support and incentive.Van Boetzelaer et al. [50] found that ensuring a manageable workload for CHWs through having SAM treatment provided on a fixed day per week was key to CHWs' ability to effectively manage uncomplicated SAM.When considering facilitators for the scalability of CHW programmes, Renzaho et al. [52], found that the following were essential: best partnership practices, standardisation and adoption of a national programme implementation strategy, a BCC component to programmes, and an overall collaborative, holistic, and multidisciplinary approach to child malnutrition.

Discussion
The confluence of the Covid-19 pandemic, climate change, and the global rise in violent, protracted conflict has led to catastrophic levels of hunger, food insecurity, and malnutrition in numerous regions and countries, such as occupied Palestinian territory, Sudan, Ukraine, and Yemen.[56].Of the hunger hotspots identified as the highest concern for October 2022 to January 2023 by WFP and the Food and Agriculture Organization (FAO), conflict has been a key driver or aggravating factor for all of them [56].Communities are being pushed onto the brink of or into famine, as is currently the case in Afghanistan, Ethiopia, the occupied territory of Gaza, Nigeria, Somalia, South Sudan, Sudan, and Yemen [56][57][58].
In these complex settings, the need for essential healthcare services is immense, but so are the challenges to delivering care to affected populations.As such, researching, designing, and implementing policy solutions that are effective, responsive, and adapted to the specific contexts and characteristics of child undernutrition in CAFS is critical.CHW policies and programmes have been deployed in both emergency and non-emergency settings to provide essential healthcare services for a variety of diseases, extend primary healthcare, and further universal health coverage.In recent years, the use of CHW models to address child malnutrition in CAFS has become more widely accepted, with the WHO endorsing the use of CHW-led treatment models for AM where possible in its revised guidelines on AM management [59].It is critical that the body of research in this area grows alongside its popularity in order to ensure strong, continued, and effective advocacy for and integration of CHW models in local, national, and global policy agendas for child undernutrition in CAFS.
This systematic review identified 25 publications that reported on the effectiveness, cost-effectiveness, ability, feasibility, and acceptability of CHW models in preventing, identifying, or managing child undernutrition in CAFS.Identified evidence suggests that CHWs were effective in screening for child acute malnutrition (both MAM and SAM) in CAFS.Notably, one study reported that CVs in DRC were able to continue child growth monitoring and acute malnutrition screening activities and achieve high screening coverage even when the health facilities were closed during times of active armed conflict [38].This finding suggests community involvement and CHWs play a key role in the building resilient child acute malnutrition screening programming in CAFS, which was echoed in a systematic review on nutrition interventions for women and children in CAFS [20].
In terms of diagnostic accuracy, one study on HEWs' diagnostic accuracy in Ethiopia, found a concerningly low sensitivity of 39% for child malnutrition.HEWs diagnosed MAM, uncomplicated SAM, and complicated SAM according to WHO growth standards, using WFH/L, oedema, MUAC, presence of medical complications, the ability to finish an RUTF for children older than six months, and the existence of breastfeeding issue for children younger for children aged six months or younger [42].Another study in Mali found CHWs classified SAM correctly in 100% of cases, measured height correctly in 100% of cases, correctly assessed oedema in 78.4% of cases, and correctly performed an appetite test in 77.8% of cases [45].The incongruent results could be due to differences in training methods, cultural context, or CHW attributes.
All of the included studies on BCC (n = 9) found that this type of intervention to be effective, to varying degrees and with varying population sizes, in improving outcomes such as child feeding practices, breastfeeding practices, EBF, child diet, child dietary diversity, and child anthropometric indicators.These findings are in line with the results of a recent systematic review on the effectiveness of nutrition social and behaviour change communication (NSBCC) interventions and the results of a recent publication on BCC's effectiveness in improving EBF and child anthropometric outcomes in Ethiopia [60,61].
Evidence on the effectiveness of CHW nutrition counselling interventions and CHW nutrition counselling plus supplementary food interventions suggest that a CHW intervention that only employs nutrition counselling is not effective in improving child growth, improving dietary diversity, or preventing the deterioration of acute malnutrition.However, CHW nutrition counselling can be impactful when used in conjunction with supplementary foods.No publications included in this systematic review researched whether the provision of supplementary foods is effective without the use of nutrition counselling and CHWs.In addition, little information was provided about the contents and implementation of the nutrition counselling in the included studies.These findings reflect similar results from a systematic review researching the use of food products to manage MAM and suggest a greater need to investigate the role of CHWs and nutrition counselling in the effectiveness of supplementary food interventions [62].
Both publications conducting economic evaluations on CHW models for the screening and management of uncomplicated SAM found the CHW interventions to be cost-effective.These results align with those of a recent systematic review on the costs and cost-effectiveness of child undernutrition treatment to households, health providers, organisations, and governments in LMICs [63].Acceptability and feasibility were also found to be high in the three publications that looked at these indicators, with community engagement, social mobilisation, and cultural appropriateness playing important roles.
When synthesising publications' findings on key barriers and facilitators to the CHW interventions, it became clear that while some elements were inherently barriers (such as insecurity, violence, and political unrest) or facilitators (such as community engagement), many of the key barriers or facilitators were not inherently challenging or supportive towards the success of the CHW model.Instead, they were key factors that, depending on the degree to which they were leveraged and integrated into programme design, became a challenge or a benefit.Some of these flexible factors include proximity of programming for beneficiary communities, degree of context specificity, level of community mobilisation, quality of training, and degree of integration of programmes.
Overall, the publications included in the review suggested that CHW models were effective, cost-effective, feasible, and acceptable in preventing, identifying, and managing child undernutrition.However, the success of the CHW model was reported to varying degrees, with some dependency on the type of intervention the CHWs delivered.In addition, the heterogeneity of the included publications with regards to study designs, outcomes, measurements, and interventions, while providing a comprehensive view of CHW child undernutrition work being done in CAFS, makes it difficult to synthesise across studies and generalise results, presenting significant challenges for policy application purposes.Future studies shall aim to harmonise data gathering across programs while still allowing for the capturing of contextspecific information.Such aim may be accomplished through convening a working group, that includes agencies, researchers, CHWs, policymakers, and affected populations, to establish a set of global definitions, indicators, and elements to report on, as well as templates, protocols and guidelines to follow.This shall include discussion of the options for CAFS classification, and recommendations for how and when to apply different classification methodologies.In addition, the use of clear definitions and in-depth descriptions of interventions, models, and implementation methods would allow for the extraction of valuable data, without confusion around what type of programme is being studied and by whom it is being delivered.
This systematic review was intentionally broad in its inclusion of different CHW interventions (screening, BCC, supplementary foods, identification, management, etc.), study designs (qualitative, quantitative, and mixedmethods), types of undernutrition addressed (AM, SAM, MAM, wasting, stunting, underweight, and micronutrient deficiencies) and outcomes (effectiveness, cost-effectiveness, ability, feasibility, acceptability, and scalability).The rationale is to generate a comprehensive review that could be leveraged for holistic policy and programme design in the field.While the review covered these various elements of CHW interventions for child undernutrition in CAFS, it revealed that little information exists on the effectiveness, cost-effectiveness, ability, feasibility, acceptability, and scalability of CHW models in preventing, identifying or managing child undernutrition in CAFS.
The identified evidence suggests that CHW models have the potential to deliver effective, empowering, equitable care for child malnutrition, while fostering resilience amongst conflicted-affected populations.The effective use of CHWs models requires an integrated, multisectoral, community-centred, and context-appropriate policy approach to be considered and embedded in policy, programme design, and implementation.To ensure strong community acceptability and mobilisation, CHWs and community members shall be engaged as codesigners, programmes shall be designed to minimise the distance needed to travel to receive services (with home visits being particularly effective), and strong, supportive links shall be created for any referral services needed from other parts of the healthcare system.To maximise CHW performance and ability, specific attention shall be paid to supervision, incentivisation, and training of CHWs.To inform policy design and implementation, anthropological research of the specific community context and desk research of similar contexts where CHW programmes have been implemented shall be undertaken.

Conclusion
This systematic review identified evidence to support the hypothesis that CHWs can play a key, beneficial role in the prevention, identification, and management of child undernutrition in fragile or conflict-affected settings, with CHW models being reported to be effective, costeffective, acceptable, feasible, and scalable.The strength of CHW interventions rests on key factors including the degree of community engagement, service proximity, presence of supervision, and quality of training.However, the sparse literature and scarcity of evidence weaken the findings; this is an issue that requires investment and attention from the field in order to build informed, sustainable, context-specific policy solutions for addressing child undernutrition in CAFS.

Fig. 1
Fig. 1 PRISMA diagram of included studies

Table 1
Final search terms and syntax for the Ovid online database search platform

Table 2
Study inclusion and exclusion criteria mum adequate diet, treatment coverage, IYCF indicators (including breastfeeding, exclusive breastfeeding, complementary feeding, and supplementary feeding), behaviour change, knowledge absorption, acceptability, feasibility, quality of care, and cost-effectiveness

Table 3 :
Summary table of study characteristics and findings (ordered alphabetically) and Appendix

Table 4 :
Summary table of study interventionsfor summaries of the included studies.

Table 3
Summary table of study characteristics and findings (ordered alphabetically)

Table 4
Summary table of study interventions