Use of Mid-Upper Arm Circumference by Novel Community Platforms to Detect, Diagnose, and Treat Severe Acute Malnutrition in Children: A Systematic Review

Limited studies suggest that with robust program inputs caregivers and CHWs can correctly use mid-upper arm circumference to detect severe acute malnutrition (SAM) and that properly trained and supported CHWs can treat uncomplicated SAM in communities.


INTRODUCTION
O f the approximately 16.4 million children aged 6-59 months worldwide estimated to experience severe acute malnutrition (SAM), roughly 7% to 13% receive treatment each year. 1,2While the growth of community-based management of acute malnutrition (CMAM) programs has considerably increased coverage of treatment for SAM over the past decade, continued gaps in coverage and a persistence of SAM have made identifying strategic methods for expanding access to care, and finding the means to leverage these methods at scale, an urgent public health need. 3he use of mid-upper arm circumference (MUAC) by health care providers to detect SAM was inextricably linked to the initial success of CMAM and is likely to remain an accurate, simple, affordable, and acceptable tool to facilitate further scale up of SAM detection and management.In the standard protocol for measuring MUAC to screen for acute malnutrition, a health care provider bends the child's left arm to locate and mark the midpoint.Then the arm is relaxed straight, the MUAC tape is wrapped around the midpoint, and the circumference of the arm is recorded to the nearest 1 millimeter. 1,2,4The platform for detecting, diagnosing, and treating SAM has typically been within CMAM programs in clinic settings; the merits and limitations of MUAC as an indicator of nutritional and mortality risk in such settings have been well described and debated in the literature. 3,5,6espite the word "community" being part of the CMAM acronym, there is seldom a measurement component at the household level.Currently, the standard protocol is being revisited with simpler, alternative protocols in mind, often involving MUAC measurement by community members and caregivers in household and community settings and/or integrating MUAC measurement into other existing platforms, such as part of growth monitoring activities, health campaigns, emergency services, and integrated community case management (iCCM) programs.Expanding the role of community health workers (CHWs) to include detection, diagnosis, and even treatment of uncomplicated SAM is also being explored as an element of decentralizing SAM care.CHWs work in the communities where they reside; we use the term inclusively, referring to both paid and volunteer workers, those working full time, and those working on an ad-hoc basis.CHWs have a decades-long history of successfully diagnosing and treating childhood illness, but their potential for addressing the burden of acute malnutrition remains largely untapped. 7he prospect of MUAC-focused management strategies led by caregivers and community members has great potential for enhancing public health impact by facilitating community sensitization and early treatment of affected children, reducing latestage clinical complications and hospitalizations, and increasing coverage of CMAM programs. 8The primary objectives of this systematic review are therefore to summarize the published and operational evidence describing (1) the use of MUAC by caregivers and CHWs in community settings for the detection and diagnosis of SAM, (2) the treatment of SAM by CHWs in community settings, and (3) health platforms where MUAC use and SAM management have been successfully integrated.

METHODS
This systematic literature review was conducted according to standards set by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 9tudies were eligible for review if they met the following criteria: Observational studies, experimental studies, intervention studies, and reviews were all eligible for inclusion.Excluded studies included those that were not available in English, those that did not directly address CHW or caregiver use of MUAC, those that occurred within clinic or hospital settings, and studies of infants younger than 6 months as MUAC is not currently a recommended indicator for acute malnutrition screening in that age group.
Five information sources were used for this review: A database search of peer-reviewed publications was conducted between September 15 and October 15, 2017, using PubMed and Google Scholar; the gray literature was searched on Emergency Nutrition Network (ENN) and Coverage Monitoring Network (CMN) websites between October 15 and October 30, 2017); records from the bibliographies of studies found in our database searches were retrieved between October 15 and October 31, 2017; and lastly, suggestions of relevant sources from experts in the field, including unpublished or operational materials, were received between September 15, 2017, and March 5, 2018.
The study selection process included 4 steps.First, a list of potential studies was compiled via our database search and from expert sources.Second, titles and abstracts were screened based on the eligibility criteria, and third, eligible articles were selected for full-text reading and further screening.The bibliographies of full-text articles were also screened for additional articles that were eligible for inclusion.Finally, articles that met the criteria were submitted for data extraction.For published literature, a standardized form was used for simultaneous data retrieval and data entry of the following items: record reference, objective of study, study design, study population, intervention, control (if any), key findings, operational lessons, study strengths and weaknesses, and comments.For gray literature sources, the reference, organizational source and setting, the objective of the document, and key messages were recorded.Data were extracted by one author and reviewed by second authors.We synthesized the findings of the included studies by summarizing key findings, identifying trends across studies, and noting operational challenges faced during implementation.Disagreements were mutually resolved between all authors.Both the published and the operational materials informed our conclusions, with recognition that individual studies may be biased toward publication of positive results.

RESULTS
We screened 1,072 records and selected 43 records for full-text screening.We included 22 studies in the review (Figure).A brief summary of each of the 22 studies reviewed is presented in Table

Caregiver Detection of SAM Using MUAC in Community Settings
We identified 3 research studies of caregiver detection of SAM using MUAC tapes (Niger   2015) found that mothers were capable of using standard MUAC tapes to classify SAM cases with >73% sensitivity and >98% specificity in a pilot study in Niger; this was comparable with CHW performance of 80% and 96%, respectively. 10Mothers responded positively to being engaged in monitoring their child's nutritional status, and their comprehension of how MUAC classification corresponded to admission (or exclusion) from SAM treatment programs improved.The simplified MUAC protocol-measurement of either arm, at a midpoint ascertained visuallyperformed as well as the standard protocol. 10,13In a follow-up study in Niger, Alé et al. (2016) observed that in areas where caregivers were using MUAC tapes to detect SAM in their own households in Niger, the median MUAC of SAM cases admitted to outpatient therapeutic programs was significantly higher than in areas where CHWs were doing the screening. 11There were also fewer complicated cases and hospital admissions among motherreferred cases.
Most recently, Grant et al. (2018) compared the performance of 3 prototype "Click-MUAC" devices with an improved MUAC insertion tape ("UniMUAC" tape) among caregivers in Kenya. 12he 3 prototypes, which resembled plastic cuffs and tapes, had internal circumferences of 115 mm or 115-125 mm and were hoped to improve casefinding sensitivity over the standard MUAC insertion tapes.Each of the prototypes yielded high sensitivity (>93%) and specificity (>98%), but the UniMUAC tape was superior in both sensitivity (100%) and level of agreement between caregiver, health facility staff, and data collection staff measurements (98%) when screening for SAM.

Operational Findings
ALIMA, the nonprofit that runs the Niger-based programs detailed earlier, has expanded its "Family MUAC" programs to Burkina Faso, Chad, and Mali. 14,15Action Against Hunger piloted "MUAC Mothers" programs in India and Mauritania with mixed results.In their India pilot, approximately 30% of mothers trained to use MUAC tapes measured their children in the 7 months following training.In Mauritania, more than 6,000 mothers were trained in both MUAC use and edema detection; outcomes were not available at the time of review. 16Several other organizations (Médecins Sans Frontières, GOAL, Concern, World Vision, International Red Cross, International Medical Corps, Cooperazione Internazionale, and Valid International) are in the process of adopting and adapting Family MUAC (also known as "Mother MUAC" and "MUAC Mothers") programming, 17 and tools to improve caregiver training and monitoring are in development. 18

CHW Detection and Diagnosis in Community Settings
We found evidence supporting high CHW capacity to accurately diagnose SAM using MUAC in 3 studies of 2 interventions in Bangladesh and Mali [19][20][21] Two studies presented mixed evidence from Niger and Pakistan, 11,13 and two from

Description of program/ materials
To identify and describe models for how CHWs currently incorporate SAM screening, referrals, and treatment into their work.While there is evidence supporting CHW capacity to conduct all SAM-related activities, there are outstanding questions regarding the conditions that foster success, as well as the optimal mix of iCCM and nutrition-related responsibilities.

Caregiver detection Description of program/ materials
To describe the expansion of "Family MUAC" concepts in Burkina Faso, Chad, Mali, and other locations.

Caregiver detection, CHW diagnosis and treatment
Pilot study To report on lab testing of an alternative MUAC strap for use with adult and child populations.Initial testing of the strap using a standardization process (not on humans, but on differently sized cylinders) showed it to be more accurate and have a higher sensitivity than the standard UNICEF strap.The next step is to test the straps on children in a field setting.

Continued
Malawi indicated success but did not report specific measured outcomes. 22,230][21] Alvarez-Moran et al. (2017) found that CHWs correctly assessed MUAC in 97% of children. 19In Bangladesh, Puett et al. (2012)  observed that CHWs completed MUAC measurements correctly >96% of the time. 20e found mixed evidence of CHW MUAC measurement in 2 other studies, one from Niger and one from Pakistan.Alé et al. observed that mothers' MUAC classifications agreed more often with nurses' than those made by CHWs in Niger: mothers' measurements agreed with nurses' 75% of the time, contrasted with only 40% of the time between CHWs and nurses.CHW performance in this case was not described as deficient, despite the discrepancy when compared with caregiver performance. 11 Rogers et  al. (2017) report that MUAC was correctly measured for just 57% of children in their study from Pakistan. 24While the authors concluded that CHWs are capable of accurate SAM diagnosis, they speculated that the low rate of correct MUAC measurements was due to operational constraints and low CHW motivation.

CHW Treatment of SAM in Community Settings
0][21][22][23] The findings of these studies, which span clinical outcomes of SAM cases, quality of care provided, and costeffectiveness, indicate that CHWs are capable of providing high-quality, effective care for uncomplicated SAM at a lower cost than inpatient care models, given adequate operational support and supervision.
CHWs provided correct medical care for uncomplicated SAM cases in 75% of cases and managed RUTF supplies correctly for all cases studied by Alvarez-Moran et al. ( 2017) in Mali. 19he majority (80%) of cases were concluded to have received high-quality treatment, defined by Alvarez-Moran as meeting essential indicators across 5 dimensions of care (interface with caregiver, evaluation, classification, treatment, and counseling).Similarly, in Puett's 2012 Bangladesh study, 90% of SAM cases were managed without error; the addition of SAM management to CHWs' regular responsibilities did not appear to affect quality of care or clinical outcomes. 20(See also Sadler 2011. 25 2017) determined that high levels of supervision likely contributed to the high quality care they observed in Bangladesh and Mali, respectively. 19,20mthor et al. ( 2009) and Linneman et al. (2007) focused on clinical outcomes of SAM cases receiving treatment from CHWs in Malawi; given the high SAM recovery rates (94% and 89%), we presume that the quality of care they received from CHWs was adequate. 22,23e found evidence of unsatisfactory CHW management of SAM in 2 instances, one in the study by Alvarez-Moran et al. in Mali (2017) and in work done by Grant et al. in Pakistan (2018). 19,24lvarez-Moran et al. (2017) found in Mali that for some tasks, such as administering antibiotics, CHW performance was deficient. 19Similarly, in the study by Rogers et al. (2018) in Pakistan, quality of CHW care for SAM was not provided at a consistently high level.While 68% of uncomplicated SAM cases received the correct medical and nutritional care (RUTF, antibiotics, and folic acid), only 4% of cases received the full package of medical and nutritional care and nutritional counseling messages.This low compliance was attributed to operational challenges-namely stock outagesand low CHW motivation owing to lack of extra remuneration. 13Nonetheless, and important to note, is the fact that the clinical outcomes of SAM cases in the Pakistan study were noninferior to traditional facility-based models.In other words, failure to deliver the full package of care did not result in low performance or SAM recovery outcomes relative to other care modalities.

Operational Findings
There are ongoing efforts to create tools that facilitate SAM diagnosis and treatment by low-literacy CHWs in Chad, India, Mali, and South Sudan.These include the use of visual materials for RUTF dosage and icons to enable reliable case documentation and monitoring. 14Simplified dosage protocols that do not depend on measuring child weight, as described in Phase 1 of the Combined Protocol for Acute Malnutrition Study (ComPAS) in Chad, Jordan, Kenya, Pakistan, and Yemen, will also facilitate CHW treatment. 26How to safely combine low-literacy tools with the ability to administer antibiotics, assess for danger signs, and refer appropriately when needed requires further exploration.

Integration of MUAC Use and/or SAM Care Into Other Platforms
Our search yielded 2 published research studies that explicitly described examples of MUAC screening and/or SAM care integration into existing health and/or nutrition platforms.Maust et al.  (2015) tested the effectiveness of an integrated treatment program for SAM and moderate acute malnutrition (MAM) in Sierra Leone, which relied on the sole use of MUAC as the indicator used for admissions, monitoring, and discharge, and on RUTF as the sole treatment food.They found the exclusive use of MUAC to be conceptually and logistically simpler, and their results of high coverage (71%) and recovery rates (83% for SAM) are promising. 27Other studies (ComPAS) are underway to examine the effectiveness of similar joint protocols in other settings. 26irandutiye et al. (2011) reported on the use of MUAC during National Nutrition Week activities in Mali; the event typically has 80% to 90% coverage nationwide and is a promising partner platform for SAM screening and referrals.The difference in screening coverage was substantial: 52% of eligible children (those in the 6-59-month age range) were screened during the event, compared with 22% screened at health centers and 5% screened in the community in the months following the event. 28The discrepancies in screening rates suggest that both facility and communitybased SAM screening have considerable room for improvement: half of the children with acute malnutrition in their survey had been at a health center within the previous 4 months, but only a quarter of them had been assessed for malnutrition.

Operational Findings
The gray literature provides additional support for different models of integration, some of which are already operational and others that are still hypothetical.Friedman and Wolfheim (2014) describe the current range of operational models, 2 of which pertain to SAM screening, referrals, and treatment by CHWs in community settings.In one model, CHWs assess and refer SAM cases as part of iCCM, and in a second model they also provide treatment 29 ; we found evidence supporting the success of both models in Bangladesh and Niger in the course of our review. 11,20,25

Caregiver Detection of SAM Using MUAC in Community Settings
Caregiver-focused models for detecting and classifying SAM using MUAC have potential for increasing coverage and detecting acute malnutrition earlier than standard MUAC protocols.Given the rapid expansion of simplified protocols for caregiver-led pilots and programs, the evidence base is likely to broaden in the next few years.While it is too early to assess the sustainability of caregiver MUAC programs, there is some evidence of high variation in the level of involvement and activity that mothers invest following training.For example, at ALIMA's initial site in Niger, 60% of mothers in the project area have been trained and >70% of CMAM admissions are now referred by mothers. 15In contrast, 7 months after an MUAC Mothers training in India, only 30% of mothers reported having ever measured their children. 16Understanding how to best motivate and engage caregivers to participate in MUAC measurements and addressing any barriers or stresses created by this responsibility are important questions going forward.
Further simplification of MUAC protocols may hold additional potential in increasing effective community/caregiver MUAC use in other settings.Potential simplifications include embracing "classification" rather than "measurement," the use of wider color-banded MUAC tapes, the use of either arm for measurements, and/or visually locating the arm midpoint. 10

CHW Detection and Diagnosis of SAM Using MUAC in Community Settings
Most of the available evidence supports the ability of CHWs to reliably measure MUAC to the standard necessary for SAM screening.It is worth noting that although MUAC is widely considered a simple indicator for measurement and interpretation, there are examples where the metric was not accepted by CHWs and/or not accurately measured by community workers in the context of growth monitoring and basic anthropometric training [30][31][32] (not included in this review as they did not meet inclusion criteria).
Thus, despite widespread assertions of the simplicity of MUAC, some users-CHWs, nurses, and other health professionals included-find measuring MUAC to be a challenging task.This may partially explain why screening rates and detection of SAM cases remain low even at clinics where staff have been trained in CMAM protocols. 28f the findings by Blackwell et al. (2015) hold true, further simplifying the MUAC protocol may address issues of familiarity and understanding of MUAC among health care providers and caregivers alike. 10

CHW Treatment of SAM in Community Settings
Most studies reviewed here indicate that CHWs are capable of performing the tasks associated with SAM diagnosis and treatment when supervision, training, and motivation are satisfactory.However, little has been published to describe the effectiveness on child outcomes or coverage relative to standard approaches.One consistent message across studies of CHWs and SAM treatment, both in the published and the gray literature, is the importance of quality training, regular refresher trainings, and high levels of supervision to ensuring sustained CHW motivation, activity, and effectiveness. 19,24,25The cost of supervision is likely to constitute a large proportion of overall program costs, particularly when implementing a new CHW-led program.Notably, there is evidence that in countries with a well-established CHW network, rapid implementation of CHW-led SAM care in an emergency setting may be highly feasible and effective. 22

Integration of MUAC Use and/or SAM Care Into Other Platforms
Several studies indicate that the use of MUAC by caregivers and CHWs is a missing link in the integration of SAM care into other platforms, be it integration with MAM programming; 27,33 integration into nutrition or health-focused events such as nutrition weeks, vaccination campaigns, nutrition events, water, sanitation, and hygiene (WASH) programming, or well-baby clinics; 28,34 or incorporating MUAC assessment and treatment into existing iCCM programs. 35As many of the studies reviewed in this article suggest, and as reported by a study of the integration of SAM care into Niger's national health system, the complexity of acute malnutrition intervention protocols (multiple indicators, treatment types, locations, and steps in care) hinders integration. 36ntegrating SAM care successfully does not necessarily require integrating every aspect of its treatment, nor is integration a solve-all for inadequate resources or staff. 37The optimal mix, or level of mixing, between MUAC use, SAM care, and other activities will likely vary widely depending on context, content, and complementarity of joint activities, among other factors.

Study Limitations
As with most systematic reviews, this review is subject to publication bias.Published research on MUAC use is likely biased in favor of successful programs; as a result this review may be missing important lessons learned from pilot programs and trials that did not show an apparent positive impact of caregiver or CHW use of MUAC for SAM detection, diagnosis, and/or treatment.Nonetheless, our review did yield work reporting negative outcomes (see Rogers et al. 2017 13 ).Most of the studies reviewed here are observational rather than interventional, so inference is limited.Furthermore, most of the published evidence describes intensively delivered, small-scale interventions, and is not readily generalizable to larger scale, public-sector health care systems.With few exceptions, the studies reviewed here contain little information about the demographic and socioeconomic characteristics of study participants.The absence of this contextual information, the narrow geographic scope of the available studies, and variation in how/whether training and remuneration were delivered limits our ability to generalize findings.Differences in settings, cultures, training protocols, and populations will have implications for the usefulness and acceptability of MUAC.
The issue of treating children classified as SAM based on low weight-for-height, rather than MUAC, at household level, is not addressed in this review.There is evidence that MUAC and weight-for-height do not detect the same children; 38 however, there is also strong evidence that MUAC is a better predictor of children at high risk of death. 5,39While community studies demonstrate that MUAC is better at predicting mortality, children with low weight-for-height still remain at risk. 5 Some studies have reassessed MUAC cutoffs in order to better capture children classified as SAM by current weight-for-height criteria. 40Others argue that MUAC and weight-forheight identify different children at risk and both should be retained as independent criteria. 41e did not address the differences between MUAC and other metrics or implications for the screening, diagnostic, or treatment models reviewed here.Vulnerable children not detected by MUAC alone still need care and should not be overlooked in efforts to scale up and integrate SAM care.Efforts to determine MUAC cutoffs that are sensitive to other metrics-such as low weight-for-height-do exist (see Fiorentino et al. 2016 40 ).Determining the impact of CHW treatment programs on children with low weight-forheight will be necessary if an MUAC-focused model is to be rolled out at scale. 8

CONCLUSIONS
There is a limited but growing amount of evidence describing the use of MUAC for detection, diagnosis, and treatment of SAM by caregivers and CHWs in community settings.The number of published research studies is small, their geographic scope is narrow, and most describe intensive, small-scale interventions and pilot studies.As such, findings should not yet be extrapolated to other settings.From our review of published research studies, case studies, and operational materials, we can conclude the following: Caregivers are able to use MUAC to detect SAM in their children with little apparent risk and many potential benefits to early case detection and coverage.
CHWs are able to diagnose SAM and provide high-quality treatment with increased coverage, particularly when using simplified protocols and when supported by strong supervision.Without adequate supervision, training, and/or remuneration, the quality of care is likely to suffer.The correct administration of antibiotics by CHWs as well as the ability to detect and refer for danger signs requires further consideration.There is currently only limited evidence on SAM treatment outcomes using this approach, and little consideration of the impact on other child health outcomes under the remit of CHWs.
While most practitioners consider MUAC relatively simple, it still requires good training and community advocacy in settings where it is unfamiliar.Ongoing research into simplified protocols, modified MUAC tapes, MUAC-based RUTF dosages, and low-literacy treatment tools will all support diagnosis and treatment of SAM by CHWs at household level.The available literature indicates that improvements in coverage are likely when SAM management protocols are simple and the proposed platform for integration is composed of complementary nutrition-related activities.
In recognizing the benefit of MUAC as a community tool, it is important to recognize that current MUAC criteria do not select for all high-risk children, including low weight-forheight children, and the optimal approach will vary across different contexts.More research is needed to identify different options to identify these high-risk children in the community and ensure successful diagnosis and treatment.
In conclusion, scaling up the use of MUAC to detect SAM in communities is a promising step toward greater coverage and use of existing CMAM services.Given adequate operational support, training, and supervision, the quality of care for self-referred and/or CHW-treated cases is likely to be comparable with current health worker and CMAM models.Further research regarding scalability and applicability of community MUAC use across a wide range of contexts is needed and warranted.In some contexts, the use of MUAC as the primary criterion for detection, diagnosis, and discharge may be appropriate and should be explored further.

TABLE 1 .
Summary of Published Research Studies Included in Review (n=11) Specificity was also high for all devices (>96%), with no significant differences observed between the insertion tape and the "Click-MUAC" devices.Continued

TABLE 1 .
Continued Continuedor alternative MUAC devices (Kenya 12 ) that indicate that caregiver-focused approaches are reliable and feasible.Blackwell et al. (

TABLE 2 .
Summary of Operational Materials Included in Review (n=11)