Mothers screening for malnutrition by mid-upper arm circumference is non-inferior to community health workers: results from a large-scale pragmatic trial in rural Niger

For the past 20 years, the diagnosis and treatment of severe acute malnutrition (SAM) has become increasingly decentralized, from a strictly hospital-based approach for all cases to the current model of outpatient care for children with uncomplicated SAM and inpatient care for children presenting with complications or failing to respond to treatment. [1] Even though this decentralization has led to a scaled-up response, only 10–15 % of the nearly 20 million children suffering from SAM have access to treatment [2].

Since the 1980s, it has been shown that mid-upper arm circumference (MUAC) can be performed by minimally trained personnel [3] leading to current recommendations for community health workers (CHWs) to screen by MUAC [4]. MUAC holds many advantages: It is a better predictor of mortality, especially when repeated frequently over time [5], compared to other practicable anthropometric measures such as the weight-for-height z-score (WHZ) [6‐8]. MUAC is simple to use; and regular screenings in the community increases early diagnosis and reduces the risk of mortality and morbidity requiring costly and specialized hospital care [9, 10].

Mothers are in the best position to detect signs of nutritional deterioration in their own children, and training mothers to regularly screen by MUAC and check for edema is a logical next step in this process of decentralization. This could lead to improved coverage, as well as earlier detection and treatment seeking that improves program outcomes and reduces cost per case treated. A recent study comparing CHWs and mothers performing MUAC by color-coded class rather than measurement showed the promise of such an approach [11]. Mothers had a high sensitivity and specificity for SAM and moderate acute malnutrition (MAM, 115 mm < = MUAC < 125 mm), there were high levels of agreement between the mothers and CHWs and a similar number of classification errors which occurred only at the boundaries between normal/MAM and MAM/SAM. Accuracy was not influenced by which arm (right or left) was measured nor by how the mid-point of the upper arm was determined (by-eye or by measurement), providing evidence that could simplify training while maintaining accuracy and precision.

Food insecurity and infections leading to childhood malnutrition is a seasonal and recurrent crisis for families in Niger, particularly in the rural areas around Maradi and Zinder where rates of global acute malnutrition (GAM, defined as the total of MAM and SAM), and stunting are consistently near or above emergency thresholds [12]. In 2014, 364,837 children were admitted for SAM treatment by all actors in the country, and even with the investments Niger has made in nutrition over the past decade, 47,225 of these children (12.9 %) still required hospitalization. [13] Such a high need for inpatient services is problematic in a country like Niger which faces a chronic shortage of qualified health personnel [14].

The objective of this study was to evaluate the efficacy (in terms of point coverage, MUAC at admission, and the need for inpatient care) and costs of two community-based strategies to screen children aged between 6 and 59 months for SAM by MUAC class. Mothers and caretakers (henceforth referred to as mothers) were compared to CHWs in separate health zones in Niger’s Mirriah District. It was hypothesized that training mothers to perform MUAC would lead to earlier SAM detection and attendance and lower rates of hospitalization, thus relieving pressure on a resource-poor health system by reducing the need for inpatient care.

This pragmatic interventional, non-randomized efficacy study was conducted according to the guidelines laid down in the Declaration of Helsinki. The study protocol was given ethical approval by the National Consultative Ethics Committee of Niger’s Ministry of Public Health on April 10, 2013 (Deliberation Number 006/2013/CCNE.) Written informed consent was obtained from all mothers or heads of households. The trial is registered with clinicaltrials.gov (Trial number NCT01863394).

The trial was performed from May 2013 to April 2014 in Mirriah, a rural district in Niger’s southeast Zinder region bordering Nigeria. Dogo (Mothers Zone) served as the intervention zone and Takieta (CHWs Zone) served as the comparison zone where screening by color-coded MUAC class (Fig. 1) was performed by mothers and CHWs, respectively.

A total of 12,893 mothers and caretakers were trained to screen children by MUAC color-coded class and check for edema in the Mothers Zone during the initial training sessions and subsequent coverage surveys. (Table 3) A further 329 previously untrained mothers were trained at the health center in the Mothers Zone during the study. In the CHWs Zone, 36 CHWs were trained and given appropriate supervision.

Referrals by MUAC <115 mm in the Mothers Zone were more likely to be in agreement with the health center agents compared to the CHWs Zone (risk ratio = 1.88 [1.69; 2.10], risk difference = 35.31 [30.39; 40.23], p < 0.0001). (Table 4) The small number of referrals for edema in either zone (42 in the Mothers Zone and 1 in the CHWs Zone) means that this analysis lacks statistical power (p = 0.4471). Table 5 shows the percentage of children admitted by each of the inclusion criteria. Table 6 shows the source of referral for those children admitted by MUAC < 115 mm, with a larger percentage in the Mothers Zone reporting that they came to the health center spontaneously suggesting that training mothers led to increased awareness of malnutrition and available treatment. The distribution of MUAC at admission for children admitted to SAM treatment by MUAC <115 mm in the two zones is shown in Fig. 3. The difference between the medians was estimated to be 1.6 mm higher in the Mothers Zone (95 % CI = 0.65; 1.87) using a smoothed bootstrap procedure; the null hypothesis of no difference between the means was rejected (p = 0.007). Consistent with earlier detection and treatment seeking, children admitted in the Mothers Zone were less likely to require inpatient care than children in the CHWs Zone, both at admission and at any point in their treatment episode, with the most pronounced difference at admission for those enrolled by MUAC < 115 mm (risk ratio = 0.09 [95 % CI 0.03; 0.25], risk difference = −7.05 % [95 % CI −9.71 %; −4.38 %], p < 0.0001). (Tables 7, 8) Point coverage was similar in both zones at the end of the study (35.14 % Mothers Zone vs 32.35 % CHWs Zone, difference 2.78 %, [95 % CI −16.34 %; 21.90 %], p = 0.9484, Yates corrected chi-square test.) (Table 9) Coverage improved in both health zones, especially after the second survey likely because the seasonal spike in malaria prompted mothers to go to health centers at the first sign of a child’s fever. During the fourth coverage survey, 6,678 of the 7,421 mothers (90.0 %) surveyed in the Mothers Zone had been trained in the use of MUAC. Of these, 6,529 (97.8 %) still possessed the MUAC tape in good condition. Lost or damaged MUAC tapes were replaced.

Even though the Mothers Zone required a higher initial investment, overall costs for the year were less than half those in the CHWs Zone ($8,600 vs $21,980.) (Table 10) While initial costs for the CHWs Zone were low, the modest monthly incentives for CHWs represented the largest part (85 %) of the higher costs.

Late presentation of SAM (e.g. children well below the admission threshold) carries a greater risk of death, and is commonly associated with medical complications that require more costly hospital-based care. [17] This study demonstrates that earlier detection of SAM can be achieved by training mothers to classify the nutritional status of their children by regular MUAC screenings. Mothers were not inferior to CHWs in terms of coverage at a substantially lower cost, and regular MUAC screening by mothers, in addition to the mass sensitization from training sessions and quarterly surveys, contributed to a higher median MUAC at admission and lower hospitalization rates at admission and during the course of treatment.

There is growing understanding that MUAC can be used safely and effectively as the sole anthropometric criterion for admission, management, and discharge from malnutrition treatment [18‐21]. A recent study also showed promising results from integrating SAM and MAM programs under one MUAC-only protocol [22]. Such an integrated approach could be important moving forward because having separate SAM and MAM programs often poses problems, leading to convoluted messaging and uneven health care delivery. It was likely confusing for mothers in both zones, for example, to be told to go to different programs based on MUAC color at screening (i.e. red for SAM and yellow for MAM). Thus a mother-centered screening strategy will be most successful in programming that integrates SAM and MAM treatment and relies on MUAC from the home to the health center, helping to facilitate a scale-up that meets current global needs.

Coverage surveys in multiple contexts have identified major reasons why mothers do not utilize available SAM treatment, including a lack of awareness about malnutrition (signs, etiology, and treatment) and existing CMAM programs, as well as previous rejection [23, 24]. Barriers to coverage were noted at baseline and not analyzed further here, but mothers welcomed an approach that allowed and expected them to be more engaged in decisions related to their children’s health. Several reported understanding for the first time why their child was admitted (or not) for malnutrition treatment in the past. Furthermore, mass sensitization from trainings and repeat opportunities for health education activities will help address these coverage barriers, especially in an area like rural Niger where even in the dry season only 34 % of people live within 5 km of a functioning health center [25].

There were several limitations and possible biases to this study. The study only lasted eleven months and it takes time for members of a community to develop new health-seeking behaviors, so community mobilization is likely to improve over time through periodic trainings and MUAC tape renewals. Although it is likely that screening by mothers contributed to the observed difference in proportion of hospitalized cases in the two zones, this is not certain as hospital referrals depend on many factors (e.g. clinicians’ level of training and/or experience). Even though the two zones had similar demographics and prevalence of malnutrition at baseline, a smaller proportion of children receiving SAM treatment in the Mothers Zone required inpatient care in a nine month period prior to the study. The hospital in Mirriah is also further away from the CHWs Zone than the Mothers Zone, but the potential for bias was reduced by the fact that transport for referrals from health centers was provided by ALIMA/BEFEN ambulances in both zones. One would expect program lengths of stay to be shorter in the Mothers Zone, but comparison was not possible because the therapeutic supplementary feeding programs (i.e. programs treating MAM) in the zones were operating at different levels of capacity, causing children to be retained in the therapeutic feeding program longer in the Mothers Zone than in the CHWs Zone. The difference in availability and quality of MAM services in the two zones is another limitation to the overall analysis because it is difficult to determine the impact or influence of MAM programming on the profile of SAM cases while the presence of a well-functioning MAM program may have influenced discharge decisions.

Screening by MUAC class itself also introduces error in favor of sensitivity since it does not take much to misclassify children near the 115 mm cut-off due to pulling too tightly. This can be reduced by using increased-width tapes. There were, however, few gross errors when mothers or CHWs measured MUAC, and most misclassifications in the both zones occurred either at the boundary between SAM and MAM or children screened as SAM determined to be MAM by health center agents, in line with a previous study [11]. With a condition such as SAM which has a high case-fatality rate when untreated, an error in favor of sensitivity is more acceptable than the opposite error provided that rejected referrals do not undermine coverage.

The length threshold for MUAC eligibility that is part of many protocols in West and Central Africa (and which changed in Niger from 65 cm to 67 cm during the study) introduced an additional bias. It systematically excludes young, stunted children who are at considerable risk of death if left untreated, which is especially problematic in an area of high stunting and in light of previous work showing such children respond well to treatment [26, 27]. There is no way of knowing how many low MUAC children were excluded because they were below the height cut-off. Removing the length restriction for MUAC at triage and screening, and instructing mothers to simply use MUAC on children older than 6 months, would end this exclusion and simplify MUAC trainings.

Some areas were identified for improvement or further exploration. Because resource-poor mothers in a rural community were shown to better identify malnutrition in their infants through pictorial rather than verbal descriptions, integrating more pictorial descriptions into the trainings would be beneficial [28]. Future messages can also more explicitly convey that bringing a child at the earliest sign of nutritional deterioration could reduce, though not eliminate, the risk of their child needing to be hospitalized. Care must be taken, though, to assure mothers do not stop treatment once they start to see a noticeable improvement in their child’s MUAC status.

While mothers and CHWs were trained to screen for edematous malnutrition, prevalence is low in Niger and the small number of referrals for edema in either zone means that the analysis here lacks statistical power. Nevertheless, more children were referred for edema and agreement was higher in the Mothers Zone, suggesting that mothers are in a better position to detect potential cases of a deadly condition with rapid onset and resolution in either death or spontaneous recovery. In the Mothers Zone, however, fewer children were admitted to the treatment program for edema than those who were determined to be in agreement with an edema referral. Because most of the children admitted for edematous malnutrition in both zones were classified as having mild edema (in both feet, according to WHO definitions [5]), this discrepancy was likely due to subjective assessor judgment between health agents and nutrition assistants, as even trained nurses have been shown to have difficulty reliably identifying edematous malnutrition in low prevalence settings [29]. It will be particularly important to further study mothers’ ability to detect edema in an area of high prevalence.

CHWs provide valuable contributions to improved health outcomes in a community, and are being relied upon to do more and more, from diagnosis to delivery in areas as varied as malaria, malnutrition, vaccination, or childbirth assistance [30]. These activities are in addition to their normal work and family responsibilities, and are sometimes expected to be performed without pay. CHWs will continue to play an important role in community-based interventions, but shifting some tasks such as screening for malnutrition to mothers can prevent CHWs from becoming overburdened. CHWs, for example, might be better utilized as MUAC trainers. Conversely, providing mothers with simple tools to encourage their active involvement need not be restricted to screening for malnutrition. Mothers in a similar context, for example, were shown to be capable of identifying and classifying respiratory tract infections in their children [31].

Even with the study’s limitations, preliminary analysis of the benefits and few risks associated with the household-level screening strategy led ALIMA to train tens of thousands of additional mothers elsewhere in Niger as well as in Burkina Faso, Mali and Chad. Thus from an operational perspective integrating household-level screening into pre-existing programs was relatively straightforward.

Involving mothers in screening their own children by MUAC and checking for edema is a key step in increasing access to care for children in any area where malnutrition poses a high risk of death or illness and can reduce cost per child treated. Making mothers the focal point of MUAC screening strategies should be included in regular CMAM programming. As mothers are taught how to screen their own children in countries where malnutrition is highly prevalent, important information will be gained by practitioners about how to most effectively extend the strategy.

MUAC, mid-upper arm circumference; CHW, community health worker; CMAM, Community Management of Acute Malnutrition; GAM, global acute malnutrition; SAM severe acute malnutrition; MAM, moderate acute malnutrition; WHZ, weight for height Z-score; RUTF ready-to-use therapeutic food; WHO World Health Organization