Article Text
PDF
Abstract
Background Childhood stunting remains common in many low-income settings and is associated with increased morbidity and mortality, as well as impaired child development.
Methods The main objective of the study was to assess whether home-installed growth charts as well as small-quantity lipid-based nutrient supplements (SQ-LNS) can reduce growth faltering among infants. All caregivers of infants between 2 and 10 months of age at baseline, and at least 6 months old at the beginning of the interventions, in 282 randomly selected enumeration areas in Choma, Mansa and Lusaka districts in Zambia were invited to participate in the study. Cluster randomisation was stratified by district. A software-generated random number draw was used to assign clusters to one of four arms: (1) no intervention (control); (2) home installation of a wall chart that contained a growth monitoring tool along with key messages on infant and young child feeding and nutrition (growth charts only); (3) 30 sachets of SQ-LNS delivered each month (SQ-LNS only) or (4) growth charts+SQ LNS. The primary outcomes were children’s height-for-age z-score (HAZ) and stunting (HAZ <−2) after 18 months of intervention. Secondary outcomes were haemoglobin (Hb), anaemia (Hb<110.0 g/L), weight-for-height, weight-for-age z-score (WAZ), underweight (WAZ<−2) and child development measured by the Global Scales of Early Development (GSED). Outcomes were analysed intention to treat using adjusted linear and logistic regression models and compared each of the three interventions to the control group. Assessors and analysts were blinded to the treatment—blinding of participating families was not possible.
Results A total of 2291 caregiver–child dyads across the 282 study clusters were included in the study. 70 clusters (557 dyads) were assigned to the control group, 70 clusters (643 dyads) to growth charts only, 71 clusters (525 dyads) to SQ-LNS and 71 clusters (566 dyads) to SQ-LNS and growth charts. SQ-LNS improved HAZ by 0.21 SD (95% CI 0.06 to 0.36) and reduced the odds of stunting by 37% (adjusted OR, aOR 0.63, 95% CI (0.46 to 0.87)). No HAZ or stunting impacts were found in the growth charts only or growth charts+SQ LNS arms. SQ-LNS only improved WAZ (mean difference, MD 0.17, 95% CI (0.05 to 0.28). No impacts on WAZ were seen for growth charts and the combined intervention. Child development was higher in the growth charts only (MD 0.18, 95% CI (0.01 to 0.35)) and SQ-LNS only arms (MD 0.28, 95% CI (0.09 to 0.46). SQ-LNS improved average haemoglobin levels (MD 2.9 g/L (0.2, 5.5). The combined intervention did not have an impact on WAZ, Hb or GSED but reduced the odds of anaemia (aOR 0.72, 95% CI (0.53 to 0.97)). No adverse events were reported.
Interpretation SQ-LNS appears to be effective in reducing growth faltering as well as improving anaemia and child development. Growth charts also show the potential to reduce anaemia and improve child development but do not seem as effective in addressing growth faltering. Further research is needed to better understand reduced effectiveness when both interventions are combined.
Trial registration number NCT051204272.
- Child health
- Nutrition
- Stunting
Data availability statement
Data are available on reasonable request. Data will be available on request to the first author.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Small-quantity lipid-based nutrient supplements (SQ-LNS) is increasingly recognised as a key intervention to reduce growth faltering among children under age 6–59 months. Furthermore, SQ-LNS appears to be generally well received by parents and met with high rates of compliance in terms of daily consumption. Home-based growth charts have recently been proposed as a relatively inexpensive intervention to increase parental awareness of children’s nutritional needs and to improve children’s growth trajectories. No study has combined SQ-LNS with home-based growth charts yet.
WHAT THIS STUDY ADDS
To our knowledge, the ZamCharts trial is the first trial to test the feasibility and impact of large-scale, systematic home-based distribution of SQ-LNS and growth charts to infants and young children in low-resource settings with high rates of malnutrition. The project was designed in close collaboration with the Zambian Ministry of Health in order to generate the evidence needed for decision-making regarding a potential implementation of this programme at scale. Three highly diverse districts were chosen for the study and infants in 282 randomly selected enumerations areas enrolled in the trial to ensure a nationally representative study sample. Our results suggest that home-based distribution of SQ-LNS leads to substantial improvements in children’s nutritional status and development. Growth charts seem to have a positive effect on parental behaviour and child development but do not appear to modify average growth outcomes in this setting.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The results of this study further strengthen the growing evidence highlighting the positive impacts of SQ-LNS on children’s health and development. In countries with high rates of growth faltering and food insecurity, SQ-LNS may be one of the most effective strategies to improve children’s nutritional and developmental outcomes. Our study suggests that SQ-LNS distribution can be done through a centralised distribution network and does not necessarily need to be integrated into health-centre-based programmes. Growth charts may offer a relatively easy and inexpensive way to provide parents with key behaviour and nutritional information.
Introduction
Despite moderate improvements in children’s average linear growth over the past decade,1 the global burden of child undernutrition remains high. According to the latest estimates, 148 million children under age 5 experienced linear growth faltering in 2022.2 Levels of stunting are particularly high in Africa, Asia and Oceania, where more than 30% of children currently experience linear growth faltering.2 Early-life malnutrition increases the risk of impaired child development,3–5 reduced educational attainment4 6–8 and chronic disease in adulthood.9
Despite a large number of behavioural,10 11 economic12 and water, sanitation and hygiene (WASH) trials,12 effective interventions to reduce linear growth faltering in low-income and middle-income countries (LMICs) remain scarce.13 One of the key challenges from a caregiving perspective is the weak link between stunting and children’s well-being perceived by caregivers, which often results in stunting not being recognised,14 especially in communities where linear growth faltering is common, and routine growth monitoring is limited.15
Home-based growth charts provide a potentially cost-effective way to help parents monitor their children’s growth trajectories.16 Growth charts are easy to set up in low-resource settings, can be used to provide standard infant and young child feeding (IYCF) messages and also serve as a visual cue reminding parents of their children’s nutritional needs.17 18
Even if parents are fully aware of their children’s nutritional needs, providing children with all of the needed macro and micronutrients can constitute a major challenge for households in LMICs.19 Small-quantity lipid-based nutrient supplements (SQ-LNS) offer a potential solution to this challenge.20 Recent evidence based on randomised trials involving more than 37 000 children suggests that SQ-LNS can reduce the prevalence of stunting21 22 and wasting.23 SQ-LNS has also been found to improve child development,24 reduce the prevalence of anaemia and micronutrient deficiencies.25 Furthermore, programmatic data suggest that SQ-LNS interventions have high rates of acceptability and compliance with regard to SQ-LNS consumption.26 27
To assess the potential impact and scalability of growth charts and systematic distribution of SQ-LNS in Zambia, we conducted a cluster randomised four-arm parallel trial (control; growth charts only; SQ-LNS only; growth charts+SQ LNS) across three purposely selected districts in Zambia. Few studies have assessed the distribution of SQ-LNS delivered through existing community-based programmes,28–33 most of which integrated SQ-LNS with intensive IYCF social and behaviour change interventions,27–29 32 33 treatment for acute malnutrition,30 31 early childhood development32 or WASH interventions.28 29 34 35 Integrated programmes likely contribute to the effectiveness of SQ-LNS by potentially increasing compliance or by removing other constraints on children’s nutritional status (ie, reducing exposure to infections or increasing caregiver resources) but also generate substantial programmatic cost.
We, thus, tested a relatively inexpensive model of direct distribution of SQ-LNS to families accompanied by standardised messaging on how to integrate SQ-LNS into the child’s diet, either alone or in combination with home installed growth charts that offered additional messages on optimal IYCF and nutrition behaviours.
Methods
Study design
This study was designed as a four-arm parallel cluster randomised controlled trial with a 1:1:1:1 allocation ratio.
Participants
Participants in the trial were caregiver–child dyads (CCDs) with a child 2–10 months old at baseline and at least 6 months at the beginning of the intervention period. Using a full list of enumeration areas from the 2010 Zambia census, 282 clusters were randomly selected for this study: 95, 95 and 93 in Lusaka, Choma and Mansa, respectively. A trained study enumerator visited all households in each cluster to assess for eligibility. All CCDs in this age range were invited to participate in the study, and on their written consent, enrolled in the study.
Study setting
The study was conducted in three districts purposely selected by a project advisory board comprising members of the Ministry of Health as well as the Scaling Up Nutrition network. Choma district is located in Southern province, about 300 km South of Lusaka and constitutes a mix of rural, periurban and urban settlements. Lusaka district covers the urban population of the capital, with an estimated population of 3 million. Mansa is located in Luapula district and is among the poorest and most rural districts of Zambia. In 2018, 13% of children aged 6–23 months of age were fed a minimum acceptable diet in Zambia, and 58% of children aged 6–59 month of age were anaemic at the national level.36 Stunting rates among children under age 5 are high in Zambia, with estimated stunting prevalences of 45%, 36% and 29% in Luapula, Lusaka and Southern provinces in 2018, respectively.36
Cluster randomisation and masking
Randomisation was stratified by district. Within each district, clusters were allocated with equal probability to the four study arms. The random sequence was generated using Stata V.16.0 SE37 after completion of the baseline survey. Assessors and analysts were blinded to the treatment. Given the nature of the interventions, blinding of participating families was not possible.
Procedures and interventions
Growth charts: Growth charts are posters designed to be placed on the walls inside children’s homes. The charts used in this trial are shown in online supplemental figure SF1 and were originally based on those used in a previous trial in Zambia,16 with a few modifications. The growth charts contain two areas: one section promoting key messages on IYCF and nutrition and second section that allows parents to monitor their children’s linear growth. On installation, parents were asked to select a career goal for their child in the form of a sticker (eg, ‘I would like my daughter to become a doctor’) and make a personal commitment to child nutrition by choosing between two stickers (‘I commit to feeding my child leafy vegetables and animal foods like kapenta or eggs at least three times per week so that they grow up successful’ or ‘I commit to feeding my child nutritious meals using recommended recipes at least three times per week so that they are well-nourished and healthy’). The charts emphasise key health and nutritional messages such as the importance of continued breastfeeding and handwashing, and feature recipes comprised diverse, nutrient-rich locally available foods for three age groups: 6–8 months, 9–11 months and 12–23 months. The necessary consistency of food is noted for each age group, allowing households to use the recipes across age groups. An image of a woman breastfeeding is included in each recipe section as another reminder to continue breastfeeding until the child is 24 months of age. The messages and recipes are in alignment with the Zambian Ministry of Health and National Food and Nutrition Commission’s recommended practices and messages. Posters also contained an illustrated search and find game at the bottom geared towards young children and designed to stimulate interest in dietary diversity while encouraging children to practice early numeracy and receptive language skills.
Supplemental material
At the centre of the growth chart was a height (linear growth) measurement area that allowed parents to directly compare the height of their children to the WHO reference growth standards38 (online supplemental file 1). For each age, we colour-coded height ranges into green (HAZ >−1), yellow (−1>HAZ ≥−2) and red (HAZ <−2) zones. After the posters were installed at children’s homes, parents were instructed by study staff to find the measurement point corresponding to their child’s age in months and to then assess whether children were within or outside of the international norms. Unlike the poster in the previous trial, this poster extended to the ground, which reduces the risk of measurement error, if after installation, a household decides to remove and rehang the poster somewhere else.
Small-quantity lipid-based nutrient supplements (SQ-LNS): We used the Nutributter+ SQ-LNS produced by Nutriset (online supplemental figure SF2). Each SQ-LNS sachet contains a 20 g nutrient supplement comprised peanuts, sugar, vegetable fat, skimmed-milk powder, and vitamin and mineral fortificants The sachet provides energy (~110 calories), protein, essential fatty acids and a wide range of micronutrients listed in online supplemental figure SF3.
At the launch of the intervention, SQ-LNS households were told that they would receive 30 sachets at the beginning of each month and that children aged 6–23 months should consume one sachet each day. Study staff visited all study clusters once per month and directly delivered 30 SQ-LNS sachets to households participating in the SQ-LNS arm. Households could either pick up the sachets at a prespecified central location in each cluster or get the sachets delivered at their home. The distribution script can be found in online supplemental figure SF4. In terms of logistics, it took on average five work days for a team of 4 with a vehicle to reach all the clusters in each district. The average time spent in its cluster varied substantially depending on the number of participants in each cluster and the support by local communities and community health workers (CHWs) provided.
Deviations from protocol
The study was originally designed to include 20 randomly selected families in each of 120 randomly selected enumeration areas. After initiation of fieldwork, it became clear that the number of eligible children was much smaller than anticipated due to smaller population sizes and declining fertility rates. We, thus, increased the number of clusters to 282 and enrolled all eligible children in these clusters into the trial.
Patient and public involvement
Several rounds of pretesting were conducted to calibrate growth charts to the local population; we also piloted the supplement distribution to ensure the supplements were well-liked by families and formally confirmed product safety through laboratory tests. Results of the trial will be shared with the public through a public dissemination event.
Study timeline
Baseline interviews were conducted between 26 April 2021 and 15 July 2021. A small pilot was conducted in July and August to test intervention logistics. Random cluster-level group assignment was completed in late August 2021. Growth charts were installed in September and October 2021. SQ-LNS distributions were implemented between November 2021 (when the youngest child reached age 6 months) and May 2023 (end of study). Endline surveys were conducted between 27 June 2023 and 13 September 2023.
Outcomes
The primary study outcomes were height-for-age z-scores (HAZ) and stunting (HAZ <−2). Height and weight were measured by trained interviewers at both baseline and endline at children’s homes. Length at baseline was measured in recumbent position using a length board. Height at endline was measured in standing position using Seca 217 stadiometers. Weight was measured using Seca 874 scales. Height and weight were measured at least twice by the same assessor and measurements were repeated if the measurement difference exceeded 100 g for weight and 7 mm for length/height. Z-scores were computed using the WHO anthro software package.39
The secondary study outcomes were weight-for-height z-scores (WHZ), weight-for-age z-scores (WAZ), underweight, Hb, anaemia and child development. Haemoglobin levels were assessed by trained study staff using a drop of blood obtained through a finger prick directly analysed at children’s homes using Hemocue 801 devices (Hemocue, Sweden). Anaemia was defined as haemoglobin level below 110 g/L. Child development was assessed using the long form of the Global Scales for Early Development (GSED)—we carefully trained our study staff on the application of this tool using the official GSED training materials.40 Dietary diversity was defined as the number of defined food groups (out of eight) consumed by the child during the previous day (elicited from caregivers during the endline survey) as suggested by the latest WHO/UNICEF guidelines.41 Weight was measured both at baseline and endline. Hb and child development (GSED) were only measured at endline.
Sample size
Based on the results of the previous growth chart trial in Zambia,16 we targeted a 0.25 SD difference in HAZ after 18 months of intervention. The study was powered to detect this 0.25 HAZ difference between any of the three intervention arms and the control arm with probability 0.9, assuming α=0.05, 15% attrition between baseline and endline and a design effect of 2. The original power calculations were made based on an intra-class correlation (ICC) of 0.05 and 20 child–caregiver dyads per cluster, and 40 clusters per group.
Statistical methods
We first provide descriptive statistics by study arm using data collected at baseline and then use multivariable regression models to estimate programme impact. Given our primary focus on child growth after 6 months of age, all models controlled for length-for-age z-scores at baseline (age 6–11 months); we also included district-specific intercepts in all models to account for the stratified (blocked) randomisation. To ensure robustness of our results and to increase precision in the linear models, we also estimated adjusted models controlling for child age, sex, twin status, caregiver age, caregiver education, household size and household asset ownership.
For all continuous outcomes, we used standard linear models with cluster-robust SEs to account for the cluster-level random treatment assignment. For all binary models, we used logistic regression models. Given the high stunting incidence in this sample, we also estimated prevalence ratios for this outcome as suggested in the literature.42 43 All primary models were estimated intention to treat. In adjusted models, we controlled for baseline HAZ, child age and sex, as well as caregiver education and household wealth as most commonly used predictors of HAZ. We also estimated per-protocol models restricted to children complying with the two interventions. Compliance with posters was defined as still having a poster hanging at endline. Compliance with supplementation was defined as caregivers (self-) reporting SQ-LNS administration for at least 12 months to children during the endline survey. We also estimated a series of exploratory models (that were not prespecified) to provide some suggestive evidence regarding the causal pathways underlying the observed treatment effects. For nutritional outcomes, we analysed breastfeeding duration as well as dietary diversity scores.44 For child development, we analyse home stimulation provided by caregivers and other adults in the household members.45
All analyses were conducted by using the Stata V.16.0 SE software package.37
The trial was registered at clinicaltrials.gov as NCT0512042 after completion of the baseline survey on 2 November 2021.
Results
The trial began on 26 April 2021 and was completed on 13 September 2023. A total of 2291 CCDs across 282 clusters were included in the study (figure 1). After completion of baseline, 70 clusters (557 CCDs) were randomised to the control group, 70 clusters (643 CCDs to growth charts only, 71 clusters (525 CCDs) to SQ-LNS only and 71 clusters (566 CCDs) to growth charts plus SQ-LNS.
CONSORT flow chart. CONSORT, Consolidated Standards of Reporting Trials; HAZ, height-for-age z-score; SQ-LNS, small-quantity lipid-based nutrient supplements.
All participants received the intended treatments initially. At endline, posters were no longer hanging in 172 homes (31.7%) in the growth charts arm, and in 131 homes (27%) in the combined intervention arm. 92 CCDs (20.5%) in the SQ-LNS only arm, and 133 CCDs in the combined intervention arm (27.4%) received less than 12 monthly doses of supplementation. At endline, 1950 CCDs (85%) were successfully assessed (n=473, 543, 448 and 486 in arms 1, 2, 3 and 4, respectively). The primary reason for loss to follow-up was the inability to locate caregivers; anthropometric data were missing or invalid for 35 cases.
Table 1 shows descriptive statistics by study arm. Children were on average 6 months old during the baseline interview; 50% of children were female and 4% were twins. At baseline, children’s mean HAZ was −0.91. Caregivers were on average 27 years old and with the exception of five caregivers female (99.7%); 53% of caregivers had some secondary education and 9.5% had attended tertiary education. 75% of caregivers were married. On average, households comprised on average approximately six household members. Baseline characteristics were highly similar across study arms.
Participant characteristics at enrolment
A total of 1884 children were analysed. On average, children were 30.7 months old at endline. Table 2 shows the estimated impacts on primary outcomes. Growth charts slightly reduced average HAZ and increased stunting rates, but the estimated differences were not statistically significant. SQ-LNS improved HAZ by 0.21 SD (95% CI 0.06 to 0.36) and reduced the odds of stunting by 37% in fully adjusted models (adjusted OR, aOR 0.63, 95% CIs (0.46 to 0.87)). The combined intervention did not have an impact on the primary outcomes.
Impact on primary outcomes
Table 3 shows estimated impacts on secondary outcomes. Growth charts alone did not have an impact on WHZ and WAZ and on being underweight but marginally improved average haemoglobin levels by 2.4 g/L (−0.1, 4.9) and marginally reduced the odds of anaemia by 25% (aOR 0.75, 95% CI (0.56 to 1.02)). Growth charts alone also improved average child GSED scores by 0.19 SD (95% CI 0.02 to 0.36). SQ-LNS alone improved WAZ (mean difference, MD 0.17, 95% CI (0.05 to 0.28), haemoglobin levels (MD 2.9, 95% CI (0.2 to 5.5)) and GSED scores (MD 0.28, 95% CI (0.10 to 0.46) and reduced the prevalence of anaemia (aOR 0.74, 95% CI (0.54 to 1.01)) and underweight (aOR 0.57, 95% CI (0.35 to 0.91)) but did not improve WHZ (which was, on average, positive in this sample). The combined intervention did not have an impact on WHZ, WAZ, haemoglobin or GSED scores but reduced the odds of anaemia (aOR 0.72, 95% CI (0.53 to 0.97)) and underweight (aOR 0.70, 95% CI (0.47 to 1.04)).
Impact on secondary outcome variables
Online supplemental figure SF5 summarises the estimated treatment effects on primary and secondary study outcomes.
Online supplemental table ST1 shows the results of the per-protocol analysis, where we restrict the treated groups to homes where the growth charts were still hanging at endline and where caregivers report at least 12 months of supplementation. Overall, the results look very similar to the main ITT estimates displayed in tables 2 and 3. Online supplemental table ST2 shows an alternative specification directly testing for intervention interactions. On average, interaction effects appear small and insignificant with the exception for GSED scores where interaction effects appear to be negative and significant.
Table 4 shows the results of our exploratory pathway analysis. For food diversity, small improvements were seen for the combined intervention group—these effects seem to stem from increased consumption of nuts and legumes as well as increased consumption of vitamin A rich fruits and vegetables. For home stimulation—activities initiated by adult caregivers to support children’s development—significant improvements were seen only for the growth charts group. None of the interventions appear to have affected breastfeeding duration. All three intervention arms were associated with improved child morbidity reports, with estimated reductions in the odds of recent illness of up to 41% for the combined intervention and having a cough in the past 2 weeks (aoR 0.59, 95% CI (0.44 to 0.78)).
Exploratory pathway analysis results
Online supplemental table ST3 shows stratified results by baseline stunting, age and gender on HAZ. Growth charts alone did not show any impact in any subsample. The SQ-LNS intervention appears to have been particularly impactful among children stunted at baseline and among younger children. We found no sex differences in intervention impact.
No study-related deaths or injures were reported by study households during the fieldwork procedures.
Discussion
In this study, we aimed to assess the feasibility and impact of home-based distribution of growth charts and SQ-LNS to families with young children across a sample of 282 enumeration areas randomly selected in three diverse districts of Zambia. The trial has yielded three main insights. First, the direct distribution of SQ-LNS to families can lead to substantial improvements in children’s nutritional status and development. We see positive impacts across all outcomes analysed, with a 0.21 SD increase in HAZ, a 37% decrease in the odds of stunting, a 26% decrease in the odds of anaemia and a 0.28 SD improvement in children’s development. Second, home installed growth charts do not seem to be an effective tool to reduce linear growth faltering but appear to have positive impacts on children’s diet and development, with a 25% reduction in the odds of anaemia, and a 0.18 SD increase in children’s development. Third, combining the two interventions does not seem to work well in practice, with combined impacts estimated smaller than the SQ-LNS alone impacts for all primary and secondary outcomes analysed other than anaemia, where the combined intervention marginally outperformed SQ-LNS alone.
The positive results of SQ-LNS are not surprising and are well aligned with the recent literature on SQ-LNS supplementation.21 23 We find that home distribution of SQ-LNS, even without intensive IYCF counselling, had large positive effects on children, in line with the effects observed in interventions that distributed SQ-LNS as part of integrated interventions.28 29 This suggests that a stand-alone distribution of SQ-LNS may be as effective as an integration of SQ-LNS into other programmes, such as IYCF platforms, that often encounter major logistical, political and financial challenges. On average, the total distribution cost for the project was US$87 per child over the 18 months period, including product cost, shipping, taxes, domestic transport and distribution. This cost could likely be reduced substantially with larger-scale rollouts, especially if existing supply chains could be used.
The lack of impact of growth charts on child growth was more surprising, as this study was partially set up as a larger-scale replication study of a previous pilot study in Zambia with very positive results among stunted and younger children.16 Relative to the previous studies, the growth charts used in this study were not only larger but also substantially more complex, providing both nutritional and aspirational messages to parents and games for the children. The nutrition messages on the growth charts appear to have worked at least partially: we observed higher average haemoglobin levels, a lower prevalence of anaemia and improvements in child development in the growth charts only group. The positive impact on haemoglobin suggests that parents may have been able to increase their children’s consumption of micronutrient-rich foods; the lack of impact on growth, however, suggests that parents were not able to sufficiently increase total caloric intake and to address other factors in the community that may have constrained child growth. The positive impact on child development is potentially attributable to increased caregiver–child interactions (seen in table 4), which have been shown to be highly predictive of children’s developmental outcomes in previous literature.45–47
Given their low cost, growth charts may be a cost-effective and scalable intervention to inform caregivers and potentially also to change their behaviour. However, more work is needed to better understand how ‘light touch’ behavioural interventions like the growth charts used in this study can be used to influence human behaviour in general and caregiving practices in particular; a recent study from Ghana suggests that animated videos and calendar reminders encouraging mothers to speak to their children can lead to substantial improvements in child development.48
While we hypothesised the two interventions would generate positive synergies, this did not happen empirically. While the HAZ and WAZ effects of the two interventions roughly add up in terms of their combined effect, we find negative synergies for GSED and haemoglobin (online supplemental table ST2), with positive impacts for each intervention separately, but no impact for the combined intervention. It seems possible in principle that participating in two interventions may have lowered parental efforts to comply with either intervention—particularly also because supplements were not mentioned on the growth charts explicitly. We see, for example, that parents receiving supplements are more likely to mix them in with other foods in the combined arm than in the SQ-LNS only arm. Mixing SQ-LNS with other foods could affect nutrient absorption, and in particular, lower the dose of nutrients provided if children do not finish their plate with the mixed-in SQ-LNS. We also see that the increased caregiver–child interactions seen in the growth chart arm are less pronounced in the combined arm than in the growth charts only arm (table 4), suggesting that parents engaged less with the posters when they also received supplements. It also seems possible that some of the food types and dishes promoted by the growth charts could have interfered with the child’s appetite or absorption of the supplements.
The trial presented has some limitations. First, we only tested one type of supplement (Nutributter+) in this trial and thus are only able to assess the relative importance of this specific product or specific micronutrient contained in this product for the positive outcomes seen. Second, we are not able to fully understand the dynamics that led to the weak impacts in the combined arm. Even though we conducted complementary 24-hour in-depth food intake monitoring among a subsample of children, we are not able to directly identify total food consumption or nutrient absorption. Our results for the growth charts can also not be interpreted as a direct failure to replicate the previous pilot study results for stunted children because we changed the design of the poster rather substantially and may thus have changed its impact despite the fact that many central design features such as the measurement space have stayed the same. While we tried to do endline surveys shortly after the end of supplementation, there was an average gap of approximately 1 month between the last supplement intake and the endline assessment, which likely lowered the measured impact on haemoglobin.
Our results in the combined treatment group highlight the fact that social and behavioural change communication that accompanies SQ-LNS is critical to the product’s effectiveness.21 We designed the growth charts as a stand-alone intervention and did not tailor the messaging to reinforce how to integrate SQ-LNS into optimal infant feeding practices in the combined intervention. This could have caused cognitive dissonance among caregivers and even ‘competition’ between SQ-LNS and some of the other behaviours promoted in the poster (ie, recipes and messages promoting specific foods and dietary diversity). The lack of integration of the two interventions may have ultimately diminished the effectiveness of both interventions.
In terms of generalisability, the three study districts reflect the general population of Zambia including urban and rural populations. Given the similarity in terms of living conditions and nutritional habits with neighbouring countries in East Africa, similar impacts in other countries of this region seem plausible. Further research will be needed to assess impacts in other settings.
In terms of national policy, stunting clearly remains a priority area for government programmes. The stunting rates seen in this trial are higher than those seen in the latest Demographic and Health Survey36 and thus highlight the urgent need for additional interventions.
Overall, the study suggests that the home-based distribution of SQ-LNS to families with young children in areas with high levels of food insecurity can yield substantial improvements in children’s physical growth and development, and that growth charts may also be an effective tool to induce some behavioural change. At the same time, this trial clearly also shows that basic supplementation and growth charts are not sufficient to address the high rates of child malnutrition in these areas—more research as well as more comprehensive interventions are urgently needed to address the continued high burden of chronic malnutrition in Zambia as well as other similarly resource-constrained settings.
Supplemental material
Data availability statement
Data are available on reasonable request. Data will be available on request to the first author.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the University of Zambia Biomedical Research Ethics Committee as protocol 1411-2020 as well as by the Ethnics Committee for Northwestern Switzerland (EKNZ) as AO_2021_00016. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We would like to thank the IPA Zambia team for all of their hard work to make this study happen. We would also like to thank all families participating in the study for their time commitment and support for this research.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
DP and PCR are joint senior authors.
Handling editor Seema Biswas
Contributors GF, PCR and DP conceptualised the work, acquired the funding and led the overall trial. DS and TB-M coordinated the trial with the Zambian Ministry of Health and provided input and feedback for the overall trial design. SH, MC, TB-M and DP led the development and refinement of the growth charts. LML and JL provided guidance on all nutritional aspects and input to the trial design and analysis. PCR conducted the initial blinded analysis. GF created all figures and tables and a first draft of the paper. All authors reviewed multiple versions of the manuscript and approved the final submission. GF is the guarantor.
Funding This project was funded by Sint Antonius Stichting.
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.