Introduction

Stunting (low height- or length-for-age) remains a major public health problem with about 149 million children under five years affected globally; 39% of these children are from Africa [1]. Stunting in early life is associated with long-term challenges including increased morbidity, mortality, chronic diseases and infections, inadequate physical functioning, low economic productivity, and poor cognition and neurodevelopment [2–4]. Children with persistent childhood stunting experience long-term poor developmental outcomes [5, 6]. Small size at birth is one of the leading contributors to stunting globally [7]. In Rwanda, stunting rates in under-five children are higher compared to global rates, 38% [8] versus 25% [1]. The situation is even worse for children born small or who become sick in the first weeks of life. “Small and sick” children (defined by the WHO as a child weighing <2500g at birth or a newborn with any medical or surgical condition [9]) have nearly double the rates of stunting compared to national averages [8, 10]. About 70% of the burden of stunting by 59 months of age is rooted from the first 24 months of life [11] and the rates of stunting increase as children age, peaking at 23 months; after this age the rates of newly stunted children tend to decrease [12].

The first 1,000 days of life, from conception to two years, is a critical period for child development when the brain is most rapidly developing and is highly sensitive to program interventions for improving child growth and developmental outcomes [7, 13]. A study of countries that have had outsized success at reducing stunting relative to economic growth has identified several interventions, particularly during pregnancy, that have contributed to stunting declines [14], but less is understood about successful interventions in early childhood. There is evidence that with sustained nutritional interventions throughout childhood until adolescence [15] infants who are stunted can have growth catch-up and recover from stunting, though recovery is thought to be most likely before age two [16]). Studies in Kenya, Brazil and South Africa have found varying rates of stunting recovery ranging from 19% to 45% [17–19]. In the South African study, factors associated with recovery from stunting included early child growth failure, birth spacing, monitoring and supporting growth, education status of the mother, birth weight of the child and household socioeconomic factors [19]. Though factors associated with recovery from stunting among general child populations have been identified [17–19], little is known about recovery from stunting among high-risk children in low- and middle-income countries.

In 2014, Partners In Health/Inshuti Mu Buzima (PIH/IMB), in partnership with the Rwanda Ministry of Health and United Nations Children’s Fund (UNICEF), initiated Pediatric Development Clinics (PDC) as a new integrated model of outpatient care at public district hospitals and health centers in rural settings to help all infants who are born small and sick survive, thrive and reach their full developmental potential. Through structured follow-up care and support, children in PDC receive integrated prevention, early detection and support interventions to improve their nutrition, health and development [20, 21]. This study aimed to examine, among children enrolled in PDCs in Rural Rwanda and stunted at 11 months, what proportion recovered from stunting at 24 months of age and which factors were associated with their recovery.

Methods

Results

In total, of 403 infants and young children who had documented stunting at 11 months, 179 met further eligibility criteria and were included in this study (Fig 1). Except the fact that ineligible kids were significantly more likely from PDCs at HCs, there was no other statistically significant differences existed between the included and excluded participants on baseline demographic or health status variables (S1 Table). Table 1 presents socio-demographic data of the study population.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Study flow chart.

This flow diagram illustrates the process for selection and enrollment of study participants.

https://doi.org/10.1371/journal.pone.0283504.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Socio-demographics of children stunted at 11 months and their mothers enrolled in PDC, N = 179 unless otherwise indicated.

https://doi.org/10.1371/journal.pone.0283504.t001

Implementation of PDC was initially started in Kayonza district then expanded in Kirehe district, thus of the 179 infants and young children, the majority were from Kayonza district (73.7%). Our study population consisted of more males (60.9%), people with low household resources (Ubudehe category 1 or 2 (62.6%). At the time of enrollment in PDC, 7.8% of children did not have any health insurance and 10.6% of children were cared for by one parent (mothers who are single, divorced or widowed). The median number of years of maternal education was 5 (IQR: 0, 10).

Table 2 presents clinical characteristics of the children in the study. Among the children enrolled in PDC with stunting at 11 months, most (77.6%) had a history of prematurity or LBW; smaller numbers of children were referred to PDC with other conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Clinical characteristics of children stunted at 11 months and their mothers enrolled in PDC, N = 179 unless otherwise indicated.

https://doi.org/10.1371/journal.pone.0283504.t002

Most children had a history of malnutrition at 6 months of age: (21.2%) infants had wasting, 67.0% were underweight and 73.7% were stunted. At 11 months of age, the percentages of infants with wasting and underweight was similar or slightly increased compared to 6 months. The median number of PDC visits between 11 and 24 months was 5 visits (IQR: 4,7).

As presented in Table 3, 37 (20.7%) recovered from stunting at 24 months of age, while 21 (21.0%) children with severe stunting improved to moderate stunting. However, 20 (25.3%) worsened from moderate to severe stunting.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Stunting dynamics at 24 months of children enrolled in PDC in Eastern Province, Rwanda between 2014 and 2018.

N = 179.

https://doi.org/10.1371/journal.pone.0283504.t003

In the bivariate analysis, normal LAZ at 6 months of age, and normal WAZ at 6 and 11 months of age positively associated with recovery from stunting, while, having no health insurance and increasing of maternal age were negatively associated with recovery (S2 Table). Lack of health insurance was completely predictive of failure to recover; no children without health insurance recovered from stunting.

In the final logistic regression model, stunting recovery at 24 months was significantly associated with stunting status at 6 months of age and severity of stunting at age 11 months when controlling for the sex of the child (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Factors associated with stunting recovery at 24 months of age for PDC children enrolled in PDC children who were stunted at 11 months of age.

https://doi.org/10.1371/journal.pone.0283504.t004

Early stunting at 6 months of age was associated with lower odds of stunting recovery, with the odds of stunting recovery being reduced by 80% (aOR: 0.2; 95%CI: 0.07–0.81) for severely stunted children and by 60% (aOR: 0.4; 95% CI: 0.16–0.97) for moderately stunted children compared to children with normal height-for-age at 6 months of age (p = 0.035). Lower odds of stunting recovery were observed among children who were severely stunted at 11 months of age (aOR: 0.3; 95%CI: 0.1–0.6) compared to these who were moderately stunted (p = 0.004).

Discussion

Nearly a quarter (21%) of children in this study who were stunted at 11 months recovered from stunting by age 24 months and factors associated with failure to recover from stunting included early stunting at 6 months of age and greater severity of stunting at age 11 months. This is perhaps the first study in Rwanda evaluating stunting recovery among children born small and sick.

The rates of recovery from stunting in our study are considerable given that the observed changes took place over the course of 12 months. While more than half of children were severely stunted at 11 months of age, recovery occurred in the critical 1,000 days’ window when the brain is most rapidly developing. The existing literature on stunting recovery reports variable rates of recovery from stunting depending on the age ranges of children at the time stunting recovery is assessed [18, 28]. A longitudinal study of children from four low- and middle-income countries (Ethiopia, India, Peru, and Vietnam) found stunting recovery rates of 27% to 53% and 30% to 47% at 5 and 8 years old, respectively [28]. Findings from other studies reported stunting recovery rates of 24% in Brazil [18], 35% in India [29] and 19% in South Africa [19]. Our results should be compared with caution to rates from these other studies, however. Most evaluated catch-up growth status of older children outside the 1,000 days’ window of critical brain development, had relatively longer timeframes for recovery, and were not conducted among similarly high-risk children [4, 30], Other studies included a general population of children [18, 19, 28, 29] and had no specific interventions to improve catch-up.

While interventions for catch-up growth are important, recovery from stunting may not be sufficient to undo the early negative impact on the brain occurring when children experience chronic undernutrition and poor growth in early life [31, 32]. Improving maternal nutrition and early integrated nutritional and caregiving interventions, including all components of nurturing care for children born small and sick, are crucial for accelerating growth catch-up and improving developmental outcomes [33]. The stunting recovery rates in our study point to the positive impact of the PDC’s integrated interventions on nutritional status, and provide important learning on potential strategies that can work to achieve the global nutrition targets for reduction of stunting by 40% by 2025 [34]. Further study could help to understand potential associated benefits for children’s development.

In our study, children who were severely stunted at 11 months were much less likely to recover completely at 24 months than their peers with moderate stunting. This finding is consistent with previous studies which reported lower likelihood of recovery from stunting among children with severe stunting [29, 35]. However, in our study, at least 29% of severely stunted children improved from severe to moderate stunting in 12 months while in the PDC program and longer evaluation could help to understand their longer-term growth.

Similarly to findings from the South African study [19], children with early onset of stunting in the first 6 months of life were significantly less likely to recover from stunting. The first 6 months of life is a critical period with fast growth velocity [36]; children exhibiting early linear growth deficiencies may experience slow growth catch-up taking longer for them to recover and may develop persistent stunting in late childhood. The PDC program includes specific early interventions focused on prevention of stunting in the first 6 months of life. In 2017, these interventions were strengthened to improve the quality of nutritional assessment, the PDC clinical protocol and tools were revised with full integration of a tool for management of small and nutritionally at-risk infants under six months and their mothers (MAMI care pathway package) [37]. In addition, decision-making support algorithms were added to provide guidance to PDC providers on appropriate systematic screening, early identification, and management of underweight, wasting and feeding issues for small and nutritionally at risk infants and their mothers [20]. Based on the timing of our study, not all children followed benefited from these revised procedures and future studies could investigate whether these early interventions reduce stunting at 6 months and contribute to overall greater stunting recovery in the PDC.

In our study, only prior stunting status was statistically significantly associated with recovery in the final model. Other factors such as maternal age, and underweight status were not independently associated with recovery when prior stunting is in the model. In our study, the maternal age in those who recovered was older, but this difference was not significant, while in a study from Kenya found that children from younger mothers were less likely to recover from stunting [17]. This discrepancy may, in part, exist because the age distribution in our study was different than the Kenyan study, which had both larger study population and a higher proportion of adolescent mothers. Insurance status has been shown in other studies to be an important predictor of care seeking [38–40] and associated with lower odds of stunting [41], and in our study lack of insurance completely predicted the outcome, resulting in complete separation of the model, so was not included in the final model. This finding is slightly surprising, as PDC care is free of charge, mothers in lower Ubudehe categories are given travel subsidies as part of social support provided to them and a high proportion (74%) of the uninsured were in low Ubudehe categories. This suggests that PDC services, while important and targeted to remove financial barriers to participation, are not entirely a replacement for other needed health care that may be lacking.

Despite the high percentage of children whose stunting status improved, a quarter of moderately-stunted children worsened to severe stunting, despite PDC interventions. Similar probabilities of progressing from moderate to severe stunting have been reported from a longitudinal study conducted in Ethiopia [42]. Nevertheless, our finding demonstrates how it can be difficult to reverse stunting given the multi-sectoral factors that contribute to stunting [43, 44] and challenges to achieve full stunting recovery among infants and young children enrolled in PDC in just a 12-month period. Our findings also imply a need for further qualitative research to understand problems faced by caregivers and PDC providers in supporting children with poor growth and solutions to strengthen the program interventions as well as other types of support to caregivers. It is also possible that observed worsening and challenges to achieve stunting recovery among children in this study could be related to other co-morbidities and challenges. For example, feeding difficulties are common among children with developmental disabilities and tend to worsen without early intervention. Feeding difficulties among children with disabilities are associated with high rates of malnutrition and early life mortality [45]. Previous studies conducted in Rwanda among children born small and sick reported high rates of potential disabilities at ages 2 to 4 years [46] and higher rates of malnutrition among children with feeding difficulties [47], thus additional specialized nutrition interventions are likely required to address these children’s needs. Based on this challenge, in 2019, the PDC program integrated early intervention using the Baby Ubuntu model [48], a group participatory early intervention program starting from 6 months for young children with disability and their families. The 10-module, participatory, parents’ peer support group program includes a module designed to provide caregivers with skills to appropriately feed a child with a disability and manage their feeding difficulties at home [49, 50].

Our study had certain limitations to note. First, as our study used data that were routinely collected from the EMR there were some variables of interest that were not available, such as information regarding the home environment including water, sanitation and hygiene; food security; child care practices at home; and female decision-making in households. However, some important socio-demographic variables were available in the EMR which we included, such as household socioeconomic status (Ubudehe category). Second, as these were routinely collected programmatic data, missing data was an issue for many different variables. Our sensitivity analyses for systematic differences between children who were eligible at 11 months and included versus those who were excluded because of lack of data at 24 months found no statistically significant differences in any sociodemographic or baseline health factors, so we do not believe that this exclusion introduced important biases. For our outcomes measures, we used z scores with cutoff points as determined by WHO and international studies; using a continuous z score would have allowed for more nuance in assessment, as the clinical difference between (for example) -2.01 and -1.99 is minimal. However, the utilization of cutoffs is the basis for provision of certain services and additionally allows us to compare to other surveys and population. Also, our study participants were children enrolled in PDC. thus, generalizability of findings to other high-risk infants who do not receive similar interventions is limited. Odds ratios may not be identical to risk in a population with common outcomes, so in our study we present odds exclusively. Finally, our study looked at change over only 12 months, which may have been too short of a timespan to detect important changes. However, since this is among the first studies conducted on a programmatic intervention targeting vulnerable children in a rural African setting, the results remain useful for designing interventions for this special population. Despite these limitations, this study is contributing to the understanding of the effects of an integrated program and early interventions on growth-catch-up among high-risk infants, and the extent of stunting recovery possible with supportive programming among infants born with perinatal complications at risk of malnutrition and developmental challenges.

Conclusion

Our study findings generate new evidence about the recovery from stunting among at infants and young children enrolled in the PDC in rural Rwandan settings. A substantial proportion of children enrolled in the PDC that were stunted at 11 months of age recovered from stunting at 24 months of age despite their history of perinatal complications. Early stunting at 6 months of age and severity of stunting at 11 months of age were negatively associated with stunting recovery at 24 months of age. Our results highlight the importance of integrated interventions to prevent undernutrition during first two years of life and observed stunting recovery rates provide promising effects of integrated and targeted nurturing care interventions to avert stunting during infancy and early childhood periods among at risk infants with history of perinatal complications. Further research is needed to explore association of environmental factors with stunting recovery at 24 months of age and in late childhood among high-risk children enrolled in post-neonatal follow-up programs. The effects of sociodemographic factors of both parents including parental stature need also to be included in future studies. In addition, other studies including controls groups could help to understand the effectiveness of PDC interventions on catch-up growth and provide more evidence to guide improvements as well as influencing policies.

Supporting information

Acknowledgments

We acknowledge Partners In Health for the support of this work. This study was developed under the Partners In Health/Inshuti Mu Buzima Intermediate Operational Research Training Program, developed and facilitated by Dale A. Barnhart, Bethany Hedt-Gauthier and Ann C. Miller. Ann C. Miller and Alphonse Nshimyiryo provided direct mentorship to this paper as part of this training. We thank Kirehe and Rwinkwavu district hospitals and their health centers for their support during this study. We also acknowledge the contributions of PDC nurses, social workers and EMR data officers for data collection and for their daily efforts to improve the lives of infants and young children in rural communities. We are grateful to the PDC children and their caregivers in the study.