Adequate nutrition in early life is essential for optimal health and development throughout the life course. Dietary deficiencies in macro- and micronutrients in early childhood can lead to poor physical growth, suboptimal neurodevelopment and increased risk of morbidity and mortality throughout the life course(1–Reference Martins, Toledo Florêncio and Grillo5). An estimated 144 million children under 5 years of age are stunted (low height/length-for-age), and 47 million suffer from wasting (low weight-for-height/length)(1). In addition, 11 % of school-age children (5–19 years) globally are estimated to be too thin for their age (BMI < −2 sd of the WHO reference population), with a prevalence of up to 25 % in South Asia(6). The magnitude of the global burden of malnutrition and the long-term consequences of poor nutrition early in life have propelled the global health community to invest substantial political, financial and human resources to address these challenges.
Child undernutrition, in particular, has garnered tremendous political momentum as a global public health problem in the last two decades. Starting in the year 2000, the global community set out to achieve the eight Millennium Development Goals (MDG) by 2015(7), including the goal to ‘Eradicate extreme poverty and hunger’. In 2010, the Scaling Up Nutrition movement was born, led by dozens of country governments, with a vision to ensure ‘a world free from malnutrition in all its forms’ by 2030(8). Shortly after, in 2012, the World Health Assembly set six global nutrition targets, to be met by 2025(9), to reduce several forms of malnutrition including stunting, wasting and anaemia. In 2015, the Sustainable Development Goals (SDG) were adopted by all UN Member States to ensure a sustainable future for the people and the planet(10), with the second SDG specifically aiming to end hunger and all forms of malnutrition by the year 2030. The UN Standing Committee on Nutrition further declared 2016–2025 the Decade of Action on Nutrition(11), identifying six action areas for nutrition, including sustainable food systems, aligned health systems, social protection, trade and investment, enabling environments, and strengthened governance and accountability.
As a result, since the year 2000, there has been greater political commitment at the national level to improve child nutrition across low- and middle-income countries. Most countries have developed national nutrition policies that outline a set of guidelines, rules, regulations and/or laws to provide key public health interventions to improve child nutritional status. Such policies include, for example, the promotion of exclusive breast-feeding and appropriate complementary feeding, fortification of staple foods and provision of micronutrient and food supplements, among others(12).
Our understanding of the number and breadth of nutrition-specific policies and their impact on child nutritional status at the population level, however, is fragmented. Previous studies have evaluated either one specific policy across multiple countries (e.g. a review of the effect of breast-feeding policies and programmes on breast-feeding trends across six countries(Reference Mangasaryan, Martin and Brownlee13)); one micronutrient intervention on one outcome across several countries (e.g. a systematic review of the effect of increasing folic acid intake through fortification on neural tube defects(Reference Heseker, Mason and Selhub14)) or the effect of multiple nutrition policies over time in one country (e.g. a review of nutrition policies in South Africa from 1994 to 2002(Reference Labadarios, Steyn and Mgijima15)). The breadth and scope of nutrition-specific policy research conducted in the MDG era to address child undernutrition is not well-mapped. Identifying the extent of nutrition policy research that specifically addresses child undernutrition will enable a greater understanding of the potential gaps and priority areas for investigation during the remaining years of the SDG. Therefore, we conducted a scoping review to map and describe the evidence base on population-level nutrition-specific policies evaluated in the literature for improving child nutritional status.
Methods
We conducted a scoping review using a systematic search of the literature to map and describe peer-reviewed studies evaluating the impact of population-level nutrition policies to address child undernutrition. Rather than focusing on a specific exposure–outcome association, scoping reviews provide a systematic overview of a broad topic and an understanding of how (and how much) existing research has been conducted on that topic, allowing the assessment and synthesis of existing evidence(Reference Munn, Peters and Stern16–Reference Peters, Marnie and Tricco18). As such, we identified the types and sources of evidence on nutrition-specific policies to address child undernutrition and the potential research gaps in nutrition-policy evaluation literature to ultimately inform further investigation(Reference Daudt, Van Mossel and Scott19). We maintained a broad definition of ‘policy’ to ensure that all relevant studies were captured in the search. This was inclusive of any nutrition-specific policy, strategy, legislation, regulation or guideline applicable at the population level (whether national, state or regional level) that was endorsed by the government. We used the Lancet 2013 Nutrition Series(Reference Black, Victora and Walker2) framework to define ‘nutrition-specific’ as factors that directly contribute to nutritional status, including macro- or micronutrient interventions, breast-feeding, complementary feeding, dietary diversity, feeding behaviours and responsive caregiving, among others. Multisectoral policies or those pertaining to ‘nutrition-sensitive’ domains, which are indirectly associated with child nutritional status, such as agriculture and food security and social safety nets, were included if they also included nutrition-specific components. Policies related to nationally instituted or mandatory nutrition-specific programs (e.g. the Special Supplemental Nutrition Program for Women, Infants, and Children in the USA) were also included; however, any programmes, trials or interventions tested for efficacy in a sample study population were excluded as our focus was on nutrition-specific policy as a population-level intervention.
Five specific interdisciplinary and field-specific electronic databases – Medline, Embase, Web of Science, Scopus and PAIS Index for public policy – were searched using a combination of keywords and MeSH terms for four key concepts: nutrition-specific; policy; children under five or school-age children and child nutritional status outcomes (see online supplementary material, Supplemental Table 1). The search was conducted in September 2017 and was intended to capture research conducted in the MDG period (allowing for a lag of 2 years for publication time). A librarian reviewed the search terms to ensure inclusivity, relevance and robustness of the search.
Titles and abstracts (n 5646) were screened in duplicate by authors B.B., A.R., V.P. and N.P. to select studies appropriate for full-text review (Fig. 1). Inclusion and exclusion criteria for title and abstract screening are summarised in Table 1. Importantly, we did not restrict study eligibility by geographic location or country wealth to be able to scope the global evidence base. In addition, we included studies which evaluated health outcomes among children up to 10 years of age so as to not inadvertently exclude any nutrition-specific policies which may typically be evaluated among school-age children (e.g. policies on iodine deficiencies). In cases where the two reviewers did not agree regarding study eligibility, or it was not clear from the abstract whether the study met the inclusion criteria, the study was included for full-text review. A total of 157 studies were selected for full-text review (reasons for exclusion are summarised in Fig. 1).
Fig. 1 Study selection process
Table 1 Inclusion and exclusion criteria for study inclusion
* Nutrition-specific as defined by Black et al. 2013, including: adolescent health and preconception nutrition, maternal dietary supplementation, micronutrient supplementation and fortification, breast-feeding and complementary feeding, dietary supplementation for children, dietary diversification, feeding behaviour and stimulation, treatment of severe acute malnutrition, disease prevention and management, nutrition interventions in emergencies.
Two reviewers independently abstracted data from each eligible study onto a standardised abstraction template developed by the authors. Any disagreements in data abstraction were resolved by consensus between two reviewers. Key study characteristics were abstracted, including study objectives, location (city and/or region, country and geographic region based on the World Bank classification), county wealth category (World Bank 2017 income levels), study design, sample size, recruitment method and age range of study population, type of nutrition policy, level of policy implementation (sub-national or national), year of policy implementation, year of outcome assessment, child age and types of health outcomes assessed. Quality of reporting/critical appraisal guidelines for the selected research articles was not used since the objective of this study was to assess the breadth of the evidence base as opposed to evaluating the quality of evidence for a specific research question. The study design, target population, type of nutrition-specific policy evaluated and overall findings were summarised quantitatively across all studies. Data on the time span between policy implementation and outcome assessment, category of the primary outcome assessed and general conclusion of the study regarding the impact of the policy on child nutritional status were post-coded following data extraction.
Results
Of the 5646 abstracts screened, eighty-three studies were included in the review (Fig. 1). The studies included were conducted across thirty-seven countries (see online supplementary material, Supplemental Figure 1 and Table 2). The included studies were in seven geographical areas (based on The World Bank classifications): eighteen in East Asia and the Pacific(Reference Li, Fan and Chen20–Reference McNair and Learoyd37), seventeen in North America(Reference Bitler and Currie38–Reference Jackson54), thirteen in Latin America and the Caribbean(Reference Martorell, Ascencio and Tacsan55–Reference Cortés, Mellado and Pardo67), ten in South Asia(Reference Jaiswal, Melse-Boonstra and Sharma68–Reference Kapil77), seven in Europe and Central Asia(Reference Rossi and Branca78–Reference Vaish, White and Daly84), seven in Sub-Saharan Africa(Reference Agaba, Pomeroy-Stevens and Ghosh85–Reference Sebotsa, Dannhauser and Jooste91), six in Middle East and North Africa(Reference Abbag, Abu-Eshy and Mahfouz92–Reference Zlotogora, Amitai and Leventhal97) and five were multi-region(Reference Andersson, Karumbunathan and Zimmermann98–Reference Rosano, Smithells and Cacciani102). The majority of studies were conducted in upper-middle or high-income countries (66 %, 54 studies); approximately 48 % (40 studies) were conducted in just five countries: the USA, China, India, Canada and Brazil. Only 7 studies (8·4 %) were from low-income countries. Most studies were published from 2010 to 2017, evaluated policies at the regional/sub-national level, assessed outcomes among children under 5 years of age and were based on community- or school-based populations (Table 2). In high-income countries, data from national registries were commonly used (e.g. registries of birth defects to examine the effect of folic acid interventions) or from national programmes such as the Special Supplemental Nutrition Program for Women, Infants, and Children in the USA which serves as a contact point for large numbers of mothers and children. The sample sizes of such registry-based studies are notably larger than studies directly sampling from the population base (see online supplementary material, Supplemental Table 2). Cross-sectional (49 %, 41 studies) and pre-/post-design (19 %, 16 studies) were the most commonly used study designs (Table 2). Iodine deficiency disorders (40 %, 33 studies) and neural tube defects (19 %, 16 studies) were the most assessed child nutrition outcomes (Table 2). The impacts of policy interventions on outcomes such as low birth weight, breast-feeding rates, vitamin A deficiencies and child growth were less frequently assessed. Nonetheless, child growth was an important primary outcome evaluated in the context of a wide range of broad, multi-sectoral and food provision strategies (see online supplementary material, Supplemental Table 3).
Table 2 Characteristics of studies included in the scoping review
* Percentages do not always total to 100 % due to rounding.
† Low-income countries (n 7), lower middle-income countries (n 18).
‡ Other study designs include: repeated cross-sectional (n 7), time trend analysis (n 5) and case–control (n 2).
§ Other study bases include: registry-based (n 5), multiple (n 4), national surveys (n 3), health record review (n 2), not recorded (n 2), surveillance system (n 1), day care centre (n 1), convenience sample (n 1) and other longitudinal study (n 1).
|| If children were recruited across our predefined age bands, the oldest category was selected. For example, if a study included children aged 4–7 years, it was placed in the 5–10 category.
¶ Outcomes assessed were post-coded according to the main outcome assessed.
** Studies evaluating the effect of nutrition policies on ‘birth outcomes’ included: birth weight (n 5), fetal growth restriction (n 1) and preterm birth (n 1).
†† Studies evaluating the effect of nutrition policies on ‘child growth’ included: chronic and acute malnutrition (n 3), height-for-age z-score (HAZ) < –2SD (stunting) (n 2), weight-for-height z-score (n 1), HAZ (n 1), growth velocity (n 1) and underweight (n 1).
‡‡ Studies evaluating the effect of nutrition policies on iodine deficiency includes: urinary iodine concentration (n 27), iodine deficiency diseases such as goitre prevalence (n 6).
§§ Studies evaluating the effect of nutrition policies on vitamin A deficiency included: xerophthalmia (n 2) and serum retinol (n 2).
|||| Other types of child outcomes assessed included: anaemia (n 1), congenital heart defect (n 1), cognitive development (n 1), morbidity (n 1), care-seeking (n 1), folate status (n 1), micronutrient deficiency (n 1) and mortality (n 1).
The most common type of policies evaluated were food fortification policies (52 %, 43 studies), followed by food provision with nutrition education (10 %, 8 studies) and supplementation policies (8·4 %, 7 studies) (Table 3). These policies were largely established between 1991 and 2000 (48 %, 40 studies); the majority of studies evaluated impact within a year (28 %, 23 studies) or within 2–5 years since policy implementation (14 %, 12 studies). Studies typically evaluated policies that were mandatory (71 %, 59 studies, spread evenly across all income categories). Sixteen studies (19 %) did not specify whether policies being evaluated were mandatory or voluntary. Studies examining food provision and nutrition education were found only in high-income countries, whereas in low- and middle-income countries, nutrition policies were primarily related to micronutrient food fortification or supplementation.
Table 3 Key characteristics of policies examined in studies reviewed
* Low-income countries (n 7), lower middle-income countries (n 18).
† Other types of policies included: international resolutions (n 1), school lunch programme (n 1) and infant feeding guidelines (n 1).
‡ These data reflect the earliest policy mentioned and the earliest year of implementation. The year of policy implementation was post-coded and represents the year the policy was first implemented, as reported by the paper. If the paper did not report a year, it was coded as ‘not specified’.
§ The duration of time between implementation and outcome assessment is reported as the earliest year that outcome data were collected on the policy being evaluated. If the year was not reported or the study was across multiple countries, it was recorded as data unavailable.
Although we did not aim to assess the effectiveness of policies, it is worth noting that over two-thirds of the studies included in this review concluded that the policy in question had a positive impact on child nutritional status (see online supplementary material, Supplemental Supplementary Table 2).
Discussion
To the best of our knowledge, this is the first exercise to map the diversity of research on the impact of nutrition-specific policies to address child undernutrition at the population level globally. Our findings suggest that a wide range of regional and sub-regional nutrition policies for child undernutrition have been evaluated. Most studies were observational in design, examined health outcomes in children 5–10 years of age as opposed to younger children and were conducted in high-income countries. The impact of policy interventions on low birth weight, breast-feeding rates, vitamin A deficiencies and child growth was assessed less frequently than those relating to iodine and folic acid. There were relatively few studies evaluating the impact of national-level policies; the majority focused on regional and sub-regional policies. In addition, there was a notable scarcity of studies evaluating nutrition-specific policies in countries in Africa, where the burden of maternal and child undernutrition is greatest.
The nutrition policy agenda has evolved dramatically in the last 50 years. Starting from a focus on the socio-economic causes of protein–energy malnutrition (namely poverty) in the 1970s to community programmes to alleviate micronutrient malnutrition in the 1980s, global nutrition was only placed at the forefront of international agenda when it was recognised as a human rights issue during a series of high-level global meetings in the 1990s(Reference Nomura, Takahashi and Reich103). This heightened visibility of undernutrition contributed to the inclusion of ‘zero hunger’ as one of the central goals in the MDG. Since then, the influential Lancet Nutrition Series(12,104,105) have systematically quantified the burden of maternal and child undernutrition to measure global progress, identified evidence-based interventions to tackle undernutrition in early life and advocated for policy reforms to incorporate nutrition as a key priority in agenda-setting at the national level. Importantly, the Lancet Nutrition Series also framed optimal nutrition from the life course perspective, highlighting its relevance at all stages of the life cycle. Global nutrition policy has therefore evolved from focusing on singular issues, such as the elimination of extreme hunger, to framing nutrition as a development goal, thereby mainstreaming nutrition in the priorities set forth by the SDG(Reference Nomura, Takahashi and Reich103). However, the scope of research evaluating the effectiveness of global nutrition policy to improve population-level health and nutrition governance has been unclear(Reference Bump106).
Governments can use a myriad of different instruments to enact nutrition policies(Reference Mozaffarian, Angell and Lang107). The findings of this review show that the type of nutrition policies that have been implemented over time in the context of undernutrition and the studies examining the effectiveness of these policies at the national level have focused primarily on micronutrient deficiencies. This may be in part because most studies included were conducted in high- and middle-income countries where the relative prevalence of stunting and underweight is low compared with lower-income countries. It may also in part be because of the nature of policies themselves. Micronutrient fortification, for example, is relatively well-defined, often implemented at the national level, requires limited cultural market adaptation (e.g. language or visual packaging adaptations) and is more likely to be a mandatory policy. In addition, an essential element in studying the impact of any nutrition policy on health outcomes is the availability of nationally representative data on health outcomes that are comparable across countries. Cross-national data on nutritional outcomes of young children are limited. The types of policies studied and gaps among the papers included in this review therefore reflect the potential downstream effects of complexities and financial barriers to collecting health outcome data at the national level. Unlike high-income settings where data on health and food systems are frequently collected, monitored and evaluated, such information systems are not readily available in low- and middle-income countries, making it challenging to study policy impact.
Nevertheless, in any discussion of nutrition policy, it is important to note that irrespective of proven policy efficacy, a number of implementation factors can limit effectiveness in achieving public health impact(108). Such factors include coverage (whether the policy is implemented at scale and to what extent); distribution and supply chain challenges; uptake at the community, institutional and individual/household levels; and enforcement. Governments are often constrained by resource limitations for implementation. For example, despite salt iodisation being one of the most successful public health nutrition interventions and currently mandated in 118 countries worldwide(109) (86 % of households worldwide and 70 % in low- and middle-income countries access iodised salt(110,Reference Mason, Shrimpton and Saldanha111) ), the number of households actually consuming it is estimated to be as low as 20 % for countries such as India(Reference Mason, Shrimpton and Saldanha111). This highlights the importance of measuring indicators of nutrition policy implementation, such as availability, access and consumption, in addition to measured nutrition outcomes at the population level. Such process evaluations could advance our context-specific understanding of bottlenecks and specific inhibitors to policy effectiveness. Standardised tools, such as the ‘Fortification Assessment Coverage Toolkit’ (FACT)(112), can be used to guide consistent and effective implementation evaluation of nutrition-specific policies in this way. In addition, as noted in a recent real-time analysis of advocacy efforts for infant and young child feeding in Southeast Asia, strong policy advocacy can be a key driver to overcoming challenges experienced in translating policies into action(Reference Michaud-Létourneau, Gayard and Pelletier113).
Some strengths of this review include the systematic approach taken in searching for, screening and extracting data from articles, as well as the global nature and broad scope of the search to identify studies of nutrition-specific policies for undernutrition. Although overnutrition is an increasingly important health concern as the double burden of malnutrition grows(114), particularly in high- and middle-income countries, overweight and obesity were not included in this review. We further restricted the scope of this review to nutrition-specific policies to comprehensively map the heterogeneity in policies and nutritional outcomes evaluated in the context of immediate causes of child undernutrition; a similar exercise to map the breath and scope of the evidence to evaluate nutrition-sensitive policies that indirectly affect child undernutrition is needed. Policies which aim to influence the underlying determinants of undernutrition(Reference Ruel and Alderman115) have the potential to have large impacts on both maternal and child nutritional status. Only peer-reviewed research published in English was included in this review. We were additionally unable to account for the extent to which policies are implemented. For example, written policy statements (even mandatory ones) may not reflect the reality of what is practiced, funded, promoted or enforced; the latter can be most effectively assessed in country-case studies using qualitative methods. While the diversity of studies captured in this scoping review reflects the breadth of the research on nutrition policies on undernutrition in the MDG era, it was difficult to group policies perfectly based on predefined characteristics. For example, some studies evaluated policies targeting the age group of interest (children under 10 years) but with a wider age range (e.g.: ages 8–12 years). Other studies evaluated a multi-component policy intervention or the effects of two simultaneous policy interventions (e.g. nutrition education alongside provision of food supplements to the same beneficiaries). As a recent study by Chen et al. (2019) demonstrated, even differences in phrasing of the same policy can translate into differences of implementation and outcome (in this case, a nutrition subsidy policy with the goal phrased either in a specific (‘anemia reduction’) or general (‘malnutrition reduction’))(Reference Chen, Pei and Bai116). Policies in the same ‘categories’ can vary widely, and these variations should be examined in future reviews. The heterogeneity across policies and studies highlights the complexity of policy design in nutrition and points to the need for common guidance on the collection of optimal indicators of nutrition policy.
As one of the Scaling Up Nutrition movement’s strategic objectives is to ‘institutionalize effective actions that contribute to good nutrition,’ including policies(3), evaluating the long-term impact of nutrition policies is essential. Data on population health and nutritional status are needed to advocate for and inform policy design, and rigorous evaluations must be planned before the introduction of new nutrition policies(Reference Mason, Shrimpton and Saldanha111,Reference Wirth, Laillou and Rohner117,Reference Mason, Deitchler and Mathys118) . This highlights the importance of increasing the capacity for policy monitoring and evaluation longitudinally, including gathering evidence in comparable non-policy areas where possible(Reference Mason, Deitchler and Mathys118). As called for by the 2020 Global Nutrition Report, in addition to monitoring and evaluating policies moving forward, it is imperative to examine nutrition inequities and to implement evidence-based, equity-sensitive nutrition policies to address such inequities(114). Future evaluations of population-level nutrition policies should therefore examine policy impacts disaggregated by sub-national parameters, such as rural/urban divide and wealth quintile, to better understand inequities and to enable appropriate resources to target marginalised populations(Reference Goh119). Finally, evaluating the political economy of a given policy to better understand the role of power, incentives, institutions and ideas on shaping policies will be equally important to inform and support nutrition security through policy efforts in the long run(Reference Reich and Balarajan120).
In summary, while the studies included in this review suggest that nutrition-specific policies contribute to improvements in child nutritional status, the scope of the current nutrition policy research evidence base is limited. Most studies were conducted in high-income settings at the sub-national level and evaluated the role of micronutrient-specific nutrition policies on a narrow set of health outcomes. Further research is needed to assess the impact of a broader range of policies (including those related to nutrition education and infant and young child feeding practices); disentangle the complexity of nutrition policy design; and present a nuanced view of policy impact considering sociodemographic inequities. In addition, measures of child nutrition, such as dietary diversity, dietary/nutrient intake data and intervention coverage, may provide useful proxy measures of child nutritional status to assess policy impact in settings where population-level data are sparse. Importantly, nutrition policy research in low-income countries is particularly scarce and should be a research priority. Tangible evaluations of the political commitments to improve nutrition through policy provide an invaluable avenue for accountability to achieving the global targets for malnutrition reduction set in the 2012 World Health Assembly, the SDG, and those which will undoubtedly be set forth in the future.
Acknowledgements
Acknowledgements: The authors gratefully acknowledge Thomasin Adams-Weber for support designing the literature search and Grace Regent for research support. Financial support: This work was supported by the 2017 Catalyst Grant awarded by the Centre for Global Child Health, Hospital for Sick Children, Toronto, Canada. The funder had no role in the study design, data collection or analysis. Conflict of interest: There are no conflicts of interest. Authorship: N.P., B.B. and S.Z. formulated the research question and designed the study. B.B., A.R., V.P. and N.P. screened articles, extracted data and compiled research results. V.P. drafted the manuscript, and all authors critically reviewed the manuscript for intellectual content and approved the final version. Ethics of human subject participation: Not applicable.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980021001890
