http://orcid.org/0000-0001-9497-7901
Figures
Abstract
Introduction
Non-communicable diseases (NCDs) disproportionately affect low- and lower-middle income countries (LLMICs) where 80% of global NCD related deaths occur. LLMICs are the primary focus of interventions to address development and poverty indicators. We aimed to synthesise the evidence of these interventions' impact on the four primary NCDs (cardiovascular disease, diabetes, chronic respiratory disease and cancer) and their common behavioural risk factors (unhealthy diets, physical inactivity, tobacco and alcohol use).
Methods
We systematically searched four online databases (Medline, Embase, Web of Science and Global Health) for primary research conducted in LLMICS, published between January 1st 1990 and February 15th 2016. Studies involved development or poverty interventions which reported on outcomes relating to NCDs. We extracted summary level data on study design, population, health outcomes and potential confounders.
Results
From 6383 search results, 29 studies from 24 LLMICs published between 1999 and 2015 met our inclusion criteria. The quality of included studies was limited and heterogeneity of outcome measures required narrative synthesis. One study measured impact on NCD prevalence, one physical activity and 27 dietary components. The majority of papers (23), involved agricultural interventions. Primary outcome measures tended to focus on undernutrition. Intensive agricultural interventions were associated with improved calorie, vitamin, fruit and vegetable intake. However, positive impacts were reliant on participant's land ownership, infection status and limited in generalisability. Just three studies measured adult obesity; two indicated increased income and consequential food affordability had the potential to increase obesity. Overall, there was poor alignment between included studies outcome measures and the key policy options and objectives of the Global Action Plan on NCDs.
Conclusions
Though many interventions addressing poverty and development have great potential to impact on NCD prevalence and risk, most fail to measure or report these outcomes. Current evidence is limited to behavioural risk factors, namely diet and suggests a positive impact of agricultural-based food security programmes on dietary indicators. However, studies investigating the impact of improved income on obesity tend to show an increased risk. Embedding NCD impact evaluation into development programmes is crucial in the context of the Sustainable Development Goals and the rapid epidemiological transitions facing LLMICs.
Citation: Pullar J, Allen L, Townsend N, Williams J, Foster C, Roberts N, et al. (2018) The impact of poverty reduction and development interventions on non-communicable diseases and their behavioural risk factors in low and lower-middle income countries: A systematic review. PLoS ONE 13(2): e0193378. https://doi.org/10.1371/journal.pone.0193378
Editor: Patrick Kolsteren, Universiteit Gent, BELGIUM
Received: August 23, 2017; Accepted: February 1, 2018; Published: February 23, 2018
Copyright: © 2018 Pullar et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This project was commissioned by the World Health Organisation (2015/577190-0). Kremlin Wickramasinghe and Nick Townsend are supported by a grant from the British Heart Foundation (006/P&C/CORE/2013/OXFSTATS).
Competing interests: The authors have declared that no competing interests exist.
Introduction
Non-communicable diseases (NCDs) predominate the global disease burden, disproportionally affecting low and lower-middle income countries (LLMICs)[1]. Latest estimates suggest 82% of premature deaths attributed to NCDs occur in LLMICs, with further increases predicted in the prevalence of the four most prevalent NCDs- cardiovascular disease (CVD), type 2 diabetes (T2DM), cancer and chronic obstructive (COPD) respiratory diseases [1]. This growing burden compounds already stressed health systems, disproportionately affecting the poorest populations and hindering countries social and economic development [2].
Since the late 1990s, the Millennium Development Goals (MDGs) have formed the basis of work in health, poverty and development within LLMICs [3]. Reflecting the key issues when they commenced, the MDGs have focused development initiatives towards supporting poverty reduction, education, gender equality, environmental sustainability, child mortality, maternal health and the control of infectious diseases. While great progress in these areas has been made, globalization and resulting epidemiological and nutrition transitions has led to the rapid emergence of NCDs within LLMICs [4].
Ten years after the inception of MDGs, the global burden of disease study identified major shifts in the magnitude of health risks within LLMICs [1]. Globally, traditional causes of ill health in children (diarrhea, infections, undernutrition) slipped down in risk ratings as NCD related risk factors rose. Though as the burden of NCDs climbs, many LLMICs continue to battle infectious disease epidemics and undernutrition [5]. For example, in Sub-Saharan Africa where malnutrition remains the leading health risk factors, the greatest rise in NCDs has been evidenced. By 2030 46% of deaths within this region are expected to be related to NCDs [6].
The mounting development challenge of NCDs was recognised by the World Health Organisation (WHO), who declared NCDs were ‘putting a break on development, undermining the MDGs and amplifying social inequality’ [7]. The 2011 Political Declaration on the prevention and control of NCDs (Paragraph 1- Resolution A/RES/66/2), concluded the way to combat NCDs was clear–by tackling the four primary modifiable behavioural risk factors for NCDs- specifically, tobacco use, physical inactivity, unhealthy diets and the harmful use of alcohol within all member states, and control of the four diseases [7].
To address these risk factors, the WHOs Global Action Plan (NCD GAP) on NCDs provides a roadmap for governments to address NCDs by outlining key policy options to attain 9 voluntary targets by 2025 [8]. In achieving these targets a 25% reduction in premature mortality from NCDs is predicted- this is particularly important for LLMICs where 82% of premature NCD related deaths occur [1,8]. The first objective of the NCD GAP explicitly outlines the aim of increasing the priority of addressing NCDs within international development agendas through strengthened co-operation and advocacy [8]. This is supported by NCDs inclusion in the Sustainable Development Goals (SDG) [8]. These goals recognize NCDs as a major barrier to sustainable development and aim to achieve a reduction in NCD-related premature mortality by one-third by 2030. The SDGs complement policy objectives and interventions of the GAP NCDs to encourage cohesive action by member states and the incorporation of NCDs into national development plans. To support this, a strong evidence base is required to understand ways of effectively linking development agendas to NCD prevention and control in practice.
The promotion of a core set of cost-effective preventative and curative interventions, commonly referred to as ‘Best Buy Interventions’, aims to assist governments in implementing NCD interventions [8–9]. However these are largely focused on health-centric interventions. Recognising many causes and solutions of the NCD burden lie outside of the health system, growing emphasis is placed on supporting multi-sector approaches to NCD prevention. In such approaches, complementary sectors such as agriculture and education work in collaboration with the health sector towards shared goals. Guided by the MDGs and SDGs, many development initiatives have the potential to impact on recipient’s risk of NCDs [5].
To the best of our knowledge, no systematic review has examined the impact of development and poverty reduction interventions on NCDs and behavioural risk factors. Such a review has the potential to provide important evidence about whether or not current development projects are impacting on NCD outcomes and to inform future projects and policy makers catalysing multi-sector approaches for NCD prevention and control. These approaches encourage inter-governmental and inter-agency collaboration to address NCDs through the establishment of synergistic objectives and interventions which can positively impact on areas such as agriculture, education, fisheries management and climate action while also supporting healthy environments for NCD prevention [10].
This systematic review aims to answer the question, ‘what is the evidence for the impact of development interventions on NCD outcomes and behavioural risk factors in LLMICs? Included poverty reduction and development interventions were defined as those that addressed economic development, social inequalities, community engagement, agriculture, fisheries, water or sanitisation or worked towards attaining participant’s human rights. For the purpose of this review we considered the four NCDs which account for 82% of all NCDs. Namely-cardiovascular disease (CVD), type 2 diabetes mellitus (T2DM), cancer and chronic obstructive pulmonary disease (COPD) [8]. We also considered the four common, modifiable risk factors for these diseases- tobacco use, unhealthy diets, physical inactivity and the harmful use of alcohol [8].
This review aims to support ongoing NCD prevention methods and provide evidence for the planning and development of future development and poverty reduction initiatives. This is the third paper in a three-part series examining NCDs in LLMICs.
Methods
Protocol and registration
Our review followed PRISMA guidelines for systematic reviews. The supporting PRISMA checklist is available as supplementary information (S1 Table), and PROSPERO registered protocol (PROSPERO: 42016039030, 4201603903). Due to the low number of papers the findings of two PROSPERO protocols (covering NCD outcomes and NCD risk factors) are reported on within this paper. Ethics approval was not required for this work.
Database and search strategies
We conducted a comprehensive literature search for papers published between January 1st 1990 and February 15th 2016 using a pre- determined search strategy (S2 Table). The search strategy was developed by the medical librarian and MESH terms formulated to cover the PICO elements of the research question. Search terms combined synonyms for poverty and development interventions with terms for NCD morbidity, mortality and behavioural risk factors as well as the 82 LLMIC countries defined according to the World Bank’s 2015 definition [11]. We conducted the search in English on the following electronic databases: MEDLINE, EMBASE, Web of Science Core Collection and Global Health. Additionally, we reviewed the first 30 hits from Google Scholar and searched the websites of WHO, World Bank, United Nations Food and Agricultural Organisation (FAO). Reference lists of key papers identified within the literature search were also reviewed.
Inclusion criteria
Included studies evaluated the impact of an intervention aimed to address a poverty and/or development indicator on at least one NCD morbidity, mortality or behavioural risk factor outcome measure within LLMICs (S3 Table). The population of interest was development programme recipients within LLMICs, with no restriction placed on age. Definitions of parameters are listed below:
Development and poverty reduction interventions (referred to as development interventions throughout):
- Economic development interventions (including financial management/ support, small business promotion, entrepreneurship, market linkage)
- Social change interventions (including projects which support gender equality, social justice, social support, human rights)
- Social protection interventions (including conditional and unconditional cash transfer, food for work, government subsidies, government taxes, microcredit programmes)
- Employment programmes (training, access, loans)
- Fisheries programmes
- Environmental protection programmes
- Agricultural programmes
- Water and sanitisation programmes
- Infrastructural and facilities improvement programmes
NCD morbidity and mortality outcomes [8]:
- Cardiovascular disease -myocardial infarction, heart failure, brain ischemia, stroke, heart disease, coronary artery disease, cerebrovascular events, vascular events, heart failure
- Diabetes- type 2 diabetes, non-insulin dependent diabetes, type 1 diabetes, insulin resistance, impaired glucose tolerance
- Cancer- neoplasms, carcinoma, tumors, malignancy, leukemia, lymphoma
- Chronic respiratory diseases-chronic obstructive pulmonary disease, chronic lung/pulmonary conditions, asthma, lung diseases
NCD behavioural risk factor outcomes [8]:
- Tobacco use -tobacco smoking or chewing, tobacco control policies
- Physical inactivity—physical activity levels, sedentary behaviour
- Unhealthy diet- diet composition, fruit, vegetable, salt, fat, sugar, calorie intake
- Harmful use of alcohol- alcohol consumption
Grey literature was excluded. No restrictions were made based on publication language, study design or participant characteristics.
Study selection and data extraction
Citations from search results were collated in Endnote, duplicates removed, and the final results exported into Excel for eligibility screening. JP and LA independently screened titles and abstracts against the predefined eligibility criteria (S3 Table). To ensure consistency, Cohen’s kappa statistic was calculated at 10% intervals (every 638 papers). Once “excellent agreement” was reached (agreement exceeding 95% and Cohen’s kappa > 0.75) JP screened all remaining records [12]. Uncertainties were brought to LA, KW and NT with disagreements resolved by group consensus. This process was repeated for full text review. For studies where the data was unclear, or more information was required the authors were contacted by email.
Following full text review, references of selected studies were examined for any additional studies of relevance. Data collection was then conducted on all eligible papers. Data collection involved the extraction of key study parameters including country, intervention design, participant population, NCD outcome and quality scoring measures.
Quality assessment
A quality assessment was conducted on all included papers by JP and LA using validated tools endorsed by Cochrane reviews (S4 Table). For randomised controlled trials (RCTs) the Cochrane Risk of Bias tool was used to evaluate the risk of biased conclusions based on selection, performance, detection, attrition, reporting and other bias [13]. Observational cohorts were assessed using the Newcastle-Ottawa Scale (NOS) [14]. NOS scores studies based on methodological rigor by assessing study selection, comparability of cohorts and outcome assessment. Under each category, studies were assessed based on the coding structure developed by Wells et al [14]. The quality of cross-sectional studies was also assessed using an adapted version of the NOS which assessed selection, comparability and outcome bias using the same letter and star system as used for cohort studies. None of the utlised bias assessment tools have a validated method of summarizing overall bias scores. Instead, the scorings are used to discuss strengths and weaknesses within each paper, as well as common areas of weaknesses at an outcome and review level.
Results
Study selection
The database search retrieved 8,094 citations with a further 114 identified through additional searches. After the removal of duplicates, 6,383 citations underwent abstract screening, and 134 full text review. Twenty-nine citations met the study inclusion criteria (Fig 1). Primary reasons for exclusion included the absence of development indicator, study not based in an LLMIC, not primary research, not involving an intervention or not including an NCD related outcome measure.
Number of papers screened and included in review process.
Study characteristics
Included studies were published between 1999 and 2015, with 66% published after 2010. Studies were conducted in 24 different LLMICs. Of these, 15 were from WHOs African Region (AFRO), nine from South East Asia (SEARO), three from Western Pacific (WPRO), one from Eastern Mediterranean (EMRO) and one from the Americas (PAHO). There were no included studies from the European region (EURO). Fig 2 shows the number of low income (LIC) and lower-middle income (LMIC) countries, along with the number represented by studies within each region.
Number of LIC and LMIC countries within each WHO region, and the number of countries represented by included papers.
One study was published in Spanish [15], while 28 were in English. The majority of studies used observational design (20); three modeling and four were randomised controlled trials (RCTs). In relation to participant characteristics, 26 studies were conducted on participants from rural and farming households [15–41], 12 measured outcomes in children [18–23,25,28,33,40–42] and 23 measured adult outcomes (with nine focusing specifically on women[17,19–20,23,26,30,36,42–42]).
Of the included studies 27 reported on dietary outcomes (with 10 including measures of over-nutrition such as calorie consumption, overweight and obesity), one on physical activity and one on diabetes and ischemic heart disease (IHD). The primary development component involved agricultural interventions (23 out of 29 papers), the remaining six involved social protection programmes. Development interventions fell into six broad categories: bio-fortification, social protection programmes, agricultural diversification, livestock and fisheries management, and multi-component interventions.
In all studies measuring dietary factors, the aim of the intervention was to prevent or address undernutrition. Outcome measures of dietary studies varied greatly; the most common measure was diet composition- measured by the intake of different food groups. Three studies used the FAOs Household Food Security and Dietary Diversity Access scale [15,34,39].
Three studies evaluated outcomes at a population level [16,41,43] the remaining 25 measured outcomes at the individual level and were focused on addressing undernutrition, physical drudgery and water contamination.
Non-communicable diseases
The only study assessing impact on NCD morbidity or mortality to meet inclusion criteria modeled the impact of mitigating arsenic exposure via the substitution of surface water for well water in Bangladesh and the resulting impact on diabetes and ischemic heart disease (IHD) prevalence [16]. Intending to update a previously used model, which suggested a reduction in disability adjusted life years relating to diabetes and IHD through mitigating arsenic exposure. However, investigators determined that further work was required to first establish and quantify the relationship between diabetes, IHD and arsenic exposure [16]. Therefore, no studies reported on the impact of development interventions on NCD morbidity or mortality.
NCD behavioural risk factors
Alcohol and tobacco use.
No included studies examined the impact of development interventions on alcohol or tobacco use.
Physical inactivity.
Only one small observational cohort, involving 120 female hill farm workers in India, measured impact on physical activity (Table 1) [17]. The introduction of an improved clod breaker effectively reduced ‘drudgery’ associated with the farming shown by reduced heart rate and self-perceived exertion. No studies measured the impact of interventions on increasing physical activity or reducing sedentary behaviour.
Unhealthy diets.
The remaining 27 studies investigated the impact of interventions on dietary measures [18–43]; these are reported below based on the type of development intervention.
Bio-fortification.
Four studies examined the impact of bio-fortification focused agricultural interventions on the consumption of vitamin A rich fruits and vegetables (Table 2). Three of these studies involved the promotion of Orange Flesh Sweet Potato (OFSP) [18–20]. All studies identified the promotion and provision of OFSP vines increased the production and consumption of OFSP- consequently increasing vitamin A intake in women and children.
Two interventions that were implemented and assessed by HarvestPlus found increased programme intensity and duration did not influence results [19–20]. The relationship between OFSP intake and serum retinol was less clear. In the pilot study for OFSP a significant 22% reduction in children diagnosed with Vitamin A Deficiency (VAD) was found after 2 years, as well as significantly higher intakes of calories and other essential vitamins [18]. However, analysis involved just 38% of children within the intervention group. Also, although the prevalence of VAD dropped from 60% within intervention children at baseline, 38% of measured recipients remained classified as VAD after 2 years [18]. Seasonal availability could be influencing this result as the percentage of children consuming OFSP 3+ days/week dropped from 54% to 8% within the intervention group outside of harvest season [18,20]. A follow-up HarvestPlus study found a significant 9% reduction in the percentage of children with low serum retinol (<1.05) after 2 years [20]. Though no difference in serum retinol was found for women, or children with diagnosed VAD. Assessment was also limited to those with low serum retinol at baseline in order to maximize sensitivity [20]. All studies highlighted the significant impact vitamin A supplementation had on serum retinol levels, though low levels of national coverage hindered impact.
Analysis of a broader vitamin A community bio-fortification project found that significant changes in dietary behaviour remained, notably increased vegetable consumption, five years after implementation [21]. However, the impact of these behaviour changes on nutrition status was hindered by the presence of helminth infection. As a result, no significant differences between intervention and control serum retinol were found.
Social protection programmes.
Eight studies assessed social protection programmes implemented to prevent food insecurity in areas affected by economic crisis (Table 3) [15,22–26,42–43]. As they focused on preventing undernutrition, the majority of papers focused on measurement of stunting, body weight and calorie intake to assess calorie deficits, as well dietary diversity to assess dietary quality.
An examination of varying methods of delivery for cash transfers found those delivered in combination with supplementary foods were twice as effective in preventing undernutrition within children, compared to cash in isolation [22]. The decline in moderate undernutrition increased as the calorie content of supplementary food increased. A large study examining the impact of food-for-work programmes on anaemia found no impact on anaemia for children or women in rural settings [42]. Only a small improvement in anemia was seen within one subset of urban women. In comparison to other settings, the women in this urban area received pinto beans (in addition the rice and oil received by other groups), had the longest programme participation and received direct nutrition education [42].
A small but significant protective effect on childhood undernutrition was identified for a microcredit programme in Ghana, which incorporated nutrition education, though it was not sufficient to overcome an overall reduction in child weight and BMI due to a prevailing drought [23].
The intensive Shamba Maisha project in Kenya supported improved agriculture through conditional loans requiring health clinic visits for HIV sufferers [24]. The project found a significant reduction in underweight and food insecurity. Another agricultural loan intervention in Honduras found no change in calorie intake despite increased crop stores [15]. Two programmes investigated the impact of creating paid employment opportunities on nutrition. Both identified a positive impact on children’s nutrition status- reducing levels of undernutrition [25] as well as increasing household food expenditure and consumption–particularly of green leafy vegetables [26]. Social protection programmes comprised of direct nutrition education, provision of nutrient rich food, and longer duration were associated with greater impact on diet.
In Egypt, an economic analysis of the impact of government food subsidies on national obesity levels found a significant association between reducing bread, wheat, flour, sugar and cooking oil prices and increased female obesity levels. Conversely, subsidies on healthy foods were modeled to reduce BMI [43].
Agricultural diversification.
Eight studies examined the impact of agricultural diversification interventions on dietary outcomes (Table 4) [27–34]. Two involved implementation of a Helen Keller International homestead garden programme. One before/after evaluation utilising cross-sectional surveys found a significant increase in nutritious food consumption (notably vegetables) after programme implementation, that did not equate to improved nutrition status [27]. The second, larger RCT identified an improvement in hemoglobin levels in young children [28]. Results from this study may have been influenced by hygiene improvements which were also promoted during the trial and could have reduced infection, another cause of low haemoglobin.
Two small pilot studies found the promotion and cultivation of local fruits and vegetables had the potential to increase consumption and sales [29–30]. This was consistent with two crop diversification projects which found an increase in vegetable intake, though not to a high enough level to reach requirements [31–32]. Two studies from Nepal involving interventions to increase market access and sale of cash crops showed improvements in crop production, including fruits and vegetables, as well as increases in vegetable intake [33–34]. These did not have a significant impact on childhood underweight when controlled for other factors [34].
Livestock and fisheries management interventions.
Three studies examined the impact of improved livestock and fisheries management on income and nutrition status (Table 5) [35–37]. Two focused primarily on dairy production (including the introduction of milk producing cows, and dairy management training) [35–36]. One showed an increase in dietary fat and iron intake [35] while the other reported an increased intake of milk and green leafy vegetables [36]. They also reported on potential indicators of over consumption with one reporting an increased calorie intake to over the recommended 2000 calories a day [35], and the other reporting an increase in obesity [36]. The remaining livestock intervention focused on the improved management of aquaculture ventures in Bangladesh which resulted in significant increases in fish consumption [37]. All livestock interventions resulted in an increased consumption of animal source foods. No studies adjusted for confounders.
Multi-component programmes.
Four programmes involved multi-faceted interventions addressing development indicators which measured dietary outcomes (Table 6) [38–41]. One was a population level analysis which aimed to examine the impact of Vietnam’s VAC programme which encouraged nationwide adoption of traditional agriculture techniques to improve the diversity of farmed crops and livestock [41]. The impact of this national initiative was evaluated by extrapolating results of the national nutrition survey which suggested an increase in dietary calories, fat, protein and a reduction in VAD [41].
Two studies evaluated the impact of improved agricultural technologies on crop outputs and food security [38–39]. In one, no difference in self-reported calorie intake was found for programme recipients, though protein intake showed a smaller reduction than was reported by control participants as a result of prevailing droughts [38]. The other was a modeling study which predicted that a food security programme focused on improving water irrigation for agriculture would result in a 30% increase in available calories (695cal/day) [39].
The final study evaluated the impact of the Millennium Development Project in nine settings within low income AFRO countries [40]. While significant improvements were seen in the number of meals, dietary diversity and stunting–these may have been influenced by the use of community health workers during implementation as opposed to solely agricultural and poverty reduction components [40]. The use of total dietary diversity scores mean we are unable to quantify increases in specific food groups.
Obesity.
Within the 25 studies that met eligibility criteria by including outcome measures related to unhealthy diets, eleven studies also included a measures of calorie intake and body mass index in adults [15,18,23–27,35–36,39,43]. These measures can be used to assess overconsumption and obesity, both risk factors for NCDs. In the six studies which measured calorie intake [15,18,35,38–39,41], four identified an increase [18,35,39,41]. Two of these studies found an increase to over 2000 calories [35,39], both were based in Ethiopia and involved the promotion of livestock projects in participants involved in agricultural work. Six studies evaluated BMI with two agricultural diversification projects finding no impact [23,27], while four showed an increase [. Two studies reported only on a reduction in underweight participants [24,26], two showed an increase in obesity when evaluating livestock management programmes [36] and government food subsidies [43], both of which aimed to improve food security. No agricultural diversification projects showed any significant increase in calories or body weight.
Quality assessment
Application of quality assessment tools identified several common weaknesses in the included studies. RCTs (S4 Table) rated as ‘unclear’ in relation to their level of reporting bias as they tended to suit the description of observational trials more than RCTs [19,20,24,28]. In particular, there was no reporting on the process of randomization or allocation concealment and study design meant participant/ researcher blinding was not possible. Both bio-fortification trials also lost over 10% of participants within intervention arms influencing outcome data [19,20]. They were also all linked to, or were conducted by the organization providing OFSP which could introduce reporting bias [18,19,20]. Weiser et al (2015) was conducted independently from the World Food Program funders and showed lower risk of bias, aside from the inability to blind participants/researchers and the possibility of cluster level variables which were not adjusted for [24].
Cohort and cross-sectional studies also showed common limitations. This was particularly apparent in their use of convenience sampling of highly motivated, high needs, easily accessible programme recipients- limiting the generalizability of results (S4 Table). Many also experienced a high attrition rate, or failed to record the attrition rate which could mean the true impact is lower than reported. As many measured dietary intake of food groups, there is also likely to be great variation between studies in this measure, with measures including self-report, self-completed FFQs, guided FFQs, structured interviews, diet recall and dietary diversity measurement using pre-validated forms. While cluster level differences, household characteristics and infection all proved to be important confounders, many studies did not adjust for these factors. Ten papers made no adjustments for confounders and only two scored highly for controlling for confounders (S4 Table).
Overall, studies led by university groups as opposed to development partners tended to report lower magnitudes of impact, adjust for potential confounders and discuss limitations.
Discussion
This review highlights the paucity of NCD related outcome measures within the design and evaluation of current development interventions within LLMICs. Very few papers included outcome measures or objectives which complemented global NCD prevention strategies which focus on reducing unhealthy behaviours, and implementing population based strategies [8,9].
For example, the paper relating to NCD morbidity was in relation to arsenic as opposed to the four modifiable behavioural risk factors [16]. The study addressing physical activity also aimed to reduce exertion as opposed to reducing sedentary behavior [17], while dietary related papers focused on the reduction of undernutrition as opposed to overnutrition [18–43].
Only 12 studies reported on outcomes consistent with the global NCD indicators stated within the NCD GAP [8], this included two on obesity and ten on fruit and vegetable intake. No studies reported on the global NCD indicators relating to the intake of saturated fat, salt, physical inactivity, alcohol or tobacco use. Seventeen studies reported on other dietary measures including undernutrition (via dietary diversity, micronutrient deficiencies and underweight), fortified crop intake, mitigating water contamination and one on reduced physical exertion from manual labor.
Overall, agricultural programmes which focused on bio-fortification and diversification of crops showed a positive impact on the consumption of nutritious foods, particularly vegetables. That said, this impact was often not sufficient to meet the WHO 400g/day recommendation for fruits and vegetables [8]. Additionally, these programmes showed inconsistent impacts on nutrition status measured primarily by vitamin A deficiency, anaemia, stunting and body weight in women and children. These mixed findings of impact on the nutrition status of women and children are supported by a previous review into the impact of nutrition-sensitive interventions [44]. With 81% of included studies involving farming households, there is also the risk that potential NCD risk factors which fall outside the primary four, have been missed. For example, agricultural projects which focus on increasing agricultural output and income via tobacco crops could unintentionally increase tobacco usage [45]. Exposure to residues from agricultural pesticides through agricultural work has also been found to increase the risk of CVD [46]. Agricultural interventions could therefore place participants at a higher risk of NCDs despite increasing their vegetable intake.
The high number of studies evaluating the impact of agricultural work interventions reflects the focus of many development interventions, especially within Africa, due to agricultures immediate impact on food security and long term benefits for economies. The physically demanding nature of agricultural work indicates many participants involved in the studies would have ample daily physical activity and high calorie requirements. It is therefore unlikely they would constitute the target population of interventions aimed at increasing physical activity. This possibility is reflected in this review’s findings of just one physical activity paper that aimed to reduce physical exertion in physically fit hill farming women [17]. It is also supported by another social protection programme which noted that the increased physical exertion brought on by employment raised concerns that the intervention may exacerbate undernutrition [25].
While not searched for specifically, an important finding within eligible studies was the association between increased access to food through (improved livestock management, provision of supplementary foods or improved affordability of nutrient poor foods, as a result of poverty reduction interventions) and increased prevalence of obesity. These findings complement findings from social protection programmes in Mexico and Brazil which have found an increase in obesity for recipients of food assistance programmes [47–48]. Within our review, cash transfer was also shown to increase body weight in undernourished children proportionate to the calorie content of supplementary foods [22]. The pattern of increasing purchasing power, supplementary foods and obesity raises an important consideration for future projects, which must consider the quality of foods as well as food access, in order to prevent contributing to the double burden of nutrition. This is an essential point of focus for future development programmes as it may supersede other noted health benefits [49].
Traditional measurements of diet within food security programmes could also contribute to this problem as they focus on calorie content over nutritional quality of available food. For example, dietary diversity scores are intended to measure diet quality by reflecting the number of different food groups consumed [50]. While reporting the total score can give an indication of improvements in diversity and provide a comparative measure–it does not reflect the quantities, nor give an indicator of the nutritional quality of consumed foods. In particular, ‘oils and fats’, ‘sweets’ and ‘spices, condiments, beverages’ are each separate food groups [50]. Hypothetically, an increase in one of these food groups would improve the total dietary diversity score. Reporting on the results of each food group would provide a clearer understanding of impact on dietary composition and nutritional quality. The importance of such considerations are highlighted by recent findings that for every 10% rise in GDP there is a 6% reduction in childhood stunting, and 7% increase in overweight prevalence [44]. With growing evidence that children affected by stunting may be at increased risk of NCDs later in life, it is essential to maximise the diet quality within LLMICs to prevent an escalation of the double burden of malnutrition [51]. Improved reporting and measurement of dietary indicators that reflect both under and over nutrition will be essential in achieving SDG 3 relating to NCDs.
Within included studies the focus on reducing undernutrition, physical hardship and exposure to harmful agents largely reflects the major, immediate health risk factors faced by included LLMICS. For example, in the African region, where the majority of included studies were conducted, the leading health risks are childhood under-weight, unsafe sex and unsafe water [52]. This suggests that while NCDs are a growing concern within the African region, differing risk profiles mean interventions outlined in the WHO Best Buy interventions may not be the most effective or urgent measures in reducing the disease burden.
Reflecting countries stage of epidemiological transition, other regions show a much higher prominence of modern risk factors. For example, in the South East Asia, where nine included studies were conducted, high blood pressure, BMI and smoking are now leading health risks [52]. The same is true for LLMICs within EURO, EMRO, WPRO and PAHO where NCD related health risks are superseding MDG related risks [5,52]. With only 14 studies from these regions identified, findings of this review highlight the lack of research on the synergistic impact of interventions addressing development indicators within LLMICs who face the highest burden of four key behavioural risk factors and consequential NCD outcomes.
Strengths and limitations
Many papers have reported on the risk of NCDs and behavioural risk factors within LLMICs [53], the impact of NCDs on international development and the need for multi-sectorial action to combat them [8]. This is the first review to assess what impact current development interventions are having on NCDs. It highlights the potential for future development interventions to impact on NCDs with intervention design, NCD sensitive objectives and outcome measurement. Our work complements previous reviews on the importance of nutrition sensitive agriculture and growing evidence of the possible negative impacts food assistance can have on NCD outcomes.
While no language restrictions were placed on included papers, the search was conducted in English, which may have been a limitation as English is an official language in only 35% of LLMICs. Exclusion of grey literature also narrowed the review’s scope, as development agencies may not have published all available results within academic journals. The decision to exclude grey literature was made by the review team in light of the overwhelming number of case-study reports and the tendency for such reports in international development interventions to be subject to outcome reporting bias [54]. This is influenced by a number of factors and generally demonstrates selective outcome reporting which favours the reporting of positive impacts which reflect favourably on the implementing partner and donor [55]. As evident in the studies included within this review, such reports also tend to employ anecdotal evidence without rigorous statistical analysis or peer review, and frequently funded by the programme funder. Many grey literature reports identified show great potential for future academic publications, such as modeling studies investigating the impact of trade agreements and improved human rights on NCD outcomes [56].
Additionally, the use of World Bank Classifications of LLMICs impacted on results, as countries often classified as ‘developing’ such as South Africa, Brazil, Thailand and Mexico, for which a greater number of studies are available, were not included [11]. Our review also failed to find any papers from the European region in which LLMICs face very high levels of NCD risks. Choice in search engines may also have swayed our citation findings towards health centric results.
The nature of development research could also have introduced a systematic bias towards positive results due to the closeness of studies to funders, advocates and political interests [57]. This could introduce publication bias as studies which identified a positive outcome were more likely to warrant time investment in the publication of results. The heterogeneity of papers means we are unable to conduct a funnel plot to assess this bias. There is also the potential for selection bias as interventions address the most at need subjects and many studies experienced a high loss to follow-up, as well as measurement bias in papers such as those evaluating bio-fortification where only children with severe deficiencies were included in analysis. Within our review, the positive impacts on health and development indicators may therefore be over-estimated.
An additional limitation of the review is the focus on specific dietary factors, as opposed to wider nutrition factors. There is growing evidence of nutrition factors such as early stunting and growth on NCD outcomes [58]. Preventing such states of undernutrition has historically been a key focus of development interventions and recent research has shown the great potential of interventions such as cash transfers [49], bio-fortification [59], and improved water and sanitisation [60] programmes to reduce and prevent stunting prevalence.
The descriptive nature of the review, due to the heterogeneity of papers, means we are unable to quantify the impact on NCDs. As many of the included studies, particularly agricultural diversification interventions, involved small sample sizes, convenience sampling of highest need populations and lack of control for confounders they are limited in their generalizability and not representative of national level programme focus. This is compounded by the fact that many countries, including all LLMICs in the European region, are not included in our review.
Conclusions
While current research on the duel impact of interventions on development, poverty reduction and NCD prevention is scarce, great potential exists for future programmes. It is essential to note that a key finding of this review is the lack of evidence for development interventions on NCD outcomes, as well as the lack of inclusions of NCDs in interventions outcome measures. A historical dominance of development interventions on addressing undernutrition has meant the measurement tools and study designs continue to focus on measures of undernutrition despite the growing double burden of malnutrition and NCDs. Measurement structures and programme focus on undernutrition do however create great potential. The promotion of healthy diets within development interventions is an important area for future focus.
Addressing health and NCDs is essential for interventions to have a sustainable impact on development and achieve the SDG targets. Failure to do so could compound current health problems and hinder development. Setting NCD sensitive objectives and measurements within poverty and development interventions could lead to simple adjustments, with great impact. Results from our review also highlight how NCD risk factors within LLMICS stray from the four commonly targeted behaviour changes. In measuring risk and impact on NCDs within future interventions, it is recommended to tailor programmes to country-specific risk factors. Improved measurement tools and guidance could enhance this process and strengthen multi-sector action and impact measurement to address NCDs within LLMICs. These improvements would also help to standardize reporting and expand the research field. As a growing number of development interventions impact on NCDs to achieve the SDGs, the evaluation of NCD outcomes and publication of results within academic literature is highly encouraged to promote best practice and inform effective resource allocation.
Supporting information
S1 Table. PRISMA checklist.
Reviews compliance to PRISMA guidelines.
https://doi.org/10.1371/journal.pone.0193378.s001
(DOCX)
S2 Table. Search strategy.
Full search strategy utilized for literature search.
https://doi.org/10.1371/journal.pone.0193378.s002
(DOCX)
S3 Table. Inclusion criteria.
Measurement and outcome criteria for review inclusion.
https://doi.org/10.1371/journal.pone.0193378.s003
(DOCX)
S4 Table. Quality assessment.
Quality assessment of included papers.
https://doi.org/10.1371/journal.pone.0193378.s004
(DOCX)
References
- 1. World Health Organisation. Global status report on noncommunicable diseases 2014. Noncommunicable diseases and mental health; 2014. Available from: http://www.who.int/nmh/publications/ncd-status-report-2014/en/
- 2. Islam SMS, Purnat TD, Phuong NTA, Mwingira U, Schacht K, Fröschl G. Non‐Communicable Diseases (NCDs) in developing countries: a symposium report. Globalization and Health. 2014;10(1). pmid:25498459
- 3.
Sachs JD. Investing in development: a practical plan to achieve the Millennium Development Goals. Overview. United Nations Millennium Project; 2005. Available from: http://www.who.int/hdp/publications/4b.pdf
- 4. World Health Organisation. Background Paper: Non-communicable diseases in low- and middle-income countries. Regional High-level Consultation in the Eastern Mediterranean Region on the Prevention and Control of Non-communicable Diseases in Low- and Middle-Income Countries, 2010. Available from: http://www.who.int/nmh/events/2010/Tehran_Background_Paper.pdf
- 5.
World Health Organization. Health in 2015: from MDGs, Millennium Development Goals to SDGs. Sustainable Development Goals. Switzerland: World Health Organization; 2015. Available from: http://www.who.int/gho/publications/mdgs-sdgs/en/
- 6. Naghavi M, Forouzanfar MH. Burden of non-communicable diseases in sub-Saharan Africa in 1990 and 2010: Global Burden of Diseases, Injuries, and Risk Factors Study 2010. The Lancet. 2013; 381.
- 7.
World Health Organisation. Background Paper: Raising the priority accorded to non-communicable diseases in development work and in related investment decisions. Side-event on Non-communicable Diseases on the occasion of the United Nations- Millenium Development Goals High-level Plenary Meeting, 2010. Available from: http://www.who.int/nmh/events/2010/ncd_background_paper_20100913.pdf
- 8.
World Health Organisation. Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013–2020. World Health Organisation. 2013. Available from: http://apps.who.int/iris/bitstream/10665/94384/1/9789241506236_eng.pdf?ua=1
- 9. World Health Organisation . From burden to "best buys": Reducing the economic impact of NCDs in low- and middle-income countries. Executive summary. 2011. Available from: http://www.who.int/nmh/publications/best_buys_summary/en/
- 10.
World Health Organisation, United Nations Development Programme. Catalyzing multisectoral action for NCDs. World Health Organisation. 2016. Available from: http://www.who.int/nmh/events/2015/ncd-multisector.pdf
- 11.
The World Bank. Country and Lending Groups: Analytic Classifications. 2015. Available from: http://data.worldbank.org/about/country-and-lending-groups.
- 12.
Fleiss J, Levin B, Paik M. Statistical Methods for Rates and Proportions. Hoboken: John Wiley & Sons, Inc; 2004: 598–626. https://doi.org/10.1002/0471445428.ch18
- 13. Higgins JPT, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaborations tool for assessing risk of bias in randomised trials. Bmj. 2011;343(oct18 2): d5928–d5928. pmid:22008217
- 14.
Wells GA, Shea B, O’Donnell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. The Ottawa Hospital Research Institute. 2000. Available at: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp
- 15. Morris SS, Medina Banegas JM. Rural development, household food safety, and nutrition in western Honduras. Arch Latinoam Nutr. 1999; 49(3): 244–52. pmid:10667264
- 16. Adamson GC, Polya DA. Critical pathway analysis to determine key uncertainties in net impacts on disease burden in Bangladesh of arsenic mitigation involving the substitution of arsenic bearing for groundwater drinking water supplies. J Environ Sci Health A Tox Hazard Subst Environ Eng. 2007;42(12): 1909–17. pmid:17952792
- 17. Kishtwaria J, Rana A. Intervention of gender friendly land preparation technologies for drudgery reduction of hill farm women. Work. 2012;41 Suppl 1: 4342–8. pmid:22317387
- 18. Low JW, Arimond M, Osman N, Cunguara B, Zano F, Tschirley D. A food-based approach introducing orange-fleshed sweet potatoes increased vitamin A intake and serum retinol concentrations in young children in rural Mozambique. J. Nutr. 2007; 137 (5): 1320–1327. pmid:17449599
- 19. Hotz C, Loechl C, deBrauw A, Eozenou P, Gilligan D, Moursi M, et al. A large-scale intervention to introduce orange sweet potato in rural Mozambique increases vitamin A intakes among children and women. Br J Nutr. 2012; 108(1): 163–76. pmid:22018075
- 20. Hotz C, Loechl C, Lubowa A, Tumwine JK, Ndeezi G, Nandutu Masawi A, et al. Introduction of β-carotene–rich orange sweet potato in rural Uganda resulted in increased vitamin A intakes among children and women and improved vitamin A status among children. J Nutr. 2012 Oct;142(10): 1871–80. pmid:22875553
- 21. Kidala D, Greiner T, Gebre-Medhin M. Five-year follow-up of a food-based vitamin A intervention in Tanzania. Public Health Nutr. 2000 Dec;3(4): 425–31. pmid:11135797
- 22. Langendorf C, Roederer T, de Pee S, Brown D, Doyon S, Mamaty AA, et al. Preventing Acute Malnutrition among Young Children in Crises: A Prospective Intervention Study in Niger. PLoS Med. 2014 Sep 2;11(9): e1001714. pmid:25180584
- 23. Marquis G, Colecraft E, Sakyi-Dawson O, Lartey A, Ben A, Birks K, et al. An Integrated Microcredit, Entrepreneurial Training, and Nutrition Education Intervention Is Associated with Better Growth Among Preschool-Aged Children in Rural Ghana. J. Nutr. 2015;jn.114.194498. pmid:25644356
- 24. Weiser SD, Bukusi EA, Steinfeld RL, Frongillo EA, Weke E, Dworkin SL, et al. Shamba Maisha: randomized controlled trial of an agricultural and finance intervention to improve HIV health outcomes. AIDS. 2015 Sep 10;29(14): 1889–94. pmid:26214684
- 25. Nair M, Ariana P, Ohuma EO, Gray R, De Stavola B, Webster P. Effect of the Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA) on malnutrition of infants in Rajasthan, India: a mixed methods study. PLoS One. 2013 Sep 25;8(9): e75089. pmid:24086447
- 26. Mascie-Taylor CG, Marks MK, Goto R, Islam R. Impact of a cash-for-work programme on food consumption and nutrition among women and children facing food insecurity in rural Bangladesh. Bull World Health Organ. 2010 Nov 1;88(11): 854–60. pmid:21076567
- 27. Olney DK, Talukder A, Iannotti LL, Ruel MT, Quinn V. Assessing impact and impact pathways of a homestead food production program on household and child nutrition in Cambodia. Food Nutr Bull. 2009 Dec;30(4): 355–69. pmid:20496626
- 28. Olney DK, Pedehombga A, Ruel MT, Dillon A. A 2-year integrated agriculture and nutrition and health behavior change communication program targeted to women in Burkina Faso reduces anemia, wasting, and diarrhea in children 3–12.9 months of age at baseline: a cluster-randomized controlled trial. J Nutr. 2015; 145(6): 1317–1324. pmid:25904734
- 29. Hanson M, Englberger L, Duncan B, Taren D, Mateak H, Johnson E. An evaluation of a nutrition intervention in Kapinga Village on Pohnpei, Federated States of Micronesia. Pac Health Dialog. 2011 Mar;17(1): 173–84. pmid:23008983
- 30.
Kariuki LW, Maundu PM, Morimoto Y. Some intervention strategies for promoting underutilised species: case of local vegetables in Kitui district, Kenya. II International Symposium on Underutilized Plant Species: Crops for the Future—Beyond Food Security, ISHS Acta Horticulturae 979. 2013. Available from: http://www.actahort.org/books/979/979_24.htm
- 31. Bamji MS, Rao MV, Satyanarayana G. Diversification from agriculture to nutritionally and environmentally promotive horticulture in a dry-land area. Sight and Life. 2011; 25 (1): 38–42.
- 32. Singh H, Kaur AP. Impact assessment of integrated nutrition garden concept for nutritional security and livelihood interventions under central Punjab conditions. Food Sci Res J. 2014; 5(1):51–55.
- 33. Sharma KR. Farm commercialization and nutritional status of children: the case of the vegetables, fruits, and cash crops programme in western Nepal. Food Nutr Bull. 1999; 20 (4): 445–453.
- 34. Jones KM, Specio SE, Shrestha P, Brown KH, Allen LH. Nutrition knowledge and practices, and consumption of vitamin A-rich plants by rural Nepali participants and nonparticipants in a kitchen-garden program. Food Nutr Bull. 2005; 26 (2): 198–208. pmid:16060221
- 35. Ahmed MM, Jabbar M, Ehui S. Household-level economic and nutritional impacts of market-oriented dairy production in the Ethiopian highlands. Food Nutr Bull. 2000; 21, 460–465.
- 36. Walingo MK. Socio-economic, food and nutrient intake and nutritional status indicators associated with successful livestock development programmes in Western Kenya. AJAR. 2012; 7 (2): 153–163.
- 37. E-Jahan KM, Ahmed M, Belton B. The impacts of aquaculture development on food security: lessons from Bangladesh. Aquacult Res. 2009; 41 (4): 481–495.
- 38. Ekesa BN, De Lange M, Macharia I, Garming H, Ouma E, Birachi E, et al. Impact of Interventions by Consortium for Improving Agriculture-based Livelihoods in Central Africa (CIALCA) on Food and Nutrition Security of Farmer Households. IGAE. 2013; 7(10): 14–22. Doi: urn:dai:10.1999/1307-6892/16930.
- 39. Abebaw D, Fentie Y, Kassa B. The impact of a food security program on household food consumption in Northwestern Ethiopia: A matching estimator approach. Food Policy. 2010; 35(4):286–293.
- 40. Remans R, Pronyk PM, Fanzo JC, Chen J, Palm CA, Nemser B. Multisector intervention to accelerate reductions in child stunting: an observational study from 9 sub-Saharan African countries. Am J Clin Nutr. 2011 Dec;94(6):1632–42. pmid:22030229
- 41. Hop IT. Programs to Improve Production and Consumption of Animal Source Foods and Malnutrition in Vietnam. J Nutr. 2003 Nov;133(11 Suppl 2):4006S–4009S.
- 42. Moench-Pfanner R, de Pee S, Bloem MW, Foote D, Kosen S, Webb P. Food-for-work programs in Indonesia had a limited effect on anemia. J Nutr. 2005 Jun;135(6):1423–9. pmid:15930447
- 43. Asfaw A. Do government food price policies affect the prevalence of obesity? Empirical evidence from Egypt. World Development. 2007; 35(7):687–701.
- 44. Ruel MT, Alderman H. Nutrition-sensitive interventions and programmes: How can they help to accelerate progress in improving maternal and child nutrition? Lancet. 2013; 382(9891): 536–551. pmid:23746780
- 45.
Panchamukhi PR. Agricultural diversification as a tool of tobacco control. The WHO International Conference on Global Tobacco Control Law: Towards a WHO Framework Convention on Tobacco Control, 2000. Available from: http://www.rctfi.org/resources/Agricultural_Diversification_as_a_Tool_of_Tobacco_Control.pdf
- 46. Wahab A, Hod R, Ismail NH, Omar N. The effect of pesticide exposure on cardiovascular system: a systematic review. Int J Community Med Public Health. 2016; 3(1): 1–10.
- 47. Silva DA, Nunes HE. Prevalence of underweight, overweight and obesity in poor children from Mato Grosso do Sul. Rev Bras Epidemiol. 2015 Apr-Jun;18(2):466–75. pmid:26083516
- 48. Fernald LC, Gertler PJ, Hou X. Cash component of conditional cash transfer program is associated with higher body mass index and blood pressure in adults. J Nutr. 2008 Nov;138(11):2250–7. pmid:18936227
- 49. Lagarde M, Haines A, Palmer N. The impact of conditional cash transfers on health outcomes and use of health services in low and middle income countries. Cochrane Database Syst Rev. 2009 Oct 7;(4):CD008137. pmid:19821444
- 50.
United Nations Food and Agricultural Organisation. Guidelines for measuring household and individual dietary diversity. FAO Corporate Document Repository, 2012. Available from: http://www.fao.org/docrep/014/i1983e/i1983e00.htm
- 51. Darnton-Hill I, Nishida C, James WP. A life course approach to diet, nutrition and the prevention of chronic diseases. Public Health Nutr. 2004 Feb;7(1A):101–21. pmid:14972056
- 52. Forouzanfar MH, Alexander L, Anderson RH, Bachman VF, Biryukov , et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015; 386(10010):2287–2323. pmid:26364544
- 53. Allen L, Williams J, Townsend N, Mikkelsen B, Roberts N, Foster C, et al. Socioeconomic status and non-communicable disease behavioural risk factors in low-income and lower-middle-income countries: a systematic review. Lancet Glob Health. 2017; 5(3):e277–e289. pmid:28193397
- 54. Waddington H, White H, Snilstveit B, Hombrados J, Vojtkova M, Davies P, et al. How to do a good systematic review of effects in international development: a tool kit. J DEV EFFECT. 2012; 4(3): 359–387.
- 55. Petruney T. Facing global development’s failure. Devex. 2014. Available from: https://www.devex.com/news/facing-global-development-s-fear-of-failure-85078.
- 56.
Mc Cafrey M, Strobel F, Jensen T. Pacific trade and health. UNDP, WPRO & OHCHR. 2014. Available from: http://www.wpro.who.int/southpacific/publications/trade_and_human_rights.pdf?ua=1.
- 57. Duvendack M, Hombrados JG, Palmer-Jones R, Waddington H. Assessing ‘what works’ in international development: meta-analysis for sophisticated dummies. J Dev Eff. 2012; 4(3):456–471.
- 58.
World Health Organistaion. Promoting healthy growth and preventing childhood stunting. Wiley Blackwell, 2013. Available from: http://www.who.int/nutrition/healthgrowthproj_maternalchildnut_journal/en/
- 59. Garcia-Casa MN, Pena-Rosas JP, Pachon H, De-Regil LM, Tablante EC, Flores-Urrutia MCF. Staple crops biofortified with increased micronutrient content: effects on vitamin and mineral status, as well as health and cognitive function in the general population. Cochrane Syst Rev. 2016.
- 60. Dangour AD, Watson L, Cumming O, Boisson S, Che Y, Velleman Y, et al. The effect of interventions to improve water quality and supply, provide sanitation and promote handwashing with soap on physical growth in children. Cochrane Syst Rev, 8, CD009382. 2013.