Child malnutrition is a major problem confronted by developing countries. The problem of undernutrition is not that of food production alone but of food supply. It is believed that agricultural interventions that are aimed at increasing household income can in turn provide better access to improved dietand nutrition for children. Some examples of these interventions are biofortification, dairy development and the promotion of vegetable gardens.
Previous reviews have assessed the effects of such agricultural interventions targeting income increases amongst the poor for better nutritional gains amongst children. However, because of difference in timings and methodologies adopted in these studies, their results show mixed or no effects of agricultural interventions. The aim of the present review was to shed light on the effectiveness of agricultural interventions.
To systematically synthesise and measure the impact of agricultural interventions on child nutrition.
The authors included all randomised and nonrandomised trials written in English assessing the effect of agricultural interventions on child nutritional outcomes in developing countries. As a minimum quality criterion, studies not using comparison groups or cross-sectional comparisons without any further attempt to correct for selection bias were excluded. Included studies were broken down into five types of intervention: biofortification (both conventional breeding and genetic engineering), home vegetable gardens, dairy development, fisheries (aquaculture and small-scale fisheries), and animal husbandry (including poultry).
The authors conducted a systematic search of the published and unpublished literature between 1990 and 2010, including among other databases EconLit, IDEAS, and PubMed. The search also involved reference checking and contacting authors. The authors critically assessed the quality of the studies. The validity assessment tool for included studies comprised four indicators: counterfactual analysis, power, intermediate outcomes and heterogeneity, and scored study quality as low, medium or high. The authors present a causal chain for the pathways of impacts of agricultural interventions. Next, they group the studies by intervention type and outcome across the causal chain and synthesise the results using narrative synthesis and meta-analysis for the studies assessing vitamin A intake.
The authors included 23 studies from Africa, Southeast Asia and South Asia involving different agricultural interventions including biofortification, home gardening, dairy development, fisheries and animal husbandry.
- The authors find that the agricultural interventions included in the review increased food production, especially for food items encouraged by these interventions, with 13 studies showing positive effects. However it is difficult to gauge impact on household income because of the poor quality of the available studies.
- The authors also obtained results from included studies that show that calorie consumption and food consumption do not change much in response to changes in income.
- Agricultural interventions increased consumption of milk, fish and vegetables, but the impact on diet as a whole is not known. Studies assessing vitamin A intake and those that assessed biofortification showed positive effects, although only a small number of studies confirmed these results. Agricultural interventions were found to have reduced wasting and underweight, both shorter-term effects of malnutrition, compared with stunting, which is a long-term effect of malnutrition.
- Overall, the review results are in line with those of previous reviews that suggest that there is little or no evidence of an impact of agricultural interventions on the nutrition of children. However, the authors clarify that this is not necessarily because the interventions are not working. Indeed, the lack of rigour in methods and insufficient sample sizes used in these studies means that the evidence base is not sufficient to make recommendations to decision makers.
The results of these studies cannot be generalised, because none of them presents information on participation rates, which partly informs about the characteristics of vulnerable groups. Most of the studies did not have sufficient statistical power and did not measure relevant information among nonparticipants. The authors recommend using more rigorous methods for further evaluations in this field.
The systematic review is based on relatively comprehensive searches for literature and appropriate methods to reduce risk of bias in terms of study selection and analysis. However, there are some limitations. The authors did not avoid language bias, as they included articles in English only. The included studies are fairly heterogeneous in terms of rigour of design, but the authors do not present any subgroup analysis by study quality.