Berti, P. R., Krasevec, J. and FitzGerald, S. (2004) A review of the effectiveness of agriculture interventions in improving nutrition outcomes. Public Health Nutrition, 7(5), pp. 599Link to Source
The authors critically review and synthesise available evidence on the nutritional outcomes of 30 agricultural interventions from low- and middle-income countries in Asia, Africa and Latin America. They analyse agricultural and nutritional outcomes as well as the effects of investing in human, financial, physical, natural and social capital. The review confirms previous findings that agricultural interventions improve agricultural outcomes. It yields mixed results on nutritional outcomes, with studies showing positive, negative and no effects, in addition to discrepancies in outcomes across different health indicators. Interventions investing in multiple types of capital in addition to the agricultural interventions were found to be more effective at improving nutritional outcomes than studies with more restricted focuses. Home gardening interventions promoting human capital, for instance through nutritional education and gender consideration, were more successful than any other type of intervention. Overall, the results highlight the importance of nutrition education in improving nutritional status, although nutritional improvements were also observed for some interventions without education components. Five of the nine studies that measured the long-term effects of the interventions (4 to 30 years after the interventions) found some long-term sustainability of effects. Contrary to expectations, these effects were not always dependent on investment and impact on financial capital.
The authors highlight that the analysis and generalisability of the results are limited by studies adopting designs that are not suitable for assessing the impact of agricultural interventions on nutritional status. The included studies also assessed the effects of interventions with different objectives and inputs, making comparisons challenging. Further research is needed in a variety of settings and interventions, and this should be conducted by multidisciplinary teams.
When agricultural interventions are designed and implemented well, they can have positive effects on productivity and food availability in developing countries. It is often assumed that these productivity and food availability improvements translate into improved nutritional status. Nevertheless, there is not much research assessing the impact of agricultural interventions such as home gardening, livestock, mixed gardening and livestock, cash cropping and irrigation on nutritional status.
To review and synthesise the evidence on the effectiveness of various agricultural interventions on improving the nutritional status of households in developing countries and to assess the programme characteristics leading to the greatest nutritional improvements
The authors included pre-post studies without comparison groups, and studies with comparison groups that assessed the effects of agricultural interventions on nutritional outcomes in developing countries. They searched both published and unpublished literature, including electronic databases such as MEDLINE, Current Contents, BIOSIS Previews, PASCAL, AGRIS and the University of Ottawa catalogue Orbis, the Web sites of relevant organisations and the personal libraries of colleagues, and they tracked the references of reviews and other relevant papers. The search covered the period between 1985 and 2001. The authors weighed studies on a methodological quality scale and analysed the results using narrative synthesis and thematic analysis. Because of heterogeneity of indicators and outcomes, they did not conduct a meta-analysis.
The review uses appropriate methods to reduce risk of bias in terms of a reasonably comprehensive search strategy and methods of data analysis. However, the review has major limitations. The authors do not explicitly state all inclusion criteria, and they do not report how articles were selected for inclusion, so it is not clear that bias was avoided. The criteria used for assessing risk of bias are not sufficiently transparent, and the authors report limited results of the quality appraisal. This is a particular concern, as the review includes a number of study designs with very high risk of bias. However, this is partially mitigated by the authors' acknowledgement of some of these limitations.