Community-level interventions for improving access to food in low- and middle-income countries

After decades of decline since 2005, the global prevalence of undernourishment reverted and since 2015 has increased to levels seen in 2010 to 2011. The prevalence is highest in low- and middle-income countries (LMICs), especially Africa and Asia. Food insecurity and associated undernutrition detrimentally affect health and socioeconomic development in the short and long term, for individuals, including children, and societies. Physical and economic access to food is crucial to ensure food security. Community-level interventions could be important to increase access to food in LMICs.

To determine the effects of community-level interventions that aim to improve access to nutritious food in LMICs, for both the whole community and for disadvantaged or at-risk individuals or groups within a community, such as infants, children and women; elderly, poor or unemployed people; or minority groups.

We searched for relevant studies in 16 electronic databases, including trial registries, from 1980 to September 2019, and updated the searches in six key databases in February 2020. We applied no language or publication status limits.

We included randomised controlled trials (RCTs), cluster randomised controlled trials (cRCTs) and prospective controlled studies (PCS). All population groups, adults and children, living in communities in LMICs exposed to community-level interventions aiming to improve food access were eligible for inclusion. We excluded studies that only included participants with specific diseases or conditions (e.g. severely malnourished children). Eligible interventions were broadly categorised into those that improved buying power (e.g. create income-generation opportunities, cash transfer schemes); addressed food prices (e.g. vouchers and subsidies); addressed infrastructure and transport that affected physical access to food outlets; addressed the social environment and provided social support (e.g. social support from family, neighbours or government).

Two authors independently screened titles and abstracts, and full texts of potentially eligible records, against the inclusion criteria. Disagreements were resolved through discussion or arbitration by a third author, if necessary. For each included study, two authors independently extracted data and a third author arbitrated disagreements. However, the outcome data were extracted by one author and checked by a biostatistician. We assessed risk of bias for all studies using the Effective Practice and Organization of Care (EPOC) risk of bias tool for studies with a separate control group. We conducted meta-analyses if there was a minimum of two studies for interventions within the same category, reporting the same outcome measure and these were sufficiently homogeneous. Where we were able to meta-analyse, we used the random-effects model to incorporate any existing heterogeneity. Where we were unable to conduct meta-analyses, we synthesised using vote counting based on effect direction.

We included 59 studies, including 214 to 169,485 participants, and 300 to 124, 644 households, mostly from Africa and Latin America, addressing the following six intervention types (three studies assessed two different types of interventions). Interventions that improved buying power: Unconditional cash transfers (UCTs) (16 cRCTs, two RCTs, three PCSs): we found high-certainty evidence that UCTs improve food security and make little or no difference to cognitive function and development and low-certainty evidence that UCTs may increase dietary diversity and may reduce stunting. The evidence was very uncertain about the effects of UCTs on the proportion of household expenditure on food, and on wasting. Regarding adverse outcomes, evidence from one trial indicates that UCTs reduce the proportion of infants who are overweight. Conditional cash transfers (CCTs) (nine cRCTs, five PCSs): we found high-certainty evidence that CCTs result in little to no difference in the proportion of household expenditure on food and that they slightly improve cognitive function in children; moderate-certainty evidence that CCTs probably slightly improve dietary diversity and low-certainty evidence that they may make little to no difference to stunting or wasting. Evidence on adverse outcomes (two PCSs) shows that CCTs make no difference to the proportion of overweight children. Income generation interventions (six cRCTs, 11 PCSs): we found moderate-certainty evidence that income generation interventions probably make little or no difference to stunting or wasting; and low-certainty evidence that they may result in little to no difference to food security or that they may improve dietary diversity in children, but not for households. Interventions that addressed food prices: Food vouchers (three cRCTs, one RCT): we found moderate-certainty evidence that food vouchers probably reduce stunting; and low-certainty evidence that that they may improve dietary diversity slightly, and may result in little to no difference in wasting. Food and nutrition subsidies (one cRCT, three PCSs): we found low-certainty evidence that food and nutrition subsidies may improve dietary diversity among school children. The evidence is very uncertain about the effects on household expenditure on healthy foods as a proportion of total expenditure on food (very low-certainty evidence). Interventions that addressed the social environment: Social support interventions (one cRCT, one PCS): we found moderate-certainty evidence that community grants probably make little or no difference to wasting; low-certainty evidence that they may make little or no difference to stunting. The evidence is very uncertain about the effects of village savings and loans on food security and dietary diversity. None of the included studies addressed the intervention category of infrastructure changes. In addition, none of the studies reported on one of the primary outcomes of this review, namely prevalence of undernourishment.

The body of evidence indicates that UCTs can improve food security. Income generation interventions do not seem to make a difference for food security, but the evidence is unclear for the other interventions. CCTs, UCTs, interventions that help generate income, interventions that help minimise impact of food prices through food vouchers and subsidies can potentially improve dietary diversity. UCTs and food vouchers may have a potential impact on reducing stunting, but CCTs, income generation interventions or social environment interventions do not seem to make a difference on wasting or stunting. CCTs seem to positively impact cognitive function and development, but not UCTs, which may be due to school attendance, healthcare visits and other conditionalities associated with CCTs.