An analysis undertaken jointly in 2009 by the UN World Food Programme, The Partnership for Child Development and the World Bank was published as Rethinking School Feeding to provide guidance on how to develop and implement effective school feeding programmes as a productive safety net and as part of the efforts to achieve Education for All. The present paper reflects on how understanding of school feeding has changed since that analysis.

Data on school feeding programme outcomes were collected through a literature review. Regression models were used to analyse relationships between school feeding costs (from data that were collected), the per capita costs of primary education and Gross Domestic Product per capita. Data on the transition to national ownership, supply chains and country examples were collected through country case studies.

School feeding programmes increase school attendance, cognition and educational achievement, as well as provide a transfer of resources to households with possible benefits to local agricultural production and local market development. Low-income countries exhibit large variations in school feeding costs, with concomitant opportunities for cost containment. Countries are increasingly looking to transition from externally supported projects to national programmes.

School feeding is now clearly evident as a major social programme in most countries with a global turnover in excess of $US 100 billion. This argues for a continuing focus on the evidence base with a view to helping countries ensure that their programmes are as cost-effective as possible. Clear policy advice has never been more important.

To describe food consumption during the school day of rural Jamaican children and participation in two government school feeding programmes. To determine factors which were related to these.

Cross sectional.

16 primary schools in rural Jamaica.

415 children in grades 2 and 5 (ages 7 and 10 years).

Consumption of sweets, sweet drinks and snacks was high. Mean intakes at lunch were: energy 1537 kJ (SD 756), protein 10.4 g (SD 7.6) and iron 1.5 mg (SD 1.2). The mean energy intake was 17–20% of the daily requirement for this age group. Two types of school feeding programmes were available in the schools, one provided a cooked meal and the other a bun and milk. Median availability of school meals (as a percentage of children enrolled in the schools) over three terms was 24.6% (range 0–85.4%). Twenty per cent of the children participated in one or other programme. Poorer children were more likely to participate in the bun and milk programme (odds ratio 2.1, 95% C1 1.3–3.5) but children with more money to purchase food participated in the more costly cooked meal programme (odds ratio 2.4, 95% CI 1.3–4.6).

Energy intakes at lunch in Jamaican children were somewhat below optimal levels and the reliance on sweets and snacks is an area of concern. Programme characteristics such as meal cost, may affect access to school feeding by poor children.