To evaluate the effectiveness of a school nutrition programme on the weight gain and growth of Vietnamese schoolchildren.

A proximate cluster evaluation of children in seven schools, in which fortified milk and biscuits supplying 300 kcal of energy were being given on school days, compared with children in 14 nearby schools with no feeding. All children were dewormed.

Twenty-one primary schools in Dong Thap Province, Vietnam.

A cohort of 1080 children in grade 1 of 21 primary schools, and a cross-sectional interview of 400 children in grade 3.

The programme gave children the equivalent of 90 kcal day− 1 over 17 months. t-Tests showed a small but statistically significant difference between groups in their average gain in weight and height: 3.19 versus 2.95 kg (P < 0.001) and 8.15 versus 7.88 cm (P = 0.008). A multiple-level model showed that the programme was statistically significant after controlling for clustering of children in schools, sex, age and initial underweight (P = 0.024). A significant impact on height was also seen in a regression model, but not when controlling for school. The most undernourished children tended to gain the least weight. There was no evidence of substitution.

The programme had a small but significant effect on weight gain, but the most undernourished children benefited the least. Methods need to be developed to target them. This design may offer a means of estimating the impact of school feeding on growth in other programme settings.