To test the impact of training for educators on the health of children enrolled in public and philanthropic day-care nurseries.

A randomised, institutional, non-blind clinical trial was conducted. An educational intervention was performed in four day-care centres and the control group consisted of four other day-care centres. Interviews with the mothers, collection of blood from the children by digital puncture and anthropometry were performed. The chosen indicator for the improvement of health was anaemia (Hb <11 g/dl). An unconditional logistic regression model was set for the risk factors for anaemia, considering associations with P ≤ 0·05 as statistically significant.

Eight day-care centres in the city of Sao Paulo, Brazil.

Two hundred and fifty-two children from day-care nurseries.

The children from the day-care centres that were not subject to intervention presented a 2·11 times greater risk (95% CI 1·04, 4·30; P = 0·40) of having anaemia at the end of the study independent of the control variables (sex, age, time in the day-care centre, anaemia at the beginning of the study, maternal age, use of oral iron supplements, number of siblings, per capita family income, use of antibiotics and the necessity of avoidable hospitalisations) used in the construction of the final logistical model.

The assessed educational intervention promoted significant changes in the health status of the children, reinforcing the importance of training for professionals who care for young children in day-care centres in developing countries in order to promote child health.