OBJECTIVE: To evaluate the impact of the "Leite é Saúde" program (PLS) in recovering the nutritional status of malnourished children in Rio de Janeiro (Rio) during the 1994/1996 period. METHOD: The evaluation was based on target children (6-23 months of age and weight-for-age < 3rd centile of the NCHS curve (wap<3) at enrollment) followed up at least for one year and for whom at least one information on weight was available for every quarter of the first year of follow-up (n=748). Impact was measured by a positive variation in weight-for-age NCHS z-scores. Regression to the mean effect was corrected by applying the correlation coefficients observed between sequential measures of well-nourished American children to z score at entrance. Age and nutritional status at enrollment, previous and current health status, presence of 24-59 month-old siblings, provision of soy oil as an additional food supplement, and level of complexity of the health unit were studied as possible intervenient factors. RESULTS: The mean variation of the weight-for-age z-score after the one year follow-up was +0.464 (IC95%=0.41-0.52), with similar gains in the first and second follow-up semesters (+0.245 and +0.219, respectively; p=0.33). Impact of the program was greater on children with a greater weight deficit at entrance; age at entrance did not significantly influence the response to the program. None of the remaining intervenient variables studied influenced the success of the intervention. CONCLUSION: The findings indicate that PLS was effective in recovering malnourished children. The internal validity of the study reinforces this conclusion. The fact that the program evaluated was implemented in the usual conditions of the city's health units and the fact that the impact was similar in health units with different levels of complexity point out the possibility of generalizing the results to other urban settings similar to Rio. The fact that the PLS implemented in Rio employed a stricter criterion (wap<3) than the original PLS criterion (weight-for-age < 10th centile of NCHS curve) to select beneficiaries argues against the generalization of the study, at least in terms of the magnitude of the impact achieved by the program.
Child malnutrition; Nutritional recovery; Supplementary feeding program; Impact evaluation; Rio de Janeiro, Brazil
