RESUMO
Programs providing cash transfers to poor families, conditioned upon uptake of preventive health services, are common in Latin America. Because of the consistent association between undernutrition and poverty, and the role of health services in providing growth promotion, these programs are supposed to improve children's growth. The impact of such a program was assessed in 4 municipalities in northeast Brazil by comparing 1387 children under 7 y of age from program beneficiary households with 502 matched nonbeneficiaries who were selected to receive the program but who subsequently were excluded as a result of quasi-random administrative errors. Anthropometric status was assessed 6 mo after benefits began to be distributed, and beneficiary children were 0.13 Z-scores lighter (weight-for-age) than excluded children, after adjusting for confounders (P = 0.024). The children's growth trajectories were reconstructed by copying up to 10 recorded weights from their Ministry of Health growth monitoring cards and by relating each weight to the child's age, gender, and duration of receipt of the program benefit in a random effects regression model. Totals of 472 beneficiary and 158 excluded children under 3 y of age were included in this analysis. Each additional month of exposure to the program was associated with a rate of weight gain 31 g lower than that observed in excluded children of the same age (P < 0.001). This failure to respond positively to the program may have been due to a perception that benefits would be discontinued if the child started to grow well. Nutrition programs should guard against giving the impression that poor growth will be rewarded.