This population-based prospective cohort study compared the risk of recurrent infections in children attending family day care ( < 3 children per family), small (10-20 children) day-care centers (DCCs), and large ( > 40 children) DCCs. The parents of a total of 1,242 children participated in the study (97% of the families initially contacted). An infectious episode was defined as the acute occurrence of a new symptom lasting for at least 48 h and resulting in specific treatment. Two episodes were counted as such only if they were separated by a symptom-free week. Surveillance was under the responsibility of a nursing director and was similar for all three types of DCCs. During the 8.5-month follow-up period, 3,639 infectious episodes were recorded. Compared to those in family day-care, children attending small DCCs presented a higher risk for > 6 total infectious episodes [odds ratio (OR) = 2.4; 95% confidence interval (CI) = 1.6-3.7]; > 5 upper respiratory tract infections (OR = 2.2; 95% CI = 1.4-3.4); > 2episodes of otitis media (OR = 2.6; 95% CI = 1.0-2.6); > 2 episodes of conjunctivitis (OR = 4.1; 95% CI = 2.1-8.2); and > 2 episodes of croup (OR = 4.1; 95% CI = 1.6-10.9). The risk for children attending large DCCs was intermediate between those in family day care and those in small DCCs. Apart from sampling variation, one explanation for this result could be that children in large DCCs are divided into groups according to their age (i.e. < 12,12-24, and > 24 months). It is possible that the homogeneity of age within each group and the absence of direct contact between groups confers some protection against the spread of infections. Children who had been in day care for at least 6 months at the beginning of the study were at a lower risk for recurrent infections than those who had entered day care earlier. This result might be explained by the acquisition of specific immunity as well as by nonspecific immunity that protects against microorganisms not previously encountered by the body. These results suggest that, for children with repeated infections in DCCs, a move to the family day-care setting, when feasible, should be contemplated. It also suggests that the development of family day-care settings should be encouraged. However, the decision of promoting one type of day-care setting rather than another requires further studies focusing on different outcomes such as the long-term health consequences, the psychological development, and also the total economic consequences related to attendance of each type of structure.