Objective: The purpose of this study was to investigate retrospectively neonatal health and maternal background among a sample of children taken into custody and placed in foster care and to investigate the relation between medical and social risk in the neonatal period. Method: The data-linkage study combined two registries: the Finnish Medical Birth Registry (MBR), from January 1, 1987 to December 31, 1989, and the Finnish Child Welfare Registry (CWR) from January 1, 1987 to December 31, 1997 using personal identification numbers (n =1,668). As controls all Finnish children born in 1987 were chosen (n = 59,727). Infant neonatal health was estimated using the following indicators: birth-weight and birth-length, birth-weight by gestational age, 1-minute Apgar scores, and discharge age from the nursery. Maternal background characteristics included maternal age, parity, marital status, number of antenatal visits at maternity clinics and smoking during pregnancy. Results: We found that the 1,668 children in the study had a lower birth-weight and birth-length, shorter gestational age, lower 1-minute Apgar scores and later discharge from the nursery than the population-based controls. The proportion of teen-age mothers was about four times higher and the proportion of unmarried women was twice as high. A majority of the women (56%) reported smoking during pregnancy, compared with only 15% of the population-based comparisons. Children subsequently taken into custody had poorer health at the time of birth than other children and these differences could only partly be explained by the differences in smoking habits during pregnancy. The women in the study made use of the free health controls at maternity clinics as much as the population-based comparisons. Conclusions: Our findings indicate that social and medical risks are related in the neonatal period. Since almost all mothers visit maternity clinics, there are possibilities for early intervention.