Growing up in a poor neighbourhood has negative effects on children and adolescents. In the literature it has been concluded that the risk of low birth weight, childhood injury and abuse, and teenage pregnancy or criminality double in poor areas. However, the validity of such studies has been questioned, as they have been associated with ecological or individualistic fallacies. Studies using multilevel technique might thus contribute important knowledge in this field. The present review clarifies the importance of neighbourhood contextual factors in child and adolescent health outcomes, through considering only studies using multilevel technique. Keyword searching of the Medline, ERIC, PsycInfo, Sociological Abstracts, and Social Citation Index databases was performed. Original studies using multilevel technique to examine the effect of neighbourhood characteristics on child and adolescent health outcomes, and focusing on populations in high-income countries were included. Neighbourhood socioeconomic status and social climate were shown to have small to moderate effects on child health outcomes, i.e. birth weight, injuries, behavioural problems, and child maltreatment. On average, 10% of variation in health outcomes was explained by neighbourhood determinants, after controlling for important individual and family variables. This review demonstrates that interventions in underprivileged neighbourhoods can reduce health risks to children, especially in families that lack resources. An analysis of methodological fallacies indicates that observed effects and effect sizes can be underestimated, and that interventions may well have greater impact than this review was able to establish.