Objective: To investigate psychosocial characteristics of children and parents as predictors and moderators of the effect of a group intervention for children with chronic illness and their parents. Methods: Data from a randomized controlled trial were used, including 194 children (8-18 years) who were assigned to a child-only intervention, a parent-child intervention, or a wait-list control group. Longitudinal multilevel regression analyses were used to test effects on change in parent and child reported internalizing and externalizing behavior problems. Results: For children with a more disengaged coping style or lower self-worth and for children who experienced a more secure parent-child relationship, the parent-child intervention was more effective than the child-only intervention in reducing behavior problems. Conclusions: Children who are more ''at risk'' appear to gain more from participating in an intervention, especially if their parents are involved as well. However, the benefit of parents' involvement may depend on the quality of the parent-child relationship.