Little is documented about the determinants of developmental outcomes for medically fragile infants who receive early intervention. In this controlled longitudinal study 65 premature infants with intraventricular hemorrhage (IVH) were randomly assigned to intervention groups beginning at 3 months adjusted age (Early) or 12 months adjusted age (Delayed). The sample was 65% African American and 35% Caucasian, and over half the youngsters were being raised by single mothers. Although cost analyses revealed that it was almost twice as expensive to begin sensorimotor intervention at the earlier age, annual comprehensive assessments revealed no significant differences in developmental outcomes based on age at start. Stepwise multiple regression analyses revealed that Maternal Education and Neonatal Medical Problems were significant predictors of outcomes at years 1, 3, 5, and 7 regardless of age at start. Related findings from other studies are discussed along with implications for policy and future research.