Background and Aims: Premature infants are at risk for later neurodevelopmental impairments. In an attempt to decrease these impairments, numerous early intervention programs target child development. These programs involve either teaching parents skills and/or involving parents in aspects of care for their infant. We conducted a systematic review to determine whether such early interventions improve neurodevelopmental outcomes of premature infants.Methods: MEDLINE, CINAHL and Cochrane Controlled Central Registry from 1966 to 2004 were searched. We included randomized controlled trials of premature infants (< 37 weeks gestation) that conducted an intervention in the first year of life and involved the parent, as compared to routine care. The primary outcome measure was Bayley neurodevelopmental outcome at 12 and 24 months corrected age. Two independent reviewers extracted data. Meta-analyses were performed using weighted mean difference (random effects model) with 95% confidence intervals.Results: 57 articles met the inclusion criteria. Of these articles, only 9 studies assessed Bayley at 12 months and 6 studies assessed Bayley at 24 months. These studies used a variety of interventions including parental education, home visits by professionals to assist/teach parents, or individualized developmental care (NIDCAP) involving parents. Meta-analysis of Bayley MDI performed at 12 months found a mean change of 8.20 [n= 1027, CI 2.81–13.58, p=0.003]. At 24 months a mean change of 7.91 [n=859, CI 3.28–12.53, p=0.0008] was found. Similar results were found for analyses that excluded NIDCAP studies (n=2).Conclusion: This review suggests that early interventions, which target the parent as well as the infant, may improve neurodevelopment of preterm infants. In order to make a firm recommendation for implementing early interventions, a large, adequately powered, well-designed randomized trial needs to be conducted to assess short and long-term neurodevelopmental outcomes. This should also incorporate cost effectiveness analysis.T
