Prediction of Developmental Outcomes in Preterm Infants: A Systematic Review

CONTEXT Increased survival of very preterm children raises concerns about later outcomes. Prediction models for developmental outcomes can support perinatal decision-making. OBJECTIVE Summarizing prediction models for developmental outcomes after very preterm birth. DATA SOURCES Medline, Embase, Web of Science, and the Cochrane database were searched for publications up to June 2024. STUDY SELECTION All developed or externally validated prognostic models predicting developmental outcomes of infants born very preterm (<32 weeks) or with a very low birth weight (≤1500 g) were included. DATA EXTRACTION Data extraction and risk of bias (RoB) assessment were done independently and in duplicate by 2 researchers using the Prediction Model Risk of Bias Assessment Tool. RESULTS Thirty-eight publications were identified describing 99 (85%) newly developed models and 18 (15%) external validations of 11 different existing models. The most predicted outcome was a composite measure of neurodevelopmental impairment, with prediction horizons ranging from 18 months to 10 years after birth. RoB was high for all models, mostly because of concerns regarding outcome and data analysis. LIMITATIONS Because of a high RoB among all included studies, no recommendation for the most accurate prediction model could be provided. CONCLUSIONS Although many prediction models for developmental outcomes after preterm birth exist, external validation is rare, and RoB is high. Future research is needed to validate and, if needed, update existing models to support perinatal decision-making.