The predictive validity of HINE, Bayley, general movements and MOS-R in infancy

Introduction: Infants born with congenital anomalies requiring surgery are at greater risk of developmental delays. Early screening tools are needed to identify infants who would benefit from early intervention. This study aimed to investigate the concurrent predictive validity of the General Movements Assessment (GMA), Motor Optimality Score - Revised (MOS-R), Hammersmith Infant Neurological Examination (HINE) and Bayley-III in identifying infants at risk of adverse neurodevelopmental outcomes. Methods: A retrospective cohort study of 95 surgical infants. Participants were assessed at 3 months using the GMA, MOS-R, HINE, and Bayley-III. Development was assessed at 1 year using the Bayley-III. Logistic regression investigated the relationship between measures at 3 months and developmental outcome, using sensitivity, specificity, positive (PPV) and negative (NPV) predictive value and area under the ROC curve (AUC). Results: All assessments had a relationship with development. Combining assessments at 3 months did not increase predictive value. MOS-R < 23 or HINE<60 had >70 % sensitivity for identifying delays in >= 2 domains of the Bayley-III at 1 year, and > 80 % accuracy in ruling out infants not at risk. MOS-R < 23 (NPV 0.957), and HINE <60 (NPV 0.971) were significantly correlated with delayed cognition. Discussion: In a surgical cohort, the concurrent predictive validity of the GMA, MOS-R, HINE and Bayley-III, administered at 3 months of age was low, however all assessment tools showed correlations with outcomes. The HINE and MOS-R might help to identify infants who need support with cognition. Further exploration of early screening tools may help to identify infants who would benefit from early intervention.