Introduction: High-risk NICU graduates are vulnerable to early neurodevelopmental delays due to cumulative prenatal, perinatal, and neonatal complications. Early identification of gross motor delay is essential for timely intervention, yet standardized predictive tools remain limited in clinical practice. Developing a reliable regression-based predictive model may support clinicians in screening infants at risk during early follow-up visits.

Objective: To develop and validate a predictive model capable of estimating 6-month gross motor outcomes using significant neonatal and perinatal clinical risk factors.

Method: A cross-sectional analytical study was conducted on 284 high-risk NICU infants. Thirty-one potential risk factors were screened, of which 17 demonstrated significant correlation with the 6-month Gross Motor score of the Ages and Stages Questionnaire (ASQ-3). Stepwise linear regression was performed, initially identifying 11 significant predictors, and subsequently refining them to 8 independent variables that contributed significantly to the final predictive model. Predicted ASQ-3 scores were categorized as High Risk or No Risk and compared with actual scores. Diagnostic accuracy indices—Sensitivity, Specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), and Overall Accuracy—were calculated.

Result: The eight significant predictors included consanguineous marriage, birth asphyxia, duration of oxygen support, neonatal seizures, periventricular leukomalacia (PVL), respiratory distress syndrome (RDS), history of miscarriage, and low Apgar score. The predictive model demonstrated Sensitivity 60%, Specificity 85.77%, PPV 62.33%, NPV 84.54%, and Overall Accuracy 78.52%, indicating good discrimination for identifying infants at risk of gross motor delay.