A machine learning model using 12 clinical and demographic variables flagged infants likely to need hip ultrasound before imaging, with strong external validation at a second hospital. The approach targets early detection of infant hip dysplasia to reduce unnecessary ultrasounds while ensuring high-risk cases are promptly identified. The work adds to a growing set of AI triage tools aiming to improve care efficiency in pediatric screening programs. External validation helps address a key deployment risk: performance drop-off across sites. If adopted, the model could streamline workflow in ultrasound-dependent pathways and support more consistent decision-making across care settings.