Toward governance of artificial intelligence in pediatric healthcare

Area of Science:

• Healthcare technology
• Pediatric medicine
• Artificial intelligence governance

Background:

• Artificial intelligence (AI) is revolutionizing healthcare delivery.
• Pediatric adoption of AI tools and associated governance frameworks remain underdeveloped.
• Existing AI governance models lack specific considerations for pediatric populations.

Purpose of the Study:

• To review current AI governance frameworks in healthcare.
• To identify pediatric-specific gaps in AI adoption and governance.
• To propose a pediatric-centric governance approach for AI integration.

Main Methods:

• Review of existing AI governance frameworks.
• Analysis of regulatory data for pediatric Software as a Medical Device (SaMD) approvals.
• Identification of key challenges in pediatric AI implementation.

Main Results:

• Significant gaps exist in pediatric AI governance, including stakeholder engagement, consent/assent processes, bias mitigation, and accountability.
• Radiology dominates FDA-cleared pediatric SaMDs, with other specialties lagging.
• Current frameworks are insufficient for the unique needs of pediatric patients.

Conclusions:

• A pediatric-centric governance model is essential for the safe and responsible integration of AI in children's healthcare.
• Key components include transparency, inclusive participation, equitable data practices, and robust post-deployment monitoring.
• Addressing identified gaps will facilitate broader and more equitable AI adoption in pediatrics.