Using Artificial Intelligence and Machine Learning to Promote Child Health Equity.

Area of Science:

• Pediatric Health
• Health Equity
• Computational Health

Background:

• Artificial intelligence (AI) and machine learning (ML) can exacerbate health inequalities if not used carefully.
• ML offers potential insights into socioeconomic factors affecting health, enabling individual and population-level predictions.
• Addressing health equity in child health requires careful consideration of AI/ML applications.

Purpose of the Study:

• To outline potential applications of ML in child health, focusing on its impact on health equity.
• To describe two novel ML use cases for promoting population health equity in diverse urban settings.
• To identify and mitigate potential inequities embedded in ML training data, models, and deployment.

Main Methods:

• Development and training of an ML algorithm using routine demographic data to predict outpatient appointment nonattendance.
• Creation of a risk-prediction tool for pediatric asthma using routine health determinant metrics.
• Application of ML in a diverse inner-city London population.

Main Results:

• Demonstrated ML's utility in predicting nonattendance at child health appointments.
• Developed a tool to identify high-risk asthma patients for preventive interventions.
• Highlighted potential for ML to target interventions and improve health equity.

Conclusions:

• ML holds significant promise for advancing child health equity when developed and deployed thoughtfully.
• Proactive strategies are necessary to prevent the inadvertent embedding of inequity in ML systems.
• Mitigation strategies are crucial for ensuring AI/ML benefits all children equitably.