Artificial Intelligence in NICU and PICU: A Need for Ecological Validity, Accountability, and Human Factors

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Pediatric Critical Care

Background:

• Neonatal and pediatric intensive care units (NICUs/PICUs) face challenges with rapid diagnosis and treatment for critically ill infants.
• Medication errors and treatment delays pose significant risks to pediatric patients in intensive care settings.
• Clinicians require efficient methods to process vast amounts of medical data for timely decision-making.

Purpose of the Study:

• To explore the application of Artificial Intelligence (AI) in NICU/PICU settings.
• To identify current limitations of AI from a clinical perspective.
• To propose recommendations for enhancing AI readiness for real-world clinical integration.

Main Methods:

• Review of AI applications in neonatal and pediatric intensive care.
• Analysis of AI limitations based on clinician viewpoints.
• Human factors evaluation, including technology readiness level and ecological validity.

Main Results:

• AI integration offers potential benefits for safeguarding pediatric patients and improving care quality.
• Existing AI systems have flaws that hinder seamless clinical workflow integration.
• Addressing AI accountability is crucial for clinician trust and acceptance.

Conclusions:

• AI has the potential to enhance patient safety and clinical decision-making in NICUs/PICUs.
• A human factors approach is essential for validating AI readiness and effectiveness.
• Recommendations are provided to improve AI's suitability for pediatric critical care environments.