Pivotal challenges in artificial intelligence and machine learning applications for neonatal care

Area of Science:

• Neonatal Medicine
• Artificial Intelligence
• Machine Learning
• Clinical Decision Support Systems

Background:

• AI/ML-based clinical decision support systems (CDSS) hold significant potential for improving neonatal care practices.
• These systems can leverage data from electronic health records, sensors, and imaging to predict outcomes like mortality and disease in neonates.

Purpose of the Study:

• To review the pivotal challenges and factors influencing the implementation of real-time supported CDSS in neonatal care.
• To highlight the gap between the potential of AI/ML in neonatology and its current limited deployment.

Main Methods:

• Literature review focusing on the development, evaluation, and deployment phases of real-time CDSS for neonatal care.
• Analysis of challenges associated with integrating AI/ML into clinical workflows.

Main Results:

• Limited adoption of AI/ML-based CDSS in neonatal healthcare settings, despite their potential.
• Key challenges identified in three main phases: model development, model evaluation, and real-time deployment.

Conclusions:

• Addressing the identified challenges is crucial for the successful implementation of real-time CDSS in neonatal care.
• Overcoming these hurdles will enable the broader application of AI/ML to improve neonate health outcomes.