Transporting an Artificial Intelligence Model to Predict Emergency Cesarean Delivery: Overcoming Challenges Posed by...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Clinical Decision Support Systems

Background:

• Artificial intelligence (AI) is poised to transform medical practice and healthcare delivery.
• Successful AI implementation requires seamless transfer of models across diverse healthcare settings.
• Interfacility variations in data reporting can impede the widespread adoption of AI tools.

Purpose of the Study:

• To evaluate the cross-facility application of an AI model designed to predict the necessity of emergency caesarean sections.
• To identify challenges encountered during the interfacility deployment of a predictive AI model in obstetrics.

Main Methods:

• Development and validation of an AI model for predicting emergency caesarean delivery.
• Deployment and testing of the AI model across multiple healthcare facilities.
• Analysis of model performance and identification of factors influencing interfacility variation.

Main Results:

• The transported AI model demonstrated utility in predicting emergency caesarean needs.
• Significant benefits were observed from the cross-facility application of the AI model.
• Interfacility differences in clinical data reporting practices were identified as a key challenge.

Conclusions:

• AI models for predicting emergency caesarean sections can be successfully applied across different healthcare facilities.
• Addressing interfacility variations in reporting practices is crucial for optimizing AI deployment in healthcare.
• Further research is needed to standardize data reporting and enhance the robustness of AI tools in clinical settings.