Integrating Large Language Models into Obstetric EHRs: The MedGPT Use Case in Portugal

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Obstetrics and Gynecology

Background:

• Obstetric electronic health record (EHR) systems like ObsCare collect vital maternal health data.
• Portugal faces challenges in obstetric care, including workforce shortages, rising C-section rates, and increasing maternal age.

Purpose of the Study:

• To integrate fine-tuned large language models (LLMs) into the ObsCare EHR system.
• To develop MedGPT for summarizing patient health journeys, providing decision support, and generating educational content.
• To address critical gaps in obstetric care through AI-driven solutions.

Main Methods:

• Integration of LLMs into the existing ObsCare EHR system.
• Development of features for health journey summarization, decision support, and personalized patient education.
• Focus on ethically aligned Generative AI solutions within a European collaboration.

Main Results:

• The project aims to reduce clinicians' documentation burden.
• Enhancement of patient engagement through personalized content.
• Preliminary results from an ongoing European collaboration.

Conclusions:

• MedGPT has the potential to significantly improve obstetric care efficiency and patient experience.
• The project demonstrates a proactive approach to leveraging AI in addressing healthcare challenges.
• Ethical considerations are central to the development of these Generative AI solutions in healthcare.