Large Language Models in Action: Smarter CDSS for UTI Diagnosis
THMT is excited to share a groundbreaking peer reviewed article that brings the promise of large language models (LLMs) into real world clinical decision support systems
(CDSS):
Integrating Large Language Models into Clinical Decision Support Systems: A Novel Approach to UTI Diagnosis and Treatment
Jain M et al., Emory, UTHSC, MIT CSAIL, and Cornell University
https://doi.org/10.30953/thmt.v10.554
This work introduces 3RDI, an adaptive AI powered CDSS built on the DETNQ framework (Diagnosis, Evidence, Treatment Plan, Notes, Quality) - currently in pilot integration with Epic EHR.
This study offers a unique, forward looking reference for anyone working at the intersection of:
✔ Large language models in clinical workflows
✔ Smart, dynamic CDSS design
✔ UTI and infectious disease management
✔
EHR-integrated AI systems
✔ Continuous learning systems in clinical settings
Key takeaways include:
- Real-time, feedback-driven AI-CDSS model integrated into Epic
- DETNQ framework for structured, explainable AI outputs
- Clinician engagement in iterative design for trust & usability
- Promising early results for diagnostic accuracy, reduced errors, and improved efficiency
Read & cite now: https://doi.org/10.30953/thmt.v10.554
This paper lays the groundwork for scalable AI-CDSS frameworks applicable across conditions — and belongs on the reference list of any serious discussion about LLMs in healthcare.
