This project conducts exploratory work on a HIPAA-compliant, large language model (LLM)-based tool that integrates structured and unstructured oncology electronic health record (EHR) data to automate the development of tailored SCPs, paired with a patient-facing chatbot to answer questions about the SCP. Unlike generic AI documentation tools, this system establishes benchmarks for SCP development and embeds user-centered design (UCD) directly into prompt engineering and model governance workflows.
Study Type
INTERVENTIONAL
Allocation
NA
Purpose
HEALTH_SERVICES_RESEARCH
Masking
NONE
Enrollment
50
A secure LLM pipeline with structured and unstructured EHR data to automatically generate Survivorship Care Plans (SCPs) and develop a patient-facing chatbot to support comprehension. The SCP chatbot will leverage GARDE-Chat, an NCI-funded open-source chatbot platform
University of Utah
Salt Lake City, Utah, United States
Feasibility: Recruitment Retention
Feasibility of the interactive AI-SCP and chatbot. This outcome measure will report the proportion of subjects who were retained from recruitment.
Time frame: up to 7 days
Acceptability/Usability: System Usability Scale
Acceptability and usability of the interactive AI-SCP and chatbot. This outcome measure will report the mean System Usability Scale (SUS) score. SUS scores range from 0-100, with higher scores indicating better usability and lower scores indicating worse usability.
Time frame: up to 7 days
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.