.png){kind=icon}
.png){kind=icon}
Return
EPPCMinerBen: A novel benchmark for evaluating large language models on electronic patient-provider communication via the patient portal
.png){kind=icon}
PRE
AI
S
Y
L
S
J
S
DOI：10.1016/j.artmed.2026.103429.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• This work introduces EPPCMinerBen, a first benchmark specifically constructed to analyze EPPC within large-scale secure messaging data, addressing a critical gap left by existing clinical dialogue datasets. • This work is based on a richly annotated, de-identified EPPC dataset that captures diverse relational communication patterns, including information seeking/sharing, socio-emotional expressions, partnership-building, and shared decision-making. The dataset provides a strong foundation for analyzing the interactive dynamics of patient-provider communication and supports the development of models that reflect real-world conversational complexity. • This work further proposes a set of tailored instruction prompts designed to guide large language models in interpreting and reasoning over EPPC data, enabling more accurate identification of communication functions and relational cues critical for enhancing patient-centered care and treatment adherence.
Keywords:
EPPC
patient-provider communication
large language models
benchmark
patient-centered care
Journal
IF:
6.2
Papers:
2.5K
Citations:
7.8K
Organization
Cited Papers