Saved in:
Bibliographic Details
Main Authors: Yao, Jianzhou, Liu, Shunchang, Drui, Guillaume, Pettersson, Rikard, Blasimme, Alessandro, Kijewski, Sara
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.00924
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915592330543104
author Yao, Jianzhou
Liu, Shunchang
Drui, Guillaume
Pettersson, Rikard
Blasimme, Alessandro
Kijewski, Sara
author_facet Yao, Jianzhou
Liu, Shunchang
Drui, Guillaume
Pettersson, Rikard
Blasimme, Alessandro
Kijewski, Sara
contents Large language models (LLMs) show promise for supporting clinicians in diagnostic communication by generating explanations and guidance for patients. Yet their ability to produce outputs that are both understandable and empathetic remains uncertain. We evaluate two leading LLMs on medical diagnostic scenarios, assessing understandability using readability metrics as a proxy and empathy through LLM-as-a-Judge ratings compared to human evaluations. The results indicate that LLMs adapt explanations to socio-demographic variables and patient conditions. However, they also generate overly complex content and display biased affective empathy, leading to uneven accessibility and support. These patterns underscore the need for systematic calibration to ensure equitable patient communication. The code and data are released: https://github.com/Jeffateth/Biased_Oracle
format Preprint
id arxiv_https___arxiv_org_abs_2511_00924
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Biased Oracle: Assessing LLMs' Understandability and Empathy in Medical Diagnoses
Yao, Jianzhou
Liu, Shunchang
Drui, Guillaume
Pettersson, Rikard
Blasimme, Alessandro
Kijewski, Sara
Computation and Language
Large language models (LLMs) show promise for supporting clinicians in diagnostic communication by generating explanations and guidance for patients. Yet their ability to produce outputs that are both understandable and empathetic remains uncertain. We evaluate two leading LLMs on medical diagnostic scenarios, assessing understandability using readability metrics as a proxy and empathy through LLM-as-a-Judge ratings compared to human evaluations. The results indicate that LLMs adapt explanations to socio-demographic variables and patient conditions. However, they also generate overly complex content and display biased affective empathy, leading to uneven accessibility and support. These patterns underscore the need for systematic calibration to ensure equitable patient communication. The code and data are released: https://github.com/Jeffateth/Biased_Oracle
title The Biased Oracle: Assessing LLMs' Understandability and Empathy in Medical Diagnoses
topic Computation and Language
url https://arxiv.org/abs/2511.00924