The Biased Oracle: Assessing LLMs' Understandability and Empathy in Medical Diagnoses
This addresses the problem of ensuring equitable patient communication for clinicians using LLMs, but it is incremental as it builds on existing evaluations of LLMs in medical contexts.
The study evaluated two leading LLMs on medical diagnostic scenarios, finding that while they adapt explanations to socio-demographic variables and patient conditions, they generate overly complex content and display biased affective empathy, leading to uneven accessibility and support.
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