Susceptibility of Large Language Models to User-Driven Factors in Medical Queries
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908289691811840 |
|---|---|
| author | Lim, Kyung Ho Kang, Ujin Li, Xiang Kim, Jin Sung Jung, Young-Chul Park, Sangjoon Kim, Byung-Hoon |
| author_facet | Lim, Kyung Ho Kang, Ujin Li, Xiang Kim, Jin Sung Jung, Young-Chul Park, Sangjoon Kim, Byung-Hoon |
| contents | Large language models (LLMs) are increasingly used in healthcare, but their reliability is heavily influenced by user-driven factors such as question phrasing and the completeness of clinical information. In this study, we examined how misinformation framing, source authority, model persona, and omission of key clinical details affect the diagnostic accuracy and reliability of LLM outputs. We conducted two experiments: one introducing misleading external opinions with varying assertiveness (perturbation test), and another removing specific categories of patient information (ablation test). Using public datasets (MedQA and Medbullets), we evaluated proprietary models (GPT-4o, Claude 3.5 Sonnet, Claude 3.5 Haiku, Gemini 1.5 Pro, Gemini 1.5 Flash) and open-source models (LLaMA 3 8B, LLaMA 3 Med42 8B, DeepSeek R1 8B). All models were vulnerable to user-driven misinformation, with proprietary models especially affected by definitive and authoritative language. Assertive tone had the greatest negative impact on accuracy. In the ablation test, omitting physical exam findings and lab results caused the most significant performance drop. Although proprietary models had higher baseline accuracy, their performance declined sharply under misinformation. These results highlight the need for well-structured prompts and complete clinical context. Users should avoid authoritative framing of misinformation and provide full clinical details, especially for complex cases. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_22746 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Susceptibility of Large Language Models to User-Driven Factors in Medical Queries Lim, Kyung Ho Kang, Ujin Li, Xiang Kim, Jin Sung Jung, Young-Chul Park, Sangjoon Kim, Byung-Hoon Computation and Language Artificial Intelligence Computers and Society Large language models (LLMs) are increasingly used in healthcare, but their reliability is heavily influenced by user-driven factors such as question phrasing and the completeness of clinical information. In this study, we examined how misinformation framing, source authority, model persona, and omission of key clinical details affect the diagnostic accuracy and reliability of LLM outputs. We conducted two experiments: one introducing misleading external opinions with varying assertiveness (perturbation test), and another removing specific categories of patient information (ablation test). Using public datasets (MedQA and Medbullets), we evaluated proprietary models (GPT-4o, Claude 3.5 Sonnet, Claude 3.5 Haiku, Gemini 1.5 Pro, Gemini 1.5 Flash) and open-source models (LLaMA 3 8B, LLaMA 3 Med42 8B, DeepSeek R1 8B). All models were vulnerable to user-driven misinformation, with proprietary models especially affected by definitive and authoritative language. Assertive tone had the greatest negative impact on accuracy. In the ablation test, omitting physical exam findings and lab results caused the most significant performance drop. Although proprietary models had higher baseline accuracy, their performance declined sharply under misinformation. These results highlight the need for well-structured prompts and complete clinical context. Users should avoid authoritative framing of misinformation and provide full clinical details, especially for complex cases. |
| title | Susceptibility of Large Language Models to User-Driven Factors in Medical Queries |
| topic | Computation and Language Artificial Intelligence Computers and Society |
| url | https://arxiv.org/abs/2503.22746 |