@inproceedings{kothari-etal-2026-contextualized,
title = "Contextualized Divergence Detection in Health Advice: A Benchmark and Analysis",
author = "Kothari, Reet Yogesh and
Munir, Sirajum and
Basak, Madhusudan and
Preum, Sarah Masud",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1828/",
doi = "10.18653/2026.findings-emnlp.1828",
pages = "39035--39060",
ISBN = "979-8-89176-462-0",
abstract = "Large language models (LLMs) are increasingly used as sources of health advice, raising urgent concerns about reliability, safety, and personalization. While prior work has studied health advice conflicts, these efforts typically ignore patient context and treat contradictions as universal. In this paper, we propose the task of contextualized divergence detection, which frames divergence not only as strict contradictions but also as context-dependent mismatches. To support this task, we introduce a novel dataset of 10,545 samples that combines user profiles, consumer medication guideline documents, and advice statements, with divergences synthetically generated across four categories: direct, conditional, temporal, and sub-typical. We benchmark a set of LLM-based inference approaches, including prompt-based inference, retrieval-augmented generation (RAG), supervised fine-tuning, and agent-based reasoning.\par We observe that relevance-filtered RAG achieves the highest accuracy and robustness for large models, while agent-based reasoning improves interpretability at the cost of accuracy, and supervised fine-tuning is essential for smaller models. A qualitative error analysis highlights recurring challenges such as hedging language, temporal mismatches, and profile-advisory inconsistencies. This work provides the first dataset and systematic evaluation of contextualized divergence detection in health advice, paving the way toward safer, more personalized interactions with LLMs."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kothari-etal-2026-contextualized">
<titleInfo>
<title>Contextualized Divergence Detection in Health Advice: A Benchmark and Analysis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Reet</namePart>
<namePart type="given">Yogesh</namePart>
<namePart type="family">Kothari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sirajum</namePart>
<namePart type="family">Munir</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Madhusudan</namePart>
<namePart type="family">Basak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sarah</namePart>
<namePart type="given">Masud</namePart>
<namePart type="family">Preum</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Large language models (LLMs) are increasingly used as sources of health advice, raising urgent concerns about reliability, safety, and personalization. While prior work has studied health advice conflicts, these efforts typically ignore patient context and treat contradictions as universal. In this paper, we propose the task of contextualized divergence detection, which frames divergence not only as strict contradictions but also as context-dependent mismatches. To support this task, we introduce a novel dataset of 10,545 samples that combines user profiles, consumer medication guideline documents, and advice statements, with divergences synthetically generated across four categories: direct, conditional, temporal, and sub-typical. We benchmark a set of LLM-based inference approaches, including prompt-based inference, retrieval-augmented generation (RAG), supervised fine-tuning, and agent-based reasoning.\par We observe that relevance-filtered RAG achieves the highest accuracy and robustness for large models, while agent-based reasoning improves interpretability at the cost of accuracy, and supervised fine-tuning is essential for smaller models. A qualitative error analysis highlights recurring challenges such as hedging language, temporal mismatches, and profile-advisory inconsistencies. This work provides the first dataset and systematic evaluation of contextualized divergence detection in health advice, paving the way toward safer, more personalized interactions with LLMs.</abstract>
<identifier type="citekey">kothari-etal-2026-contextualized</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1828</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1828/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>39035</start>
<end>39060</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Contextualized Divergence Detection in Health Advice: A Benchmark and Analysis
%A Kothari, Reet Yogesh
%A Munir, Sirajum
%A Basak, Madhusudan
%A Preum, Sarah Masud
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F kothari-etal-2026-contextualized
%X Large language models (LLMs) are increasingly used as sources of health advice, raising urgent concerns about reliability, safety, and personalization. While prior work has studied health advice conflicts, these efforts typically ignore patient context and treat contradictions as universal. In this paper, we propose the task of contextualized divergence detection, which frames divergence not only as strict contradictions but also as context-dependent mismatches. To support this task, we introduce a novel dataset of 10,545 samples that combines user profiles, consumer medication guideline documents, and advice statements, with divergences synthetically generated across four categories: direct, conditional, temporal, and sub-typical. We benchmark a set of LLM-based inference approaches, including prompt-based inference, retrieval-augmented generation (RAG), supervised fine-tuning, and agent-based reasoning.\par We observe that relevance-filtered RAG achieves the highest accuracy and robustness for large models, while agent-based reasoning improves interpretability at the cost of accuracy, and supervised fine-tuning is essential for smaller models. A qualitative error analysis highlights recurring challenges such as hedging language, temporal mismatches, and profile-advisory inconsistencies. This work provides the first dataset and systematic evaluation of contextualized divergence detection in health advice, paving the way toward safer, more personalized interactions with LLMs.
%R 10.18653/2026.findings-emnlp.1828
%U https://aclanthology.org/2026.findings-emnlp.1828/
%U https://doi.org/10.18653/2026.findings-emnlp.1828
%P 39035-39060
Markdown (Informal)
[Contextualized Divergence Detection in Health Advice: A Benchmark and Analysis](https://aclanthology.org/2026.findings-emnlp.1828/) (Kothari et al., Findings 2026)
ACL