A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Mahyar Abbasian, Saba A. Farahani, Arshia Ilaty, Hung Cao, Ramesh Jain, Amir M. Rahmani
• Venue: arXiv: cs.CL
• Published: 2026-08-20
• Collected: 2026-08-21
• URL: http://arxiv.org/abs/2608.19875v1
• PDF: https://arxiv.org/pdf/2608.19875v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

The paper introduces a framework to help healthcare chatbots better understand underspecified patient queries by using a knowledge graph to generate plausible hypotheses and asking targeted follow-up questions. The framework was tested on two benchmarks and showed significant improvements in accuracy for diagnosis retrieval and dietary safety classification compared to direct prompting or rephrasing.

What the Document Says

• The paper discusses a knowledge-guided agentic framework aimed at mitigating ambiguity in patient queries submitted to healthcare chatbots.
• The framework interprets the initial query, uses a task-specific knowledge graph to generate plausible hypotheses, identifies missing patient-context variables, asks targeted follow-up questions, and combines the original query with acquired context for the downstream language model.
• The paper evaluates the framework across five language models on two controlled ambiguity-mitigation benchmarks: diagnosis retrieval and dietary-safety classification.
• In diagnosis retrieval, the framework increased exact Top-1 accuracy by at least 57.1 percentage points and selective exact Recall@5 by at least 77.7 percentage points across five models compared to direct prompting.
• In dietary-safety classification, the framework improved accuracy across all five models and achieved the highest Matthews correlation coefficient for four of them.

Key Concepts Mentioned

• healthcare chatbots
• patient queries
• knowledge graph
• plausible hypotheses
• targeted follow-up questions
• diagnosis retrieval
• dietary safety classification

Claims Made by the Document

• The framework can mitigate ambiguity in patient queries.
• The framework increases exact Top-1 accuracy and selective exact Recall@5 in diagnosis retrieval.
• The framework improves accuracy in dietary-safety classification.

Evidence Used

• Results from evaluations on two controlled benchmarks.

Methods / Approach

• Interpreting initial query, using a task-specific knowledge graph, generating plausible hypotheses, identifying missing patient-context variables, asking targeted follow-up questions, combining original and acquired context for the downstream model

Limitations Stated by the Authors

Not stated in source

Limitations Noted by the Agent

Agent interpretation — not from the source.

The framework is only tested on five language models.

Open Questions

• How does the framework perform on real-world, less controlled datasets?
• What is the impact of the framework on user satisfaction and perceived helpfulness?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The paper focuses on improving the accuracy of healthcare chatbots in interpreting patient queries by mitigating ambiguity using a knowledge-guided agentic framework.

Raw Abstract / Excerpt

Patients often submit short, underspecified queries to healthcare chatbots that lack the patient-specific information needed to determine an appropriate response. Although these queries may be linguistically clear, they can support multiple plausible answers depending on undisclosed factors such as symptoms, diagnoses, medications, allergies, or dietary restrictions. A language model answering such a query directly may therefore rely on unsupported assumptions about the patient. We introduce a knowledge-guided agentic framework for mitigating patient-context ambiguity before final response generation. The framework operates between the patient and an otherwise unchanged downstream language model. It interprets the initial query, uses a task-specific knowledge graph to construct a set of plausible hypotheses, identifies the missing patient-context variables needed to distinguish among them, and asks targeted follow-up questions. The original query and the acquired context are then combined into a clarified prompt for the downstream model. We evaluated the framework across five language models using two controlled ambiguity-mitigation benchmarks: diagnosis retrieval from 1,034 symptom queries with clinically relevant evidence systematically masked, and dietary-safety classification from 487 queries with decisive health context omitted. The framework was compared with direct answering of the underspecified query and with rephrasing the same query without acquiring new patient information. In diagnosis retrieval, it increased overall exact Top-1 accuracy by at least 57.1 percentage points and selective exact Recall@5 by at least 77.7 percentage points across the five evaluated models compared with direct prompting. In dietary-safety classification, it improved accuracy across all five models and achieved the highest Matthews correlation coefficient for four...