Assisting patients with reliable information gathering: A guide to building efficient QA systems for medical use
- 1Florida International University, Miami, Florida, United States of America
- 2University of Maryland Baltimore County, Baltimore, Maryland, United States of America
- 3Carnegie Mellon University, Pittsburgh, Pennsylvania, United States of America
Abstract
The application of question-answering (QA) systems in the medical domain has rapidly advanced, significantly improving patients’ access to reliable health-related information. However, current approaches face notable challenges, including the difficulty in obtaining large-scale and unbiased medical datasets, significant privacy concerns, and inefficiencies due to manual dataset annotation. To address these issues, This study introduce a novel methodology leveraging publicly accessible online health forums to systematically build an unbiased, privacy-conscious QA dataset, and it employs Topic-guided Semantic Modeling (TGSM) for automated topic identification, enabling efficient and targeted annotation of relevant patient-generated content. Subsequently, this study propose a two-stage QA pipeline based on a Retriever–Reader architecture, which is further enhanced through fine-tuning state-of-the-art transformer-based models on the constructed domain-specific QA dataset. Experimental results demonstrate that our fine-tuned BioBERT significantly outperforms existing benchmarks, offering accurate patient-derived insights and providing a replicable framework for building efficient medical QA systems. Author summary: In recent years, question-answering (QA) systems leveraging natural language processing (NLP) have significantly improved how patients access reliable health information. However, developing effective QA systems in healthcare remains challenging due to difficulties in obtaining large-scale unbiased datasets, privacy concerns, and the labor-intensive nature of manual dataset annotation. In this paper, we propose a method to systematically build an unbiased and privacy-conscious QA dataset by leveraging publicly accessible online health forums, where individuals openly share their medical experiences. To enhance dataset construction, this work applies a Topic-guided Semantic Modeling (TGSM) technique, efficiently identifying topics and corresponding aspects from unstructured patient-generated content, thus simplifying and accelerating the annotation process. This study then implements a two-stage QA pipeline featuring a Retriever-Reader architecture and further optimizes it by fine-tuning advanced transformer-based models on our dataset. Our experimental results demonstrate that our fine-tuned BioBERT model significantly outperforms existing benchmarks, highlighting the effectiveness and robustness of our approach. By providing accurate patient-derived insights and a replicable framework, our method can contribute meaningfully to the development of effective QA systems for medical applications.
Related
- Engineering Neural Interface Materials for Brain–Computer Interfaces: Toward Reliable Neural Information and Functional Recovery
- Prevention of SARS-CoV-2 transmission during a large, live, indoor gathering (SPRING): a non-inferiority, randomised, controlled trial
- Suboptimal human inference reflects an efficient and flexible information bottleneck
- Clustered heterogeneity, spiking history and efficient silencing maximize mutual information in adaptive time encoding
- INFORME: coupling information-theoretic experimental design with nonlinear mixed-effects modeling for efficient observation scheduling
BibTeX
@article{yin2026assisting,
title = {{Assisting patients with reliable information gathering: A guide to building efficient QA systems for medical use}},
author = {Zhipeng Yin and Zichong Wang and Min Chen and Ian Stockwell and Xin Ning and Jun Liu and Wenbin Zhang},
journal = {PLOS Digital Health},
year = {2026},
month = sep,
doi = {10.1371/journal.pdig.0001634},
url = {https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001634},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Assisting patients with reliable information gathering: A guide to building efficient QA systems for medical use
AU - Zhipeng Yin
AU - Zichong Wang
AU - Min Chen
AU - Ian Stockwell
AU - Xin Ning
AU - Jun Liu
AU - Wenbin Zhang
PY - 2026
DA - 2026/09/17
JO - PLOS Digital Health
DO - 10.1371/journal.pdig.0001634
UR - https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001634
AB - The application of question-answering (QA) systems in the medical domain has rapidly advanced, significantly improving patients’ access to reliable health-related information. However, current approaches face notable challenges, including the difficulty in obtaining large-scale and unbiased medical datasets, significant privacy concerns, and inefficiencies due to manual dataset annotation. To address these issues, This study introduce a novel methodology leveraging publicly accessible online health forums to systematically build an unbiased, privacy-conscious QA dataset, and it employs Topic-guided Semantic Modeling (TGSM) for automated topic identification, enabling efficient and targeted annotation of relevant patient-generated content. Subsequently, this study propose a two-stage QA pipeline based on a Retriever–Reader architecture, which is further enhanced through fine-tuning state-of-the-art transformer-based models on the constructed domain-specific QA dataset. Experimental results demonstrate that our fine-tuned BioBERT significantly outperforms existing benchmarks, offering accurate patient-derived insights and providing a replicable framework for building efficient medical QA systems. Author summary: In recent years, question-answering (QA) systems leveraging natural language processing (NLP) have significantly improved how patients access reliable health information. However, developing effective QA systems in healthcare remains challenging due to difficulties in obtaining large-scale unbiased datasets, privacy concerns, and the labor-intensive nature of manual dataset annotation. In this paper, we propose a method to systematically build an unbiased and privacy-conscious QA dataset by leveraging publicly accessible online health forums, where individuals openly share their medical experiences. To enhance dataset construction, this work applies a Topic-guided Semantic Modeling (TGSM) technique, efficiently identifying topics and corresponding aspects from unstructured patient-generated content, thus simplifying and accelerating the annotation process. This study then implements a two-stage QA pipeline featuring a Retriever-Reader architecture and further optimizes it by fine-tuning advanced transformer-based models on our dataset. Our experimental results demonstrate that our fine-tuned BioBERT model significantly outperforms existing benchmarks, highlighting the effectiveness and robustness of our approach. By providing accurate patient-derived insights and a replicable framework, our method can contribute meaningfully to the development of effective QA systems for medical applications.
SN - 2767-3170
ER -