Mitigating Hallucinations in Healthcare LLMs with Granular Fact-Checking and Domain-Specific Adaptation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918519680008192 |
|---|---|
| author | Zeba, Musarrat Mamun, Abdullah Al Tithee, Kishoar Jahan Sutradhar, Debopom Raiaan, Mohaimenul Azam Khan Mukta, Saddam Mohamed, Reem E. Islam, Md Rafiqul Sebastian, Yakub Hussain, Mukhtar Azam, Sami |
| author_facet | Zeba, Musarrat Mamun, Abdullah Al Tithee, Kishoar Jahan Sutradhar, Debopom Raiaan, Mohaimenul Azam Khan Mukta, Saddam Mohamed, Reem E. Islam, Md Rafiqul Sebastian, Yakub Hussain, Mukhtar Azam, Sami |
| contents | In healthcare, it is essential for any LLM-generated output to be reliable and accurate, particularly in cases involving decision-making and patient safety. However, the outputs are often unreliable in such critical areas due to the risk of hallucinated outputs from the LLMs. To address this issue, we propose a fact-checking module that operates independently of any LLM, along with a domain-specific summarization model designed to minimize hallucination rates. Our model is fine-tuned using Low-Rank Adaptation (LoRa) on the MIMIC III dataset and is paired with the fact-checking module, which uses numerical tests for correctness and logical checks at a granular level through discrete logic in natural language processing (NLP) to validate facts against electronic health records (EHRs). We trained the LLM model on the full MIMIC-III dataset. For evaluation of the fact-checking module, we sampled 104 summaries, extracted them into 3,786 propositions, and used these as facts. The fact-checking module achieves a precision of 0.8904, a recall of 0.8234, and an F1-score of 0.8556. Additionally, the LLM summary model achieves a ROUGE-1 score of 0.5797 and a BERTScore of 0.9120 for summary quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_16189 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Mitigating Hallucinations in Healthcare LLMs with Granular Fact-Checking and Domain-Specific Adaptation Zeba, Musarrat Mamun, Abdullah Al Tithee, Kishoar Jahan Sutradhar, Debopom Raiaan, Mohaimenul Azam Khan Mukta, Saddam Mohamed, Reem E. Islam, Md Rafiqul Sebastian, Yakub Hussain, Mukhtar Azam, Sami Computation and Language In healthcare, it is essential for any LLM-generated output to be reliable and accurate, particularly in cases involving decision-making and patient safety. However, the outputs are often unreliable in such critical areas due to the risk of hallucinated outputs from the LLMs. To address this issue, we propose a fact-checking module that operates independently of any LLM, along with a domain-specific summarization model designed to minimize hallucination rates. Our model is fine-tuned using Low-Rank Adaptation (LoRa) on the MIMIC III dataset and is paired with the fact-checking module, which uses numerical tests for correctness and logical checks at a granular level through discrete logic in natural language processing (NLP) to validate facts against electronic health records (EHRs). We trained the LLM model on the full MIMIC-III dataset. For evaluation of the fact-checking module, we sampled 104 summaries, extracted them into 3,786 propositions, and used these as facts. The fact-checking module achieves a precision of 0.8904, a recall of 0.8234, and an F1-score of 0.8556. Additionally, the LLM summary model achieves a ROUGE-1 score of 0.5797 and a BERTScore of 0.9120 for summary quality. |
| title | Mitigating Hallucinations in Healthcare LLMs with Granular Fact-Checking and Domain-Specific Adaptation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2512.16189 |