{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:47:02Z","timestamp":1792086422298,"version":"4.4.0"},"reference-count":29,"publisher":"Public Library of Science (PLoS)","issue":"10","license":[{"start":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T00:00:00Z","timestamp":1792022400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100012456","name":"National Social Science Fund of China","doi-asserted-by":"publisher","award":["25CKX003"],"award-info":[{"award-number":["25CKX003"]}],"id":[{"id":"10.13039\/501100012456","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosdigitalhealth.org"],"crossmark-restriction":false},"short-container-title":["PLOS Digit Health"],"abstract":"<jats:p>Large language models (LLMs) can produce patient-friendly adaptations, such as simplifying specialized medical literature for readers with a middle school education. However, any distortion of knowledge during this process may have serious consequences. Current evaluations overlook such risks. In this study, we proposes a knowledge triplet comparison (KTC) method to assess knowledge distortion by systematically quantifying the retention and alteration of medical knowledge before and after text adaptation. Using the Chemical-Disease Relations dataset, this study evaluates the performance of six LLMs across eight adaptation scenarios (e.g., adjusting reading level, modifying text length) and draws the following key conclusions: First, from a quantitative perspective, knowledge distortion is prevalent in all adaptation scenarios, with the models retaining only approximately half of the original medical knowledge. Additionally, about 10% more knowledge is added. Second, the distribution of knowledge distortion follows a clear categorical distribution pattern, with etiology and pathological mechanisms, as well as complications and risk factors, being the most frequently distortion types. Third, targeted corrections through prompt engineering have no significant impact on the quantity and categorical distribution patterns of knowledge distortion, indicating that knowledge distortion is persistent and regular. This study shows that, in addition to the more obvious risk of hallucinations caused by a small amount of added knowledge, patient-friendly rewrites by LLMs also pose a significant risk of subtle yet potentially serious knowledge omissions. Furthermore, this research reveals the regularity of knowledge distortion, laying the groundwork for further investigation into the patterns of knowledge distortion and the development of solutions to this problem.<\/jats:p>","DOI":"10.1371\/journal.pdig.0001739","type":"journal-article","created":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:21:51Z","timestamp":1792084911000},"page":"e0001739","update-policy":"https:\/\/doi.org\/10.1371\/journal.pdig.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Prevalent knowledge omissions: An unseen risk of LLMs in patient-friendly adaptation of medical literature"],"prefix":"10.1371","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4613-2626","authenticated-orcid":true,"given":"Mei","family":"Chen","sequence":"first","affiliation":[{"name":"School of Information Engineering, Minzu University of China","place":["Beijing, China"]},{"name":"Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance of MOE, Minzu University of China","place":["Beijing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shibin","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Information Engineering, Minzu University of China","place":["Beijing, China"]},{"name":"Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance of MOE, Minzu University of China","place":["Beijing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,10,15]]},"reference":[{"issue":"1","key":"pdig.0001739.ref001","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1038\/s43856-023-00370-1","article-title":"The future landscape of large language models in medicine","volume":"3","author":"J Clusmann","year":"2023","journal-title":"Commun Med (Lond)"},{"issue":"1","key":"pdig.0001739.ref002","doi-asserted-by":"crossref","first-page":"183","DOI":"10.1038\/s41746-024-01157-x","article-title":"The ethics of ChatGPT in medicine and healthcare: a systematic review on Large Language Models (LLMs)","volume":"7","author":"J Haltaufderheide","year":"2024","journal-title":"NPJ Digit Med"},{"key":"pdig.0001739.ref003","doi-asserted-by":"crossref","unstructured":"Huo H, Fu Z, Yu H, et al. Chinese Text Simplification Based on Large Language Models. 2024 International Conference on Computational Linguistics and Natural Language Processing (CLNLP), 2024: 52\u20136. https:\/\/ieeexplore.ieee.org\/abstract\/document\/10743168","DOI":"10.1109\/CLNLP64123.2024.00018"},{"issue":"1","key":"pdig.0001739.ref004","doi-asserted-by":"crossref","first-page":"288","DOI":"10.1038\/s41746-024-01282-7","article-title":"Medical large language models are susceptible to targeted misinformation attacks","volume":"7","author":"T Han","year":"2024","journal-title":"NPJ Digit Med"},{"issue":"10","key":"pdig.0001739.ref005","doi-asserted-by":"crossref","first-page":"1156","DOI":"10.1038\/s42256-024-00899-3","article-title":"Poisoning medical knowledge using large language models","volume":"6","author":"J Yang","year":"2024","journal-title":"Nat Mach Intell"},{"key":"pdig.0001739.ref006","unstructured":"Chen Q, Hu Y, Peng X, et al. A systematic evaluation of large language models for biomedical natural language processing: benchmarks, baselines, and recommendations (2024\u201309\u201330). https:\/\/arxiv.org\/abs\/2305.16326"},{"issue":"9","key":"pdig.0001739.ref007","article-title":"Towards trustworthy LLMs: a review on debiasing and dehallucinating in large language models","volume":"57","author":"Z Lin","year":"2024","journal-title":"Artif Intell Rev"},{"key":"pdig.0001739.ref008","article-title":"User-Controlled Knowledge Fusion in Large Language Models: Balancing Creativity and Hallucination","author":"C Zhang","year":"2023","journal-title":"arXiv."},{"key":"pdig.0001739.ref009","doi-asserted-by":"crossref","unstructured":"Li L-H, Nugroho AC. Evaluating Large Language Models: Challenges, Limitations, and Future Directions. 2025 Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT &amp; NCON), 2025: 29\u201334. https:\/\/ieeexplore.ieee.org\/abstract\/document\/10961969","DOI":"10.1109\/ECTIDAMTNCON64748.2025.10961969"},{"issue":"5","key":"pdig.0001739.ref010","doi-asserted-by":"crossref","first-page":"104216","DOI":"10.1016\/j.ipm.2025.104216","article-title":"A survey on biomedical automatic text summarization with large language models","volume":"62","author":"Z Huang","year":"2025","journal-title":"Information Processing & Management"},{"key":"pdig.0001739.ref011","article-title":"Fact-Controlled Diagnosis of Hallucinations in Medical Text Summarization","volume":"2025","author":"H-C Shing","year":"2025","journal-title":"arXiv"},{"key":"pdig.0001739.ref012","doi-asserted-by":"crossref","DOI":"10.52034\/lanstts.v8i.246","article-title":"Automated error analysis for multiword expressions: Using BLEU-type scores for automatic discovery of potential translation errors","volume":"8","author":"B Babych","year":"2021","journal-title":"LANS-TTS"},{"key":"pdig.0001739.ref013","doi-asserted-by":"crossref","first-page":"102571","DOI":"10.1016\/j.jocs.2025.102571","article-title":"Assessing the effectiveness of ROUGE as unbiased metric in Extractive vs. Abstractive summarization Techniques","volume":"87","author":"CA Auriemma","year":"2025","journal-title":"Journal of Computational Science"},{"key":"pdig.0001739.ref014","article-title":"BERTScore: Evaluating Text Generation with BERT","author":"T Zhang","year":"2019","journal-title":"arXiv"},{"issue":"2","key":"pdig.0001739.ref015","doi-asserted-by":"crossref","first-page":"78","DOI":"10.3390\/info11020078","article-title":"A Framework for Word Embedding Based Automatic Text Summarization and Evaluation","volume":"11","author":"TT Hailu","year":"2020","journal-title":"Information"},{"issue":"7","key":"pdig.0001739.ref016","doi-asserted-by":"crossref","first-page":"1611","DOI":"10.1007\/s13042-020-01069-8","article-title":"From static to dynamic word representations: a survey","volume":"11","author":"Y Wang","year":"2020","journal-title":"Int J Mach Learn & Cyber"},{"issue":"4","key":"pdig.0001739.ref017","doi-asserted-by":"crossref","first-page":"861","DOI":"10.1162\/coli_a_00418","article-title":"The (Un)Suitability of Automatic Evaluation Metrics for Text Simplification","volume":"47","author":"F Alva-Manchego","year":"2021","journal-title":"Computational Linguistics"},{"key":"pdig.0001739.ref018","doi-asserted-by":"crossref","unstructured":"Jourdan L, Hernandez N, Boudin F, Dufour R. Identifying Reliable Evaluation Metrics for Scientific Text Revision. In: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2025. 6731\u201356. https:\/\/doi.org\/10.18653\/v1\/2025.acl-long.335","DOI":"10.18653\/v1\/2025.acl-long.335"},{"key":"pdig.0001739.ref019","doi-asserted-by":"crossref","first-page":"658","DOI":"10.1109\/OJCS.2025.3560333","article-title":"A Detailed Comparative Analysis of Automatic Neural Metrics for Machine Translation: BLEURT & BERTScore","volume":"6","author":"A Mukherjee","year":"2025","journal-title":"IEEE Open Journal of the Computer Society"},{"key":"pdig.0001739.ref020","article-title":"Towards Explainable Evaluation Metrics for Natural Language Generation","volume":"2022","author":"C Leiter","year":"2022","journal-title":"arXiv"},{"issue":"00","key":"pdig.0001739.ref021","doi-asserted-by":"crossref","first-page":"432","DOI":"10.1162\/tacl_a_00653","article-title":"Do Text Simplification Systems Preserve Meaning? A Human Evaluation via Reading Comprehension","volume":"12","author":"S Agrawal","year":"2024","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"pdig.0001739.ref022","author":"T Guidroz","year":"2025"},{"key":"pdig.0001739.ref023","first-page":"76","volume-title":"Comparative analysis of evaluation measures for scientific text simplification. InInternational Conference on Theory and Practice of Digital Libraries","author":"D Davari","year":"2024"},{"key":"pdig.0001739.ref024","doi-asserted-by":"crossref","first-page":"104368","DOI":"10.1016\/j.jbi.2023.104368","article-title":"Causal feature selection using a knowledge graph combining structured knowledge from the biomedical literature and ontologies: A use case studying depression as a risk factor for Alzheimer\u2019s disease","volume":"142","author":"SA Malec","year":"2023","journal-title":"J Biomed Inform"},{"key":"pdig.0001739.ref025","article-title":"BioCreative V CDR task corpus: A resource for chemical disease relation Extraction","volume":"2016","author":"J Li","year":"2016","journal-title":"Database"},{"issue":"4","key":"pdig.0001739.ref026","article-title":"Prompting large language models to extract chemical\u2012disease relation precisely and comprehensively at the document level: an evaluation study","volume":"20","author":"M Chen","year":"2025","journal-title":"PLoS One"},{"key":"pdig.0001739.ref027","author":"Y Tao","year":"2014"},{"issue":"20","key":"pdig.0001739.ref028","doi-asserted-by":"crossref","first-page":"7895","DOI":"10.1021\/acs.jcim.4c00882","article-title":"How Does a Generative Large Language Model Perform on Domain-Specific Information Extraction?\u2500A Comparison between GPT-4 and a Rule-Based Method on Band Gap Extraction","volume":"64","author":"X Wang","year":"2024","journal-title":"J Chem Inf Model"},{"issue":"1","key":"pdig.0001739.ref029","doi-asserted-by":"crossref","first-page":"41","DOI":"10.1038\/s41746-024-01029-4","article-title":"Prompt engineering in consistency and reliability with the evidence-based guideline for LLMs","volume":"7","author":"L Wang","year":"2024","journal-title":"NPJ Digit Med"}],"container-title":["PLOS Digital Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001739","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:21:58Z","timestamp":1792084918000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001739"}},"subtitle":[],"editor":[{"given":"Po-Chih","family":"Kuo","sequence":"first","affiliation":[{"name":"National Tsing Hua University","place":["TAIWAN"]}],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,10,15]]},"references-count":29,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,10,15]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pdig.0001739","relation":{},"ISSN":["2767-3170"],"issn-type":[{"value":"2767-3170","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,15]]}}}