{ "abstract": "Objective\n\nClinical narratives in electronic health records (EHRs) hold diagnostic, therapeutic and temporal information often missing from structured fields; manual chart review remains the de facto standard but is slow, costly and variable. We aimed to develop and evaluate Clinical LLM‑based Information Extraction and Structuring Agent (CLINES), a pipeline that extracts and normalises clinical concepts from free-text notes into ontology-grounded, auditable, schema-ready data without task-specific fine-tuning.Methods and analysis\n\nWe developed CLINES, a modular multistep pipeline: semantic chunking of long notes; extraction by reasoning-capable large language models (LLM); attribute assignment (assertion, value+unit); Unified Medical Language System normalisation; explicit and relative date resolution; and aggregation into an i2b2-style schema. Evaluation was conducted without task-specific supervised training on deidentified MIMIC-III, 4CE and CORAL (breast, pancreas) notes. Comparators include rule/lexicon systems, transformer encoders, single-prompt LLM baselines, alongside chain-of-thought LLMs and full-size clinical named entity recognition encoders (BERT-base, GatorTron-base). Outcomes were F1 for entity, assertion, value+unit and date processing, reported with 95% bootstrap CIs and paired-bootstrap, false discovery rate (FDR)-corrected tests. Inter-annotator agreement, hallucination taxonomy, cost and component ablations were also assessed.Results\n\nAcross all datasets and tasks, CLINES ranked first on every task-dataset combination. F1 ranges were 0.69-0.87 (entity), 0.84-0.93 (assertion), 0.77-0.90 (value+unit) and 0.73-0.79 (date, where gold annotations were available). Gains over the strongest single-prompt LLM were +0.21-0.38; transformer encoders trailed by +0.28-0.68 F1 on entity extraction. Performance remained stable as notes lengthened, while transformer baselines lost recall. Code field gains and most assertion/value comparisons were significant at FDR-adjusted p<0.05 on paired-bootstrap tests; mention-level inter-annotator F1 on a cross-annotated subset was 0.82 (assertion κ=0.85); a seven-class taxonomy on 700 stratified false positives found 54% of apparent false positives to be annotation gaps rather than true hallucinations.Conclusion\n\nCLINES translates narrative EHR text into ontology-grounded, auditable, schema-ready data, offering a practical route to scale chart review-like extraction for cohort discovery and real-world evidence. CLINES substantially reduces but does not eliminate the need for manual review: the residual hallucination quantified here means human verification remains necessary for safety-critical use. CLINES is model agnostic-open and closed models can be substituted to meet cost, performance and privacy goals.", "abstract_shown": true, "affiliations": [ "Department of Biomedical Informatics, Harvard Medical School, Boston, Massachusetts, USA.", "Department of Statistics and Data Science, Tsinghua University, Beijing, China.", "Division of Biostatistics, University of California Berkeley, Berkeley, California, USA.", "Harvard Medical School, Boston, Massachusetts, USA.", "Division of Interdisciplinary Medicine and Biotechnology, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.", "Harvard T H Chan School of Public Health, Boston, Massachusetts, USA." ], "article_type": "research article", "arxiv": null, "author_total": 13, "authors": [ { "affs": [ 0 ], "name": "Zongxin Yang", "orcid": "0000-0001-8783-8313" }, { "affs": [ 1 ], "name": "Hongyi Yuan" }, { "affs": [ 0 ], "name": "Raheel Sayeed" }, { "affs": [ 0 ], "name": "Amelia Tan" }, { "affs": [ 0 ], "name": "Enci Cai" }, { "affs": [ 0 ], "name": "Mohammed Moro" }, { "affs": [ 2 ], "name": "Xiudi Li" }, { "affs": [ 1 ], "name": "Huaiyuan Ying" }, { "affs": [ 3 ], "name": "Nicholas Brown" }, { "affs": [ 3, 4 ], "name": "Griffin Weber" }, { "affs": [ 1 ], "name": "Sheng Yu" }, { "affs": [ 0 ], "name": "Isaac Kohane" }, { "affs": [ 0, 5 ], "name": "Tianxi Cai" } ], "bibtex": "https://research.draconitic.com/paper/9840kfa5e0dtz.bib", "cited_by": 0, "doi": "10.1136/bmjdh-2026-000027", "fulltext_url": "https://europepmc.org/articles/PMC13711561", "id": "9840kfa5e0dtz", "issns": [ "3049-575X" ], "landing_url": "https://europepmc.org/article/MED/42948623", "language": "eng", "licence": "CC-BY-NC", "licence_label": "CC BY-NC", "licence_source": "article", "licence_url": "https://creativecommons.org/licenses/by-nc/4.0/", "metadata_licence": null, "native_subjects": [ [ "heading", "1506" ], [ "heading", "Original Research" ], [ "lcc", "R" ], [ "lcc", "T58.5-58.64" ], [ "openalex", "1312|Biomedical Text Mining and Ontologies" ] ], "openalex": "W7223656390", "pdf_url": "https://europepmc.org/articles/PMC13711561?pdf=render", "pmcid": "PMC13711561", "pmid": "42948623", "published": "2026-10-26", "publisher": "BMJ Publishing Group", "ris": "https://research.draconitic.com/paper/9840kfa5e0dtz.ris", "sources": [ { "first_seen": "2026-10-31T00:20:34Z", "native_id": "MED:42948623", "source": "europepmc", "url": "https://europepmc.org/article/MED/42948623" }, { "first_seen": "2026-10-31T12:22:19Z", "native_id": "PMC13711561", "source": "pmc", "url": "https://doi.org/10.1136/bmjdh-2026-000027" }, { "first_seen": "2026-10-31T14:21:38Z", "native_id": "b8799e4902d446af8267884975d6b128", "source": "doaj", "url": "https://doi.org/10.1136/bmjdh-2026-000027" } ], "subjects": [ { "field": "biochemistry", "origin": "openalex:1312|Biomedical Text Mining and Ontologies", "subfield": "molecular-biology" } ], "title": "CLINES: Clinical LLM-based Information Extraction and Structuring Agent.", "title_text": "CLINES: Clinical LLM-based Information Extraction and Structuring Agent.", "type": "peer-reviewed", "updated_at": "2026-10-31T14:21:38Z", "url": "https://research.draconitic.com/paper/9840kfa5e0dtz", "venue": "BMJ digital health & AI", "versions": [] }