{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,13]],"date-time":"2026-03-13T05:11:29Z","timestamp":1773378689924,"version":"3.50.1"},"reference-count":34,"publisher":"Oxford University Press (OUP)","issue":"3","license":[{"start":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T00:00:00Z","timestamp":1766016000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"name":"National Key R&D Program","award":["2022YFF1203002"],"award-info":[{"award-number":["2022YFF1203002"]}]},{"name":"National Key R&D Program","award":["2022YFF0712000"],"award-info":[{"award-number":["2022YFF0712000"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["72074006"],"award-info":[{"award-number":["72074006"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"National High-Level Hospital Clinical Research Funding","award":["2024IR33"],"award-info":[{"award-number":["2024IR33"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,3,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Objectives<\/jats:title>\n                    <jats:p>To develop and evaluate a knowledge graph-augmented large language model (LLM) framework that synthesizes epidemiological evidence to infer life-course exposure-outcome pathways, using gestational diabetes mellitus (GDM) and dementia as a case study.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Materials and Methods<\/jats:title>\n                    <jats:p>We constructed a causal knowledge graph by extracting empirical epidemiological associations from scientific literature, excluding hypothetical assertions. The graph was integrated with GPT-4 through four graph retrieval-augmented generation (GRAG) strategies to infer bridging variables between early-life exposure (GDM) and later-life outcome (dementia). Semantic triples served as structured inputs to support LLM reasoning. Each GRAG strategy was evaluated by human clinical experts and three LLM-based reviewers (GPT-4o, Llama 3-70B, and Gemini Advanced), assessing scientific reliability, novelty, and clinical relevance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>The GRAG strategy using a minimal set of abstracts specifically related to GDM\u2013dementia bridging variables performed comparably to the strategy using broader sub-community abstracts, and both significantly outperformed approaches using the full GDM- or dementia-related corpus or baseline GPT-4 without external augmentation. The knowledge graph-augmented LLM identified 108 maternal candidate mediators, including validated risk factors such as chronic kidney disease and physical inactivity. The structured approach improved accuracy and reduced confabulation compared to standard LLM outputs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>Our findings suggest that augmenting LLMs with epidemiological knowledge graphs enables effective reasoning over fragmented literature and supports the reconstruction of progressive risk pathways. Expert assessments revealed that LLMs may overestimate clinical relevance, highlighting the need for human-AI collaboration in interpretation and application.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Integrating semantic epidemiological knowledge with LLMs via GRAG strategies provides a promising framework for life-course epidemiology, enabling early detection of modifiable risk factors and guiding variable selection in cohort study design.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/jamia\/ocaf219","type":"journal-article","created":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T12:34:08Z","timestamp":1764592448000},"page":"632-640","source":"Crossref","is-referenced-by-count":1,"title":["Knowledge graph-augmented large language models for reconstructing life course risk pathways: a gestational diabetes mellitus-to-dementia case study"],"prefix":"10.1093","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9430-3197","authenticated-orcid":false,"given":"Shuang","family":"Wang","sequence":"first","affiliation":[{"name":"National Institute of Health Data Science, Peking University , Beijing 100191,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yang","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Endocrinology, Peking University First Hospital , Beijing 100034,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9156-9483","authenticated-orcid":false,"given":"Ying","family":"Gao","sequence":"additional","affiliation":[{"name":"Department of Endocrinology, Peking University First Hospital , Beijing 100034,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xin","family":"He","sequence":"additional","affiliation":[{"name":"National Institute of Health Data Science, Peking University , Beijing 100191,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Guanghui","family":"Deng","sequence":"additional","affiliation":[{"name":"National Institute of Health Data Science, Peking University , Beijing 100191,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8436-778X","authenticated-orcid":false,"given":"Jian","family":"Du","sequence":"additional","affiliation":[{"name":"National Institute of Health Data Science, Peking University , Beijing 100191,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2025,12,18]]},"reference":[{"key":"2026031216444474500_ocaf219-B1","doi-asserted-by":"crossref","first-page":"e261","DOI":"10.1016\/S2468-2667(24)00018-5","article-title":"Life course epidemiology and public health","volume":"9","author":"Wagner","year":"2024","journal-title":"Lancet Public Health"},{"key":"2026031216444474500_ocaf219-B2","doi-asserted-by":"crossref","first-page":"0176","DOI":"10.34133\/hds.0176","article-title":"Moving beyond medical statistics: a systematic review on missing data handling in electronic health records","volume":"4","author":"Ren","year":"2024","journal-title":"Health Data Sci"},{"key":"2026031216444474500_ocaf219-B3","doi-asserted-by":"crossref","first-page":"1667","DOI":"10.1158\/1055-9965.EPI-17-0232","article-title":"Developing the WCRF International\/University of Bristol methodology for identifying and carrying out systematic reviews of mechanisms of exposure-cancer associations","volume":"26","author":"Lewis","year":"2017","journal-title":"Cancer Epidemiol Biomarkers Prev"},{"key":"2026031216444474500_ocaf219-B4","doi-asserted-by":"crossref","first-page":"1583","DOI":"10.1158\/1055-9965.EPI-17-0230","article-title":"A comparative study on the WCRF International\/University of Bristol methodology for systematic reviews of mechanisms underpinning exposure-cancer associations","volume":"26","author":"Ertaylan","year":"2017","journal-title":"Cancer Epidemiol Biomarkers Prev"},{"key":"2026031216444474500_ocaf219-B5","first-page":"1025","article-title":"Does testosterone mediate the relationship between vitamin D and prostate cancer progression? A systematic review and meta-analysis","volume":"33","author":"Robles","year":"2022","journal-title":"Cancer Causes Control"},{"key":"2026031216444474500_ocaf219-B6","doi-asserted-by":"crossref","first-page":"11","DOI":"10.1158\/1055-9965.EPI-21-0435","article-title":"Linking physical activity to breast cancer: text mining results and a protocol for systematically reviewing three potential mechanistic pathways","volume":"31","author":"Lynch","year":"2022","journal-title":"Cancer Epidemiol Biomarkers Prev"},{"key":"2026031216444474500_ocaf219-B7","doi-asserted-by":"crossref","first-page":"28","DOI":"10.1158\/1055-9965.EPI-21-0438","article-title":"Linking physical activity to breast cancer via sex steroid hormones, part 2: the effect of sex steroid hormones on breast cancer risk","volume":"31","author":"Drummond","year":"2022","journal-title":"Cancer Epidemiol Biomarkers Prev"},{"key":"2026031216444474500_ocaf219-B8","doi-asserted-by":"crossref","first-page":"16","DOI":"10.1158\/1055-9965.EPI-21-0437","article-title":"Linking physical activity to breast cancer via sex hormones, part 1: the effect of physical activity on sex steroid hormones","volume":"31","author":"Swain","year":"2022","journal-title":"Cancer Epidemiol Biomarkers Prev"},{"key":"2026031216444474500_ocaf219-B9","doi-asserted-by":"crossref","first-page":"1593","DOI":"10.1038\/s41591-023-02366-9","article-title":"AI-generated text may have a role in evidence-based medicine","volume":"29","author":"Peng","year":"2023","journal-title":"Nat Med"},{"key":"2026031216444474500_ocaf219-B10","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":"Yang","year":"2024","journal-title":"Nat Mach Intell"},{"key":"2026031216444474500_ocaf219-B11","first-page":"9459","author":"Lewis","year":"2020"},{"key":"2026031216444474500_ocaf219-B12","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1162\/tacl_a_00638","article-title":"Lost in the middle: how language models use long contexts","volume":"12","author":"Liu","year":"2024","journal-title":"Trans Assoc Comput Linguist"},{"key":"2026031216444474500_ocaf219-B13","author":"Edge","year":"2024"},{"key":"2026031216444474500_ocaf219-B14","doi-asserted-by":"crossref","first-page":"319","DOI":"10.1001\/jama.2024.21700","article-title":"Testing and evaluation of health care applications of large language models: a systematic review","volume":"333","author":"Bedi","year":"2025","journal-title":"JAMA"},{"key":"2026031216444474500_ocaf219-B15","doi-asserted-by":"crossref","first-page":"102","DOI":"10.1038\/s41746-024-01091-y","article-title":"Optimization of hepatological clinical guidelines interpretation by large language models: a retrieval augmented generation-based framework","volume":"7","author":"Kresevic","year":"2024","journal-title":"NPJ Digit Med"},{"key":"2026031216444474500_ocaf219-B16","doi-asserted-by":"crossref","first-page":"798","DOI":"10.1001\/jamaophthalmol.2024.2513","article-title":"Development and evaluation of a retrieval-augmented large language model framework for ophthalmology","volume":"142","author":"Luo","year":"2024","journal-title":"JAMA Ophthalmol"},{"key":"2026031216444474500_ocaf219-B17","doi-asserted-by":"crossref","first-page":"100102","DOI":"10.1016\/j.gloepi.2023.100102","article-title":"Causal reasoning about epidemiological associations in conversational AI","volume":"5","author":"Cox","year":"2023","journal-title":"Glob Epidemiol"},{"key":"2026031216444474500_ocaf219-B18","author":"Kerner","year":"2024"},{"key":"2026031216444474500_ocaf219-B19","author":"Dubey","year":"2024"},{"key":"2026031216444474500_ocaf219-B20","author":"Reid","year":"2024"},{"key":"2026031216444474500_ocaf219-B21","author":"Nature","year":"2024"},{"key":"2026031216444474500_ocaf219-B22","doi-asserted-by":"crossref","first-page":"3158","DOI":"10.1093\/bioinformatics\/bts591","article-title":"SemMedDB: a PubMed-scale repository of biomedical semantic predications","volume":"28","author":"Kilicoglu","year":"2012","journal-title":"Bioinformatics"},{"key":"2026031216444474500_ocaf219-B23","author":"Touvron","year":"2023"},{"key":"2026031216444474500_ocaf219-B24","doi-asserted-by":"crossref","first-page":"774","DOI":"10.1111\/joim.13787","article-title":"Gestational diabetes mellitus is associated with greater incidence of dementia during long-term post-partum follow-up","volume":"295","author":"Zhang","year":"2024","journal-title":"J Intern Med"},{"key":"2026031216444474500_ocaf219-B25","doi-asserted-by":"crossref","first-page":"572","DOI":"10.1016\/S0140-6736(24)01296-0","article-title":"Dementia prevention, intervention, and care: 2024 report of the lancet standing commission","volume":"404","author":"Livingston","year":"2024","journal-title":"Lancet"},{"key":"2026031216444474500_ocaf219-B26","doi-asserted-by":"crossref","first-page":"401","DOI":"10.2337\/dc23-1092","article-title":"Kidney disease in women with previous gestational diabetes mellitus: a nationwide register-based cohort study","volume":"47","author":"Christensen","year":"2024","journal-title":"Diabetes Care"},{"key":"2026031216444474500_ocaf219-B27","doi-asserted-by":"crossref","first-page":"466","DOI":"10.1016\/j.kint.2019.10.032","article-title":"Crosstalk between the nervous system and the kidney.","volume":"97","author":"Tanaka","year":"2020","journal-title":"Kidney Int"},{"key":"2026031216444474500_ocaf219-B28","doi-asserted-by":"crossref","first-page":"1032","DOI":"10.1093\/jamia\/ocaf059","article-title":"Detecting emergencies in patient portal messages using large language models and knowledge graph-based retrieval-augmented generation","volume":"32","author":"Liu","year":"2025","journal-title":"J Am Med Inform Assoc"},{"key":"2026031216444474500_ocaf219-B29","doi-asserted-by":"crossref","first-page":"e67677","DOI":"10.2196\/67677","article-title":"Improving dietary supplement information retrieval: development of a retrieval-augmented generation system with large language models","volume":"27","author":"Hou","year":"2025","journal-title":"J Med Internet Res"},{"key":"2026031216444474500_ocaf219-B30","doi-asserted-by":"crossref","DOI":"10.1093\/bioinformatics\/btae560","article-title":"Biomedical knowledge graph-optimized prompt generation for large language models","volume":"40","author":"Soman","year":"2024","journal-title":"Bioinformatics"},{"key":"2026031216444474500_ocaf219-B31","doi-asserted-by":"crossref","DOI":"10.1093\/bioinformatics\/btae353","article-title":"KRAGEN: a knowledge graph-enhanced RAG framework for biomedical problem solving using large language models","volume":"40","author":"Matsumoto","year":"2024","journal-title":"Bioinformatics"},{"key":"2026031216444474500_ocaf219-B32","doi-asserted-by":"crossref","first-page":"354","DOI":"10.1287\/msom.2023.0279","article-title":"A manager and an AI walk into a bar: does ChatGPT make biased decisions like we do?","volume":"27","author":"Chen","year":"2025","journal-title":"M&SOM"},{"key":"2026031216444474500_ocaf219-B33","author":"Qian","year":"2025"},{"key":"2026031216444474500_ocaf219-B34","author":"Adnan","year":"2025"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jamia\/advance-article-pdf\/doi\/10.1093\/jamia\/ocaf219\/66004941\/ocaf219.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/33\/3\/632\/66004941\/ocaf219.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/33\/3\/632\/66004941\/ocaf219.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,12]],"date-time":"2026-03-12T20:44:50Z","timestamp":1773348290000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/33\/3\/632\/8383560"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,18]]},"references-count":34,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2025,12,18]]},"published-print":{"date-parts":[[2026,3,1]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocaf219","relation":{},"ISSN":["1067-5027","1527-974X"],"issn-type":[{"value":"1067-5027","type":"print"},{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,3]]},"published":{"date-parts":[[2025,12,18]]}}}