{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,11]],"date-time":"2026-07-11T01:06:07Z","timestamp":1783731967181,"version":"3.55.0"},"reference-count":38,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"\n Background<\/jats:title>\n Inflammatory bowel disease (IBD) is a chronic autoimmune disorder with an increasing prevalence in the general population. Internet-based communities have become vital for communication among patients with IBD, especially throughout the COVID-19 pandemic. However, these internet-based patient-to-patient communications remain largely underexplored.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aims to analyze community posts from 3 of the largest IBD support groups on Reddit between March 1, 2020, and December 31, 2022, using a pretrained transformer model, and to validate the classification system\u2019s results via comparison to human scoring.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n We collected posts (N=53,333) from subreddits r\/CrohnsDisease, r\/UlcerativeColitis, and r\/IBD and classified them using OpenAI\u2019s GPT-3.5 Turbo model to determine sentiment, categorize topics, and identify demographic information and mentions of the COVID-19 pandemic. A subset of posts (n=397) was manually scored to measure interrater agreement between human raters and the GPT-3.5 Turbo model.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n Fleiss \u03ba and Gwet AC1 coefficients indicated a high level of agreement between raters, with values ranging from 0.53 to 0.91. The raters demonstrated almost perfect agreement on the classification of gender, with a Fleiss \u03ba of 0.91 (P<.001). Medications (14,909\/53,333) and symptoms (14,939\/53,333) emerged as the most discussed topics, and most posts conveyed a neutral sentiment. While most users did not disclose their age, those who did primarily belonged to the 20-29 years (2392\/4828) and 30-39 years (859\/4828) age groups. Based on self-reported gender, we identified 1509 men and 1502 women among our IBD Reddit users. When comparing the users on the IBD subreddits to the general IBD population, there was a significant difference in gender distribution (N=3,090,011; \u03c722=69.53; P<.001; \u03c6<0.001). After an initial spike in posts within the first month, most posts did not reference the COVID-19 pandemic.<\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n Our study showcases the potential of generative pretrained transformer models in processing and extracting insights from medical social media data. Future research can benefit from further subanalyses of our validated dataset or use OpenAI\u2019s model to analyze social media data for other conditions, particularly those for which patient experiences are challenging to collect.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/53332","type":"journal-article","created":{"date-parts":[[2025,1,26]],"date-time":"2025-01-26T13:28:54Z","timestamp":1737898134000},"page":"e53332","source":"Crossref","is-referenced-by-count":4,"title":["Exploring Inflammatory Bowel Disease Discourse on Reddit Throughout the COVID-19 Pandemic Using OpenAI\u2019s GPT-3.5 Turbo Model: Classification Model Validation and Case Study"],"prefix":"10.2196","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-7458-7186","authenticated-orcid":false,"given":"Tyler","family":"Babinski","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-9823-5592","authenticated-orcid":false,"given":"Sara","family":"Karley","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3822-5809","authenticated-orcid":false,"given":"Marita","family":"Cooper","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6416-4598","authenticated-orcid":false,"given":"Salma","family":"Shaik","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2829-3776","authenticated-orcid":false,"given":"Y Ken","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2025,7,3]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.15585\/mmwr.mm6542a3"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1093\/ibd\/izaa278"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.3389\/fmed.2021.765474"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyg.2021.652086"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1177\/17562848221127474"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1002\/smi.3080"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1093\/ibd\/izad247"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1093\/crocol\/otab044"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1007\/s42979-023-01772-7"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.2196\/45381"},{"key":"ref11","first-page":"1877","volume":"33","author":"Brown, T","year":"2020","journal-title":"Adv Neural Inf Processing Syst"},{"key":"ref12","unstructured":"GaoLBidermanSBlackSGoldingLHoppeTFosterCPhangJHeHThiteANabeshimaNPresserSLeahyCThe pile: an 800GB dataset of diverse text for language modelingArXiv20201231https:\/\/arxiv.org\/abs\/2101.00027"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyg.2023.1199058"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.3389\/feduc.2023.1272229"},{"key":"ref15","unstructured":"LamichhaneBEvaluation of ChatGPT for NLP-based mental health applicationsArXiv20230328https:\/\/arxiv.org\/abs\/2303.15727"},{"key":"ref16","unstructured":"WakeNKanehiraASasabuchiKTakamatsuJIkeuchiKBias in emotion recognition with ChatGPTArXiv20221018https:\/\/arxiv.org\/abs\/2310.11753"},{"key":"ref17","doi-asserted-by":"crossref","first-page":"161","DOI":"10.1007\/978-981-97-0293-0_12","author":"Nadi, F","year":"2024","journal-title":"Data Science and Emerging Technologies: Proceedings of DaSET 2023"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1101\/2024.03.19.24304544"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.31234\/osf.io\/8g36r"},{"key":"ref20","unstructured":"Reddit by the numbersReddit Inc2025-03-03https:\/\/www.redditinc.com\/press"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.23750\/abm.v91i1.9397"},{"key":"ref22","first-page":"830","author":"Baumgartner, J","year":"2020","journal-title":"Vol. 14 (2020): Fourteenth International AAAI Conference on Web and Social Media"},{"key":"ref23","unstructured":"Watchful1PushshiftDumpsGitHub2025-03-27https:\/\/github.com\/Watchful1\/PushshiftDumps"},{"key":"ref24","unstructured":"Alteryx2025-03-27https:\/\/www.alteryx.com\/"},{"issue":"3","key":"ref25","doi-asserted-by":"crossref","first-page":"276","DOI":"10.11613\/BM.2012.031","volume":"22","author":"McHugh, M","year":"2012","journal-title":"Biochem Med (Zagreb)"},{"key":"ref26","unstructured":"HartlingLHammMMilneAVandermeerBSantaguidaPLAnsariMTsertsvadzeAHempelSShekellePDrydenDMValidity and Inter-Rater Reliability Testing of Quality Assessment Instruments [Internet]20122025-05-06Rockville, MDAgency for Healthcare Research and Quality (US)https:\/\/www.ncbi.nlm.nih.gov\/books\/NBK92293\/pdf\/Bookshelf_NBK92293.pdf"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1186\/1471-2288-13-61"},{"key":"ref28","first-page":"1","author":"Israel, Glenn","year":"1992","journal-title":"University of Florida Cooperative Extension Service, Institute of Food and Agriculture Sciences"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1016\/j.theriogenology.2010.01.003"},{"key":"ref30","first-page":"916","author":"Yamane, T","year":"1967","journal-title":"Statistics: An Introductory Analysis"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.2427\/12267"},{"key":"ref32","unstructured":"GwetKilemirrCAC: computing chance-corrected agreement coefficients (CAC)The Comprehensive R Archive Network201910222025-03-27https:\/\/cran.r-project.org\/web\/packages\/irrCAC\/index.html"},{"key":"ref33","unstructured":"RStudio TeamRStudioPosit20202025-03-27Boston, MAhttp:\/\/www.rstudio.com\/"},{"key":"ref34","unstructured":"Social media fact sheetPew Research Center202411132024-11-13https:\/\/www.pewresearch.org\/internet\/fact-sheet\/social-media\/"},{"key":"ref35","unstructured":"Tableau DesktopTableau2025-03-27https:\/\/www.tableau.com\/products\/desktop\/download"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.2196\/44307"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1057\/s41599-021-00798-7"},{"key":"ref38","unstructured":"OpenAIAchiamJAdlerSAgarwalSAhmadLAkkayaIGPT-4 technical reportArXiv20240304https:\/\/arxiv.org\/abs\/2303.08774"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2025\/1\/e53332","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T17:38:56Z","timestamp":1782841136000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2025\/1\/e53332"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,3]]},"references-count":38,"URL":"https:\/\/doi.org\/10.2196\/53332","relation":{"has-preprint":[{"id-type":"doi","id":"10.2196\/preprints.53332","asserted-by":"object"}]},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,3]]}}}