Cure8 research brief
Registries that automatically extract clinical data with AI/NLP can speed research, help identify patient cohorts for studies or trials, and may eventually support better clinical decision-making for people with IBD. Improved data capture can highlight gaps in outcomes, treatment responses, and adverse events.
Researchers building IBD cohorts, clinicians and data teams interested in registry-based research, and patients curious about how electronic health data may be used to support IBD research or trial recruitment.
A new IBD registry called the IBD Data Lake was developed using machine learning and natural language processing (NLP) to extract structured and unstructured data from electronic medical records and assemble a searchable database.
The authors validated the system against manual chart review (208 patients, 104 IBD and 104 matched non-IBD) and report high accuracy for identifying IBD cases and for extracting key attributes such as diagnosis (Crohn’s vs ulcerative colitis), smoking status, and extraintestinal manifestations.
The report describes use of a cloud-based secure infrastructure, a custom user interface, and an NLP service (Comprehend Medical) to process clinical documents.
Performance metrics vs chart review included sensitivity ~98% and specificity ~97% for IBD identification; disease classification and several clinical features showed similarly high sensitivity and specificity. The article frames this registry as a tool to enable real-time cohort identification and research recruitment.
This brief is grounded in the article abstract provided; Cure8 did not review the full peer-reviewed manuscript beyond the supplied text.
This classification and summary are based on the source abstract (structured content depth: abstract). Comprehend Medical and other commercial NLP tools vary by site and data quality; performance reported here was validated against chart review in a single dataset (208 patients). Results may differ in other health systems or larger populations.
Review the original publication for the complete reporting, methods, and context.
Conflict statement: We have read the journal’s policy and the authors of this manuscript have the following competing interests. JLCK, CM, ECJ, IK, and GL have no relevant conflicts of interest to disclose. BB has served on advisory boards and received speaker fees from Ferring, Janssen, Abbvie, Takeda, Pfizer, BMS, Merck, Sandoz, Organon, Lifelabs, Celltrion, Alimentiv, Gilead, Iterative Health, Celgene, Merck, Amgen, Pendopharm, Eli Lilly, Fresenius Kabi, Mylan, Viatris, Bausch Health, BioJamp Pharma, and Eupraxia. BB has received research support, though not directly for this project, from Janssen, Abbvie, GSK, BMS, Amgen, Genentech, and Merck. BB has Qu Biologic stock options. YL has served on advisory boards and received speaker fees from Janssen, Abbvie, Organon, Pfizer, Takeda, BMS, Amgen, Celltrion, and Eli Lilly. GR has received honoraria from Abbvie, Frensius-Kabi, Janssen, Pfizer, Takeda, Merck, Amgen, Viatris, Organon and Ferring as a speaker, adviser, and consultant. GR has research grant support, though not directly for this study, from Abbvie, Pfizer, Ferring and Crohn’s and Colitis Canada.
This Cure8 brief is based on source text from the linked article. Cure8 is informational only and is not a substitute for professional medical advice, diagnosis, or treatment.
Genetic ancestry can influence IBD risk and disease course, and most prior studies focused on European populations.
A validated AI biomarker that predicts which IBD treatments are likely to work could reduce trial-and-error in treatment choices, potentially shortening time to effective therapy for people with Crohn's disease or...
This project aims to improve prediction of autoimmune disease onset and immunotherapy-related autoimmunity by combining genetics and plasma proteins.
