Agentic LLM reasoning in pediatric diabetes outcome prediction from multimodal medical records
In the authors' words
Large language model (LLM) agents are proposed for clinical prediction but are untested in pediatrics, where cohorts are small and outcomes are uncommon. Using records from 2,489 youth with type 2 diabetes (T2D) and 4,018 with type 1 diabetes (T1D), spanning structured data, note-derived social determinants and, in T1D, continuous glucose monitoring data, we predicted seven T2D outcomes, four also in T1D, at 24 months from 18 months of follow-up. CEDAR (Clinical Evidence-Driven Agent for Risk-prediction), an evidence-citing, self-auditing LLM pipeline, outperformed trained machine learning in both cohorts (pooled area under the receiver operating characteristic curve 0.815 versus 0.656 to 0.700 in T2D and 0.756 versus 0.603 to 0.631 in T1D), most on the sparsely screened outcomes. Only 4 of 70,246 values it cited did not match the chart. Reading only the evidence and reasoning cited by CEDAR, two pediatric endocrinologists reached the same prediction as CEDAR in about nine of ten cases, agreeing with each other on 90.5%.
In plain words AI summary
The authors built CEDAR, an AI system that cites evidence and checks its own work, to predict diabetes outcomes in children. They used medical records from 2,489 youth with type 2 diabetes and 4,018 with type 1 diabetes. The system cited 70,246 pieces of evidence, and only 4 did not match the actual medical charts. Two pediatric diabetes specialists agreed with CEDAR's prediction in about nine out of ten cases when shown only the evidence CEDAR used.
Written by Kimi K2.5 using only the abstract above. Jev checked that every sentence is supported by the abstract sentences it cites, and removed any that were not. Point at a sentence to highlight its source. 1 sentence was removed for not being supported by the abstract.
Appeared: Friday, October 30. arXiv. Preprint, not yet peer-reviewed.
Authors' comment: 18 pages main text plus 40 pages Supplementary Information