Multi-Large Language Model Orchestrated Severity Assessment of Clinical Records (MOSAIC)
2026-07-06 • Computation and Language
Computation and Language
AI summaryⓘ
The authors developed MOSAIC, a two-step AI system using large language models to assess how severe type 2 diabetes is from electronic health records. They tested MOSAIC on synthetic patient data and found it could capture complex information, like blood sugar levels and social factors, better than traditional rule-based methods. MOSAIC's results corresponded reasonably well with existing clinical severity scores and were linked to patient outcomes like mortality. The study shows that such AI systems can reason beyond fixed rules to produce meaningful health severity categories. The authors suggest this approach could be useful for other diseases too.
Electronic Health Records (EHR)Large Language Models (LLM)Type 2 Diabetes (T2D)Severity PhenotypingDCSI (Diabetes Complications Severity Index)Agentic SystemsMortalitySynthetic CohortLog-rank TestSocial Determinants of Health
Authors
Manuela Del Castillo Suero, Arnault-Quentin Vermillet, Nicole Sonne Heckmann, Darmendra Ramcharran, Maurizio Sessa
Abstract
Background: Disease severity is a multidimensional construct difficult to capture with rule-based approaches in Electronic Healthcare Records (EHR). Agentic large language model (LLM) systems could synthesise clinical evidence and reason over EHRs, but remain unevaluated for this task. Methods: MOSAIC is a two-phase agentic LLM framework for severity phenotyping, using type 2 diabetes (T2D) as a proof-of-concept. MOSAIC was evaluated on a synthetic cohort (SyntheticMass; open-weight N = 4,886; closed-weight N = 200) against three algorithmic ground truths (DCSI, DiSSCo, Cooper) and against all-cause mortality and incident complications. Open-weight (locally deployable) and proprietary pipelines were also compared. Results: The generated framework spanned domains absent from the comparators, including biomarker-based glycaemic staging, beta-cell function, and social determinants of health. Open-weight MOSAIC matched the proprietary pipeline (closed- vs open-weight weighted kappa = 0.773) and reached moderate agreement with Cooper (kappa = 0.597) and DCSI (kappa = 0.534) and fair agreement with DiSSCo (kappa = 0.320). Agent-based (Type 1) tiers showed significant separation of all-cause mortality (log-rank p < 0.001; crude hazard ratios 1.6-2.4 for non-Baseline tiers), with non-monotonic separation at the upper tiers, and an inverse gradient for incident complications (log-rank p < 0.001) consistent with depletion of susceptibles. Agentic classification also diverged from deterministic execution of the same rubric (MOSAIC Frozen; kappa = 0.428), indicating reasoning beyond fixed rules. Conclusion: MOSAIC shows agentic LLM systems can generate and apply clinically meaningful severity phenotypes from structured EHR data in T2D. Extending it to other diseases with similarly multidimensional severity warrants further research.