The combined objective is derived, which proof steps transfer from the classification setting without modification and which require adaptation, and which require adaptation on the PhysioNet 2019 Sepsis Challenge, treating the two hospital systems as sequential training fragments and the unseen system as an out-of-distribution test.
Abstract
Medical world models aim to learn a latent state of patient or organ physiology and a transition function that forecasts how that state evolves under interventions, supporting downstream tasks from imaging-based diagnosis to digital-twin treatment planning. Two failure modes threaten the reliability of such models in clinical deployment: (i)~\emph{covariate shift}, because training data are fragmented across hospitals, scanners, and time, so the feature distribution seen by the latent-dynamics predictor differs across fragments and from the distribution at deployment; and (ii)~\emph{confidence misalignment}, because multi-step forecasts are often overconfident exactly where clinical risk is highest. We argue that both problems admit a unified treatment via a single lightweight regularisation objective, \textbf{CalTwin}, which combines a Fisher-Information-based shift penalty adapted from our prior work on fragmented covariate-shift remediation~\cite{khan2025mitigating,khan2025causal} with a Confidence Misalignment Penalty adapted from our prior work on calibrated vision-language classification~\cite{khan2025confidence}, applied here to a GRU-based medical world model's latent transition predictor. We derive the combined objective, establish which proof steps transfer from the classification setting without modification and which require adaptation, and evaluate it on the PhysioNet 2019 Sepsis Challenge, treating the two hospital systems as sequential training fragments and the unseen system as an out-of-distribution test. CalTwin reduces OOD next-step latent-state MSE by 9.1\% relative to the no-penalty baseline (FIM penalty alone accounts for 7.0\%); the ECE reduction from the Confidence Misalignment Penalty is real but small (0.7\% for CalTwin, 1.3\% for CMP alone).
The Normalised Sensitivity Ratio (NSR) is introduced, a post-hoc, model-agnostic diagnostic for this question under a structured-shift regime: environments differ primarily in the mean of spurious features while the causal mechanism and causal marginals remain stable, as in multi-site clinical data or multi-batch genomics.
Athanasios Vlontzos, G. Papanastasiou, Bernhard Kainz et al.· 0 citations
BERT-LER is presented, a BERT-style model for coded EHR timelines pretrained and fine-tuned from a de-identified EHR dataset of 75 million patients, that encodes laboratory test results as discrete tokens while retaining graded information through percentile-based binning, paired with Integrated Gradients for token-level attributions grounded in the input EHR sequence.
Jun Ni Du, Lukas Adamek, Maxim A Kryukov et al.· 0 citations
Generative modeling of longitudinal Electronic Health Records is increasingly important for privacy-preserving research, yet standard autoregressive models tend to underrepresent the co-occurrence structure of tail events (i.e., diseases, symptoms), reducing the fidelity and faithfulness of generated data for rare subpopulations. To this end, we propose AdaPCLA framework, which enables generative models to adaptively fit and generate EHR data through a data distribution-aware training strategy; this is achieved by internalizing data knowledge parameters by simulated annealing training. It also supports training-free adaptation to a diverse clinical population for generation through zero-shot distribution control. Moreover, our theoretical analysis characterizes rare-code logit updates through the label-wise empirical NTK and derives a prior-internalization bound for how annealing speed and NTK conditioning affect retained prior signals. Experiments on real-world data show that AdaPCLA achieves consistent gains in tail plausibility, downstream utility, and zero-shot control; in particular, it improves TailPairSeen over HALO by 114.2% on MIMIC-III and 65.1% on MIMIC-IV, outperforms GPT-style generation by 3.5% F1 for zero-shot cross-population adaptation.
Shuai Cui, Wenxuan Chen, Wenjie Du et al.· 0 citations
Predicting chronic kidney disease progression from longitudinal records is a multi-task sequence problem: the targets are correlated, the series is irregularly sampled and sparsely observed, and standard learners fit each target independently and can therefore return mutually contradictory estimates for the same instance. We present TIDE, a Treatment-Informed Dynamics Encoder. A diagonal continuous-time state space uses the true inter-observation interval as its integration step and lets the exogenous intervention channel modulate the decay eigenvalues, so an intervention alters the rate law rather than the level. A structural coherence layer emits one latent trajectory from which five of the seven targets are derived analytically, and cause-specific hazards give the two time-to-event targets. On a 6 000-patient simulated longitudinal cohort with 107 978 visits, TIDE was compared with six baselines including gradient boosting, random forest, XGBoost and CatBoost under one pipeline, identical splits and an equal tuning budget. TIDE improved rapid-progression discrimination to 0.810 against 0.769 for the best tabular model (difference +0.040 (95% CI +0.020 to +0.062, P < 0.001)), reduced future-eGFR error from 2.48 to 2.16 mL/min/1.73 m2, raised the concordance index to 0.939, and cut internally contradictory predictions from 0.061 to 0.0005. Annual-decline regression remained below the boosted ensembles and is reported as such.
Juslin Franklin, Gopalakrishna Kanathur· Journal of Intelligent Decis...· 0 citations
This tutorial presents OneEHR, an open-source toolkit that defines a unified experiment contract for modern EHR modeling and enables head-to-head comparison among conventional, neural, LLM-based, and agentic methods through a single configuration-driven interface.
Yinghao Zhu, Zixiang Wang, Lei Gu et al.· Proceedings of the 32nd ACM...· 0 citations