Multimodal Forecasting of Sparse Intraoperative Hypotension Events Powered by Language Model

ICLR 2026 Conference Submission15646 Authors

Keywords: Time Series Forecasting; Language Models; Multimodal Representation Learning

Abstract: Intraoperative hypotension (IOH) is a common complication of general anesthesia and is strongly associated with adverse outcomes such as myocardial injury and increased mortality. Despite its significance, IOH prediction is hindered by event sparsity and the challenge of integrating heterogeneous static attributes and dynamic physiological signals. In this paper, we propose a multimodal language model framework IOHFuseLM. To accurately identify and differentiate sparse hypotensive events, we leverage a two-stage training strategy. The first stage involves domain adaptive pretraining on IOH physiological time series augmented through diffusion methods, thereby enhancing the sensitivity to patterns associated with hypotension. Subsequently, task fine-tuning is performed on the original clinical dataset to further enhance the ability to distinguish normotensive from hypotensive states. To enable multimodal fusion for each patient, we align structured clinical descriptions with the corresponding physiological time series at the token level. Such alignment enables the model to capture individualized temporal patterns alongside their corresponding clinical semantics. In addition, we transform static patient attributes into structured text to enrich personalized information. Experiments on two intraoperative datasets and one arrhythmia dataset demonstrate that IOHFuseLM outperforms baselines in IOH identification and generalizes effectively to abnormal heartbeat detection, underscoring its potential as a versatile solution across physiological domains. Our code is publicly available to promote reproducibility at https://anonymous.4open.science/r/IOHFuseLM-C5A4.

Primary Area: learning on time series and dynamical systems

Submission Number: 15646