Icu delirium prediction improved by combining vital signs and clinical notes
DeliMF: Gated Cross-Attention Fusion of Structured Time Series and Clinical Text for Early Intensive Care Delirium Prediction
Machine Learning
Summary
Delirium is a common, serious problem for patients in intensive care, and spotting it early can help doctors treat it better. The authors created a system called DeliMF that looks at two kinds of patient information: numbers like vital signs and test results, and doctors' written notes. Their system smartly combines both kinds of information and can still work if some notes are missing. It predicts delirium better than methods using only numbers, which means it could help catch delirium earlier.
What this means in practice
- •For hospital data teams: Integrate multimodal ICU patient data including clinical notes with vital signs for better delirium risk prediction within 24 hours of admission.
- •For healthcare software developers: Develop ICU monitoring tools that incorporate gated cross-attention fusion of structured data and free-text reports to enhance early detection of delirium.$Commercial implications: Enables creation of advanced ICU decision support products that outperform existing vital-sign-only delirium predictors.
Tested on one dataset.
Authors
Yibo Yang, Zerui Huang, Haomiao Yang, Zixia Wang
Abstract
Delirium affects a large fraction of intensive care unit (ICU) patients and foreshadows longer stays, lasting cognitive decline, and higher mortality. Flagging the patients who will become delirious within the first day of admission would give clinicians a window to intervene. Yet state-of-the-art predictors rely only on the structured electronic health record (EHR) of vital signs, laboratory values, and medications, while overlooking the free-text clinical notes written in the same window, which carry information the numbers cannot encode. To this end, we present DeliMF, a multimodal framework that fuses the structured time series with clinical text. It encodes the structured signals with a temporal Transformer and the time-stamped radiology reports with a clinical text encoder, and fuses the two through a gated cross-attention module. The gate learns a per-patient weight on the text; when a patient has no note the gate closes and DeliMF reduces to its structured-only form, giving robustness to missing text by construction. On an incident-only MIMIC-IV cohort, DeliMF improves significantly over the baselines, particularly on patients who carry clinical text, while losing nothing on those without. Ablation studies confirm the contribution of multimodal fusion and gated cross-attention, and a case study further shows that fusing clinical text with the structured stream recovers delirium cases that the structured data alone misses. We hope DeliMF encourages early ICU prediction to make fuller use of the overlooked clinical text and contributes to the early identification and timely intervention of delirium.