Search

The Berlin definition of ARDS provides standardized diagnostic criteria based on acute onset within one week of a known insult, bilateral chest imaging opacities not explained by other causes, respiratory failure not due to cardiac issues or fluid overload, and impaired oxygenation measured by the PaO₂/FiO₂ ratio, enabling consistent identification in intensive care; however, its clinical use is limited by variability in imaging interpretation and the need for rapid decision-making, often causing delays and inconsistent diagnoses. Current practice relies heavily on subjective assessment of chest X-rays and limited integration of clinical notes and laboratory trends, resulting in moderate inter-observer agreement and reduced diagnostic reliability. To overcome these challenges, a multimodal transformer framework is proposed that integrates chest X-rays, clinical notes, and laboratory data using vision transformers, BERT-based text encoders, and temporally aware lab embeddings, with cross-modal attention enabling interaction across data types and a fusion module producing final ARDS probability estimates. This integrated approach improves diagnostic accuracy by combining complementary information, enhances interpretability through attention mechanisms, and offers a more objective and timely method for ARDS detection, with potential to support earlier intervention and better outcomes in critically ill patients.

Adverse drug reactions (ADRs) are a major global health issue, contributing to significant morbidity, mortality, and healthcare costs. Many ADRs are detected only after widespread drug use, reflecting limitations of pre-market trials in capturing real-world patient variability. Although electronic health records (EHRs) collected between 2017 and 2023 provide rich data for post-market surveillance, they remain underused for systematic ADR detection. Current pharmacovigilance methods rely heavily on spontaneous reporting systems, which suffer from underreporting and bias, while supervised machine learning approaches require labeled ADR data that are often unavailable for rare or novel events. This paper proposes a variational autoencoder (VAE)-based unsupervised framework to detect ADR signals from multimodal EHR data, including clinical notes and laboratory results. The model learns normal patient data distributions and identifies deviations as potential safety signals without requiring labeled ADR examples. A multimodal architecture combines natural language processing of clinical notes with structured laboratory encoders, forming a shared latent space for anomaly detection based on reconstruction error. The framework enables detection of unknown ADRs by flagging abnormal patterns in patient records across large datasets from 2017 to 2023. Its unsupervised nature makes it suitable for identifying rare or previously unrecognized drug safety issues. Overall, this approach offers a scalable, proactive pharmacovigilance strategy that shifts drug safety monitoring from reactive reporting to predictive detection using routine EHR data.