Multimodal fusion network for ICU patient outcome prediction.

Over the past decades, massive Electronic Health Records (EHRs) have been accumulated in Intensive Care Unit (ICU) and many other healthcare scenarios. The rich and comprehensive information recorded presents an exceptional opportunity for patient outcome predictions. Nevertheless, due to the diversity of data modalities, EHRs exhibit a heterogeneous characteristic, raising a difficulty to organically leverage information from various modalities. It is an urgent need to capture the underlying correlations among different modalities. In this paper, we propose a novel framework named Multimodal Fusion Network (MFNet) for ICU patient outcome prediction. First, we incorporate multiple modality-specific encoders to learn different modality representations. Notably, a graph guided encoder is designed to capture underlying global relationships among medical codes, and a text encoder with pre-fine-tuning strategy is adopted to extract appropriate text representations. Second, we propose to pairwise merge multimodal representations with a tailored hierarchical fusion mechanism. The experiments conducted on the eICU-CRD dataset validate that MFNet achieves superior performance on mortality prediction and Length of Stay (LoS) prediction compared with various representative and state-of-the-art baselines. Moreover, comprehensive ablation study demonstrates the effectiveness of each component of MFNet.

Evaluating the Predictive Features of Person-Centric Knowledge Graph Embeddings: Unfolding Ablation Studies.

Developing novel predictive models with complex biomedical information is challenging due to various idiosyncrasies related to heterogeneity, standardization or sparseness of the data. We previously introduced a person-centric ontology to organize information about individual patients, and a representation learning framework to extract person-centric knowledge graphs (PKGs) and to train Graph Neural Networks (GNNs). In this paper, we propose a systematic approach to examine the results of GNN models trained with both structured and unstructured information from the MIMIC-III dataset. Through ablation studies on different clinical, demographic, and social data, we show the robustness of this approach in identifying predictive features in PKGs for the task of readmission prediction.

Graph-guided deep hashing networks for similar patient retrieval.

With the rapid growth and widespread application of electronic health records (EHRs), similar patient retrieval has become an important task for downstream clinical decision support such as diagnostic reference, treatment planning, etc. However, the high dimensionality, large volume, and heterogeneity of EHRs pose challenges to the efficient and accurate retrieval of patients with similar medical conditions to the current case. Several previous studies have attempted to alleviate these issues by using hash coding techniques, improving retrieval efficiency but merely exploring underlying characteristics among instances to preserve retrieval accuracy. In this paper, drug categories of instances recorded in EHRs are regarded as the ground truth to determine the pairwise similarity, and we consider the abundant semantic information within such multi-labels and propose a novel framework named Graph-guided Deep Hashing Networks (GDHN). To capture correlation dependencies among the multi-labels, we first construct a label graph where each node represents a drug category, then a graph convolution network (GCN) is employed to derive the multi-label embedding of each instance. Thus, we can utilize the learned multi-label embeddings to guide the patient hashing process to obtain more informative and discriminative hash codes. Extensive experiments have been conducted on two datasets, including a real-world dataset concerning IgA nephropathy from Peking University First Hospital, and a publicly available dataset from MIMIC-III, compared with traditional hashing methods and state-of-the-art deep hashing methods using three evaluation metrics. The results demonstrate that GDHN outperforms the competitors at different hash code lengths, validating the superiority of our proposal.

Stable clinical risk prediction against distribution shift in electronic health records.

The availability of large-scale electronic health record datasets has led to the development of artificial intelligence (AI) methods for clinical risk prediction that help improve patient care. However, existing studies have shown that AI models suffer from severe performance decay after several years of deployment, which might be caused by various temporal dataset shifts. When the shift occurs, we have access to large-scale pre-shift data and small-scale post-shift data that are not enough to train new models in the post-shift environment. In this study, we propose a new method to address the issue. We reweight patients from the pre-shift environment to mitigate the distribution shift between pre- and post-shift environments. Moreover, we adopt a Kullback-Leibler divergence loss to force the models to learn similar patient representations in pre- and post-shift environments. Our experimental results show that our model efficiently mitigates temporal shifts, improving prediction performance.