Disease Concept-Embedding Based on the Self-Supervised Method for Medical Information Extraction from Electronic Health Records and Disease Retrieval: Algorithm Development and Validation Study

Abstract

Background The electronic health record (EHR) contains a wealth of medical information. An organized EHR can greatly help doctors treat patients. In some cases, only limited patient information is collected to help doctors make treatment decisions. Because EHRs can serve as a reference for this limited information, doctors’ treatment capabilities can be enhanced. Natural language processing and deep learning methods can help organize and translate EHR information into medical knowledge and experience. Objective In this study, we aimed to create a model to extract concept embeddings from EHRs for disease pattern retrieval and further classification tasks. Methods We collected 1,040,989 emergency department visits from the National Taiwan University Hospital Integrated Medical Database and 305,897 samples from the National Hospital and Ambulatory Medical Care Survey Emergency Department data. After data cleansing and preprocessing, the data sets were divided into training, validation, and test sets. We proposed a Transformer-based model to embed EHRs and used Bidirectional Encoder Representations from Transformers (BERT) to extract features from free text and concatenate features with structural data as input to our proposed model. Then, Deep InfoMax (DIM) and Simple Contrastive Learning of Visual Representations (SimCLR) were used for the unsupervised embedding of the disease concept. The pretrained disease concept-embedding model, named EDisease, was further finetuned to adapt to the critical care outcome prediction task. We evaluated the performance of embedding using t-distributed stochastic neighbor embedding (t-SNE) to perform dimension reduction for visualization. The performance of the finetuned predictive model was evaluated against published models using the area under the receiver operating characteristic (AUROC). Results The performance of our model on the outcome prediction had the highest AUROC of 0.876. In the ablation study, the use of a smaller data set or fewer unsupervised methods for pretraining deteriorated the prediction performance. The AUROCs were 0.857, 0.870, and 0.868 for the model without pretraining, the model pretrained by only SimCLR, and the model pretrained by only DIM, respectively. On the smaller finetuning set, the AUROC was 0.815 for the proposed model. Conclusions Through contrastive learning methods, disease concepts can be embedded meaningfully. Moreover, these methods can be used for disease retrieval tasks to enhance clinical practice capabilities. The disease concept model is also suitable as a pretrained model for subsequent prediction tasks.