Multi-layer Representation Learning and Its Application to Electronic Health Records

Abstract

Electronic Health Records (EHRs) are digital records associated with hospitalization, diagnosis, medications and so on. Secondary use of EHRs can promote the clinical informatics applications and the development of healthcare undertaking. EHRs have the unique characteristic where the patient visits are temporally ordered but the diagnosis codes within a visit are randomly ordered. The hierarchical structure requires a multi-layer network to explore the different relational information of EHRs. In this paper, we propose a Multi-Layer Representation Learning method (MLRL), which is capable of learning effective patient representation by hierarchically exploring the valuable information in both diagnosis codes and patient visits. Firstly, MLRL utilizes the multi-head attention mechanism to explore the potential connections in diagnosis codes, and a linear transformation is implemented to further map the code vectors to non-negative real-valued representations. The initial visit vectors are then obtained by summarizing all the code representations. Secondly, the proposed method combines Bidirectional Long Short-Term Memory with self-attention mechanism to learn the weighted visit vectors which are aggregated to form the patient representation. Finally, to evaluate the performance of MLRL, we apply it to patient's mortality prediction on real EHRs and the experimental results demonstrate that MLRL has a significant improvement in prediction performance. MLRL achieves around 0.915 in Area Under Curve which is superior to the results obtained by baseline methods. Furthermore, compared with raw data and other data representations, the learned representation with MLRL shows its outstanding results and availability on multiple different classifiers.

Improving the Performance of Outcome Prediction for Inpatients With Acute Myocardial Infarction Based on Embedding Representation Learned From Electronic Medical Records: Development and Validation Study

“…In addition to the representation algorithm, features used to represent a patient were also critical. Many previous studies focused on some features in the original form of medical codes, such as disease diagnoses, procedures, and medications [ 1 , 11 , 14 , 37 ]. For laboratory tests that contained much diagnosis and prognosis-relevant information about patients, we included the normal status of the laboratory tests into the feature sets, rather than simply using the number of laboratory tests and test co-occurrences [ 12 , 38 ].…”

“…It includes patient health information, such as demographics, vital signs, laboratory test results, medications, procedures, diagnosis codes, and clinical notes. The informative MIMIC-III data set was widely used in some medical machine learning modeling and algorithm evaluations, providing strong data support for researchers to establish models and evaluate algorithms [ 14 , 18 ].…”

“…EMR data can be used to not only reflect the health status of patients and record the treatment trajectory, but also help doctors in making clinical decisions [1][2][3][4][5][6] and improving the efficiency of diagnosis and treatment [1,7,8]. One of the most prevalent and practical tasks of the secondary use of EMR data is building models to predict the disease status [8][9][10] and treatment outcomes [11][12][13][14][15][16][17] for a patient, using machine learning algorithms.…”

Enhancing Parkinson’s Disease Prediction Using Deep Learning-Based Convolutional Neural Networks

“…4: CNN Architecture Efforts have gone into not only designing this proprietary CNN architecture but also methodically tweaking its hyperparameters and implementing a durable training approach. These rigorous techniques have been applied to guarantee not only the model's functionality as well as its steadfast commitment towards accuracy in prediction of PD [18]…”

“…Fig.4: CNN Architecture Efforts have gone into not only designing this proprietary CNN architecture but also methodically tweaking its hyperparameters and implementing a durable training approach. These rigorous techniques have been applied to guarantee not only the model's functionality as well as its steadfast commitment towards accuracy in prediction of PD[18][25]. A complete set of performance indicators to evaluate the model's usefulness, including accuracy, a Receiver Operating Characteristic (ROC) curve, F1-score, precision, and recall are employed.…”