The application of unsupervised deep learning in predictive models using electronic health records

Abstract

Background:The main goal of this study is to explore the use of features representing patient-level electronic health record (EHR) data, generated by the unsupervised deep learning algorithm autoencoder, in predictive modeling. Since autoencoder features are unsupervised, this paper focuses on their general lower-dimensional representation of EHR information in a wide variety of predictive tasks. Methods: We compare the model with autoencoder features to traditional models: logistic model with least absolute shrinkage and selection operator (LASSO) and Random Forest algorithm. In addition, we include a predictive model using a small subset of response-specific variables (Simple Reg) and a model combining these variables with features from autoencoder (Enhanced Reg). We performed the study first on simulated data that mimics real world EHR data and then on actual EHR data from eight Advocate hospitals. Results: On simulated data with incorrect categories and missing data, the precision for autoencoder is 24.16% when fixing recall at 0.7, which is higher than Random Forest (23.61%) and lower than LASSO (25.32%). The precision is 20.92% in Simple Reg and improves to 24.89% in Enhanced Reg. When using real EHR data to predict the 30-day readmission rate, the precision of autoencoder is 19.04%, which again is higher than Random Forest (18.48%) and lower than LASSO (19.70%). The precisions for Simple Reg and Enhanced Reg are 18.70 and 19.69% respectively. That is, Enhanced Reg can have competitive prediction performance compared to LASSO. In addition, results show that Enhanced Reg usually relies on fewer features under the setting of simulations of this paper. Conclusions: We conclude that autoencoder can create useful features representing the entire space of EHR data and which are applicable to a wide array of predictive tasks. Together with important response-specific predictors, we can derive efficient and robust predictive models with less labor in data extraction and model training.

Patient Representation From Structured Electronic Medical Records Based on Embedding Technique: Development and Validation Study

“…Representation for structured medical data is critical for data mining tasks in the medical domain [ 3 , 5 , 6 , 14 ]. The one-hot code scheme is a simple and widely used representation.…”

“…The past decade has witnessed an explosion in the amount of digital information stored in electronic medical records (EMRs), which contain massive quantities of information on the clinical history of patients. The wide secondary use of this information for various clinical applications has become a prevalent trend [1], helping to make diagnostic decisions [2][3][4], predict patient outcomes [5][6][7][8], and provide treatment recommendations [9][10][11].…”

“…Ning et al [ 15 ] trained vector representations for medical concepts from biomedical journal articles through Skip-gram and proposed a fully automated feature extraction method for disease phenotyping based on the medical concept vector representation. Moreover, some researchers learned patient representation through deep learning [ 3 , 5 , 12 ]. Zhe Wang et al [ 5 ] designed a feature rearrangement representation based on the convolutional neural network for heart failure mortality prediction.…”

“…Zhe Wang et al [ 5 ] designed a feature rearrangement representation based on the convolutional neural network for heart failure mortality prediction. Lei Wang et al [ 3 ] used autoencoder, an unsupervised deep learning algorithm, to generate lower-dimensional representations from EMR data in various predictive tasks such as readmission prediction and pneumonia prediction. A similar study [ 12 ] used the recurrent neural network–based denoising autoencoder to encode patient records into low-dimensional and dense vectors for heart failure prediction.…”

Clinical applications of artificial intelligence in cardiology on the verge of the decade

“…Such an approach may lead to the development of more specific treatment strategies according to the paradigm of precision medicine [26]. Very recently, Wang et al [27] used an unsupervised autoencoder based on a deep learning algorithm to represent data from electronic health records and compared the classification based on these learned features with more conventional approaches.…”