Predicting hypertension onset from longitudinal electronic health records with deep learning

Abstract

Objective Hypertension has long been recognized as one of the most important predisposing factors for cardiovascular diseases and mortality. In recent years, machine learning methods have shown potential in diagnostic and predictive approaches in chronic diseases. Electronic health records (EHRs) have emerged as a reliable source of longitudinal data. The aim of this study is to predict the onset of hypertension using modern deep learning (DL) architectures, specifically long short-term memory (LSTM) networks, and longitudinal EHRs. Materials and Methods We compare this approach to the best performing models reported from previous works, particularly XGboost, applied to aggregated features. Our work is based on data from 233 895 adult patients from a large health system in the United States. We divided our population into 2 distinct longitudinal datasets based on the diagnosis date. To ensure generalization to unseen data, we trained our models on the first dataset (dataset A “train and validation”) using cross-validation, and then applied the models to a second dataset (dataset B “test”) to assess their performance. We also experimented with 2 different time-windows before the onset of hypertension and evaluated the impact on model performance. Results With the LSTM network, we were able to achieve an area under the receiver operating characteristic curve value of 0.98 in the “train and validation” dataset A and 0.94 in the “test” dataset B for a prediction time window of 1 year. Lipid disorders, type 2 diabetes, and renal disorders are found to be associated with incident hypertension. Conclusion These findings show that DL models based on temporal EHR data can improve the identification of patients at high risk of hypertension and corresponding driving factors. In the long term, this work may support identifying individuals who are at high risk for developing hypertension and facilitate earlier intervention to prevent the future development of hypertension.

Predicting disease onset from electronic health records for population health management: a scalable and explainable Deep Learning approach

“…Our study confirmed that SHAP analysis has the potential to improve the explainability of disease prediction models trained using embedded EHR data. In line with similar research (Datta et al, 2022 ), our analysis found a combination of established disease predictors and novel features identified by the model. For example, our COPD predictors included tobacco dependence and expected comorbidities (e.g., hypertension) but also features such as chest x-rays which may correlate with likelihood of a COPD diagnosis.…”

“…The Hi-BEHRT authors included type 1 diabetes in their cohort and used a 5 year prediction window (Li et al, 2022 ). The ROC AUC for predicting the 3-year risk of hypertension (0.92) is comparable to that achieved by Datta et al ( 2022 ) predicting 2-year risk of hypertension with a similar model architecture (0.90). Further research would be required to systematically evaluate the relative performance of our approach to others.…”

Prognostic risk models for incident hypertension: A PRISMA systematic review and meta-analysis

“…The top eight most used variables in the literature were the same used in the Framingham risk model [ 18 ]. Note, in five studies applying ML methods, complete information about the variables used in final models was not presented [ 45 , 48 , 50 , 53 , 70 ]. A summarized view of predictors used in studies can be seen in Fig 2 .…”

“…This suggests that reporting should be even more rigorous. As an example of the opposite, in three studies [ 50 , 53 , 70 ] developing risk models using EHRs, a complete list of variables used in the final models was not reported.…”

Prediction of cardiovascular and renal risk among patients with apparent treatment‐resistant hypertension in the United States using machine learning methods

“…Recent studies have used EHRs and machine learning methods to predict risks in a variety of health care settings, including risk of hospital readmissions after stroke, 27 development of transthyretin amyloid cardiomyopathy in patients with HF, 28 hospitalization in children with complex health needs, 29 and personalized breast cancer prediction. 30 In the field of cardiovascular medicine, recent studies have used EHR‐based models to predict the risk of developing vascular complications in patients with prediabetes or diabetes, 31 developing hypertension, 32 , 33 and occurrence of stroke in patients with hypertension. 34 Similar studies that developed predictive risk models using insurance claims data have investigated risks for adverse cardiovascular and chronic renal outcomes among patients with type 2 diabetes.…”