An Application of Machine Learning to Etiological Diagnosis of Secondary Hypertension: Retrospective Study Using Electronic Medical Records (Preprint)
BACKGROUND Secondary hypertension is a kind of hypertension with definite etiology and may be cured. Patients with suspected secondary hypertension can benefit from detection and treatment in time and, conversely, will have higher risk of morbidity and mortality than patients with primary hypertension. OBJECTIVE The aim of this study was to develop and validate machine learning (ML) prediction models of common etiologies in patients with suspected secondary hypertension. METHODS The analyzed dataset was retrospectively extracted from electronic medical records (EMRs) of patients discharged from Fuwai hospital between January 1, 2016 and June 30, 2019. A total of 7532 unique patients were included and divided into two datasets by time: 6302 patients in 2016-2018 as training dataset for model building and 1230 patients in 2019 as validation dataset for further evaluation. Extreme Gradient Boosting (XGBoost) was adopted to develop five prediction models of four etiologies of secondary hypertension and occurrence of any of them, including renovascular hypertension (RVH), primary aldosteronism (PA), thyroid dysfunction and aortic stenosis. Both univariate logistic analysis and Gini impure method were used for feature selection, while grid search and 10-fold cross-validation were used to select the optimal hyperparameters for each model. RESULTS Validation of the composite outcome prediction model showed good performance with an area under the receiver-operating characteristic curve (AUC) of 0.924 in the validation dataset, while the four prediction models of RVH, PA, thyroid dysfunction and aortic stenosis achieved AUC of 0.938, 0.965, 0.959, 0.946, respectively, in the validation dataset. 79 clinical indicators were identified in all and finally used in our prediction models. The result of subgroup analysis on the composite outcome prediction model demonstrated high discrimination with AUCs all higher than 0.890 among all age groups of adults. CONCLUSIONS The ML prediction models in this study showed good performance in detecting four etiologies of patients with suspected secondary hypertension, thus they may potentially facilitate clinical diagnosis decision making of secondary hypertension in an intelligent way. CLINICALTRIAL