Ensemble machine learning for personalized antihypertensive treatment

Abstract

Due to its prevalence and association with cardiovascular diseases and premature death, hypertension is a major public health challenge. Proper prevention and management measures are needed to effectively reduce the pervasiveness of the condition. Current clinical guidelines for hypertension provide physicians with general suggestions for first‐line pharmacologic treatment, but do not consider patient‐specific characteristics. In this study, longitudinal electronic health record data are utilized to develop personalized predictions and prescription recommendations for hypertensive patients. We demonstrate that both binary classification and regression algorithms can be used to accurately predict a patient's future hypertensive status. We then present a prescriptive framework to determine the optimal antihypertensive treatment for a patient using their individual characteristics and clinical condition. Given the observational nature of the data, we address potential confounding through generalized propensity score evaluation and optimal matching. For patients for whom the algorithm recommendation differs from the standard of care, we demonstrate an approximate 15.87% decrease in next blood pressure score based on the predicted outcome under the recommended treatment. An interactive dashboard has been developed to be used by physicians as a clinical support tool.