Highly precise risk prediction model for new‐onset hypertension using artificial intelligence techniques

Abstract

It has been estimated that hypertension is the cause of approximately 13% of all deaths worldwide each year. 1 In light of the world's population growth and aging demographic, hypertension is a global burden, along with other cardiovascular and age-related diseases. 2,3 Early intervention with lifestyle modifications and treatment of "prehypertension" may reduce the incidence and long-term consequences of clinical hypertension. [4][5][6][7] Recent guidelines lowered the recommended thresholds for diagnosing hypertension or abnormal "elevated blood pressure (BP)" and the BP goal during antihypertensive therapy. 8-10 Therefore, the ability to predict an individual's risk of developing hypertension would be helpful for clinicians. They could then plan and prescribe personalized lifestyle modifications or make therapeutic decisions designed to prevent or postpone the development of hypertension. There are several models available to predict the risk of new-onset hypertension; these have been developed in Western and Asian countries using traditional statistical methods (eg, Cox regression or logistic regression). 11,12 Arterial stiffness is increasingly being recognized as making an important contribution to increases in systolic BP (SBP) and the development of hypertension in general populations, independent of traditional hypertension risk factors. [13][14][15][16] In addition, arterial AbstractHypertension is a significant public health issue. The ability to predict the risk of developing hypertension could contribute to disease prevention strategies. This study used machine learning techniques to develop and validate a new risk prediction model for new-onset hypertension. In Japan, Industrial Safety and Health Law requires employers to provide annual health checkups to their employees. We used 2005-2016 health checkup data from 18 258 individuals, at the time of hypertension diagnosis [Year (0)] and in the two previous annual visits [Year (−1) and Year (−2)].Data were entered into models based on machine learning methods (XGBoost and ensemble) or traditional statistical methods (logistic regression). Data were randomly split into a derivation set (75%, n = 13 694) used for model construction and development, and a validation set (25%, n = 4564) used to test performance of the derived models. The best predictor in the XGBoost model was systolic blood pressure during cardio-ankle vascular index measurement at Year (−1). Area under the receiver operator characteristic curve values in the validation cohort were 0.877, 0.881, and 0.859 for the XGBoost, ensemble, and logistic regression models, respectively. We have developed a highly precise prediction model for future hypertension using machine learning methods in a general normotensive population. This could be used to identify at-risk individuals and facilitate earlier non-pharmacological intervention to prevent the future development of hypertension.

The HOPE Asia Network activity for “zero” cardiovascular events in Asia: Overview 2020

“…The HOPE Asia Network has published a number of papers on important Asia‐specific topics in hypertension . This Asia‐specific issue of the Journal of Clinical Hypertension provides an important resource including original research and review articles on the management of hypertension from a variety of Asian viewpoints, including topics such as central BP, environmental factors, comorbidities such as diabetes, atrial fibrillation, heart failure, and dementia, recent Asian guidelines and so forth . This issue also includes reports from twelve countries/regions in Asia to help define and highlight the current status of hypertension, its management, and their initiatives.…”

Applications of artificial intelligence for hypertension management

“…Health data provide valuable information for AI applications in assisting health assessment and disease prognosis. Current evidence has revealed AI's efficacy on identification of hypertensive status, [12][13][14] prediction for incidence, [15][16][17] and outcomes from hypertension. 18,19 Masked hypertension, with a reported prevalence ranging from 9% to 30% 20,21 globally, persists as an obstacle for accurate hypertension diagnosis which mainly relies on elevated office blood pressure (BP) level.…”

“…Health data provide valuable information for AI applications in assisting health assessment and disease prognosis. Current evidence has revealed AI’s efficacy on identification of hypertensive status, 12‐14 prediction for incidence, 15‐17 and outcomes from hypertension 18,19 …”

Management of Hypertension in the Digital Era

“… 31 Using data from 18 258 individuals, machine learning methods were able to develop a highly precise model to predict the development of hypertension in a general population. 34 This could be used to identify individuals at risk of developing hypertension and allow early intervention with tools such as lifestyle modification to prevent future hypertension development. Another artificial intelligence-based prediction model using multi-input multi-output deep neural networks based on time-series BP data and related contextual information showed that this approach has the potential to predict both mean BP and BPV.…”

“… 32 In addition, the availability of big data collected by wearable BP monitoring could facilitate time-series analyses and be used to inform artificial intelligence strategies to predict hypertension. 33 , 34 …”