Machine learning-based analyses of contributing factors for the development of hypertension: a comparative study

Area of Science:

• Cardiovascular disease research
• Biostatistics and machine learning applications
• Public health and preventative medicine

Background:

• Hypertension is a significant global health concern.
• Predictive modeling for hypertension using longitudinal data is underexplored.
• Machine learning (ML) offers advanced analytical capabilities for complex health data.

Purpose of the Study:

• To investigate the predictive performance of various ML models for new-onset hypertension.
• To identify key predictors of hypertension development from a large dataset.
• To evaluate the efficacy of ML in hypertension risk assessment.

Main Methods:

• Utilized longitudinal data from 15,965 Japanese participants undergoing annual health examinations.
• Employed five ML models: logistic regression, random forest, naïve Bayes, extreme gradient boosting, and artificial neural network.
• Assessed 58 candidate predictors, including the fatty liver index (FLI), using receiver operating characteristic curve analysis.

Main Results:

• Systolic blood pressure, age, and FLI were identified as significant predictors of hypertension via random forest feature selection.
• ML models achieved Area Under the Curve (AUC) values ranging from 0.765 to 0.825.
• The artificial neural network model demonstrated superior discriminatory capacity compared to logistic regression.

Conclusions:

• ML models effectively predict hypertension development using systolic blood pressure, age, and FLI.
• These findings support the use of ML for simple and accurate hypertension prediction.
• Developing multiple ML models may enhance the practicality of hypertension risk prediction.