Heterogeneous logistic regression for estimation of subgroup effects on hypertension

Area of Science:

• Biostatistics
• Cardiovascular Disease Research
• Personalized Medicine

Background:

• Hypertension is a significant modifiable risk factor for cardiovascular disease.
• Effective hypertension prevention requires understanding complex influencing factors and subgroup characteristics.
• Personalized medicine necessitates identifying individual factor effects for tailored treatments.

Purpose of the Study:

• To develop a statistical method for analyzing population-based survey data on hypertension.
• To identify key predictors and estimate subgroup-specific effects of hypertension risk factors.
• To improve prediction accuracy for personalized hypertension prevention and treatment strategies.

Main Methods:

• Utilized heterogeneous logistic regression combined with a concave fusion penalty.
• Analyzed population-based survey data including common hypertension influencing factors.
• Employed a two-step analytic approach: predictor identification and heterogeneous effect estimation.

Main Results:

• Achieved prediction accuracy exceeding 99% for primary hypertension data.
• Demonstrated significantly higher accuracy compared to classical logistic regression (0% accuracy).
• Identified important predictors and estimated subgroup-specific effects of hypertension risk factors.

Conclusions:

• The proposed method offers a practical guide for implementing individualized hypertension management.
• Heterogeneous effects analysis is crucial for effective personalized prevention strategies.
• The advanced statistical approach significantly enhances prediction accuracy in hypertension research.