Journals Information
Universal Journal of Public Health Vol. 13(2), pp. 417 - 426
DOI: 10.13189/ujph.2025.130213
Reprint (PDF) (659Kb)
Cost-Sensitive Boosting Approach to Analyzing Risk Factors Affecting Health-Related Quality of Life in Hypertensive Diabetic Patients: Insights from a National Population-Based Study
Haewon Byeon *
Worker's Care & Digital Health Lab, Department of Future Technology, Korea University of Technology and Education, Cheonan 31253, South Korea
ABSTRACT
This study seeks to conduct an in-depth analysis of the key factors influencing the health-related quality of life (HRQoL) in patients suffering from both hypertension and diabetes, and to develop a predictive model based on these factors to propose personalized health management strategies for individual patients. In particular, the study seeks to address existing data imbalance issues and improve prediction accuracy using Cost-Sensitive Boosting techniques, thereby enabling a more precise evaluation of various health conditions. This research employed data from the 5th Korea National Health and Nutrition Examination Survey (KNHANES) conducted by the Korea Disease Control and Prevention Agency from 2010 to 2012. A total of 1,240 patients diagnosed with both hypertension and diabetes were selected as the study subjects. A predictive model for HRQoL was constructed utilizing Cost-Sensitive Boosting techniques, and its performance was compared with traditional machine learning methods such as SVM, Bagging, Boosting, and Random Forest. Additionally, variable importance analysis was conducted to ascertain the key factors affecting HRQoL, and multiple quantile regression analysis was employed to deeply interpret the individual and interactive impacts of these components. The Cost-Sensitive Boosting model exhibited enhanced performance across all evaluation metrics compared to other models, achieving an accuracy of 0.85 and an F1 score of 0.86, indicating high reliability in HRQoL prediction. Variable importance analysis revealed that weight gain, depressive mood, activity limitations, perceived stress, and subjective health status were the most significant factors impacting HRQoL. Quantification of these factors' effects confirmed the need for intervention strategies focusing on weight management, mental health support, enhancing activity, stress reduction, and patient education. This study identifies the key factors affecting the HRQoL in patients with both hypertension and diabetes, providing crucial foundational data for future personalized health management and policy formulation. The high predictive performance of the Cost-Sensitive Boosting model suggests it can be a powerful tool for predicting and managing complex health conditions. These findings require further validation through longitudinal studies across diverse populations, which can contribute to the development of more sophisticated and effective health management programs.
KEYWORDS
Health-Related Quality of Life, Cost-Sensitive Boosting, Hypertension, Diabetes, Predictive Modeling
Cite This Paper in IEEE or APA Citation Styles
(a). IEEE Format:
[1] Haewon Byeon , "Cost-Sensitive Boosting Approach to Analyzing Risk Factors Affecting Health-Related Quality of Life in Hypertensive Diabetic Patients: Insights from a National Population-Based Study," Universal Journal of Public Health, Vol. 13, No. 2, pp. 417 - 426, 2025. DOI: 10.13189/ujph.2025.130213.
(b). APA Format:
Haewon Byeon (2025). Cost-Sensitive Boosting Approach to Analyzing Risk Factors Affecting Health-Related Quality of Life in Hypertensive Diabetic Patients: Insights from a National Population-Based Study. Universal Journal of Public Health, 13(2), 417 - 426. DOI: 10.13189/ujph.2025.130213.