Predicting Risk Factors for Dyslipidemia Based on Health Behaviors by Age in Adults Using Machine Learning
Source abstract
Abstract
According to the 2022 Korean Society of Lipidology and Atherosclerosis, in Korea, dyslipidemia is a common disease that occurs in 40.2% of adults aged 20 or older, and its prevalence increases with age. Although dyslipidemia has a high prevalence of 47.8% in adults aged 30 or older, it is known to be preventable and manageable through lifestyle improvements in areas including eating habits, alcohol consumption, smoking, and physical activity. In this study, we propose a model for predicting age-specific dyslipidemia risk factors according to adult health behavior characteristics and diet. By analyzing the correlation between age-specific health behaviors and diet and the presence or absence of dyslipidemia, we aimed to predict dyslipidemia risk factors through a combination of multiple factor variables. This study utilized data from the 8th National Health and Nutrition Examination Survey, and selected 12,028 adults who received a doctor’s diagnosis of dyslipidemia as the subjects. In order to compare the characteristics of the dyslipidemia diagnosis group and the non-diagnosed group, a Rao–Scott χ2 test was performed, and machine learning-based logistic regression and decision tree analyses were performed to predict the dyslipidemia risk factors. Analyzing the difference in the dyslipidemia prevalence according to the general characteristics and health status showed no significant difference between the men and women in the 19–34, 35–49, and 50–64 age groups, but there was a significant difference in the dyslipidemia prevalence in the 65 and older group. It was found that the dyslipidemia risk also increased with age. In terms of health behavior characteristics, the alcohol intake frequency and aerobic exercise frequency were found to have statistically significant effects and, in terms of eating habits, the breakfast frequency and dining out frequency were found to be significant factor variables in the dyslipidemia prevalence. As a result of the decision tree analysis, the most important dyslipidemia predictive factor showed differences according to the age group. The most important predictive variable for the presence or absence of dyslipidemia in the 19–34 age group was the BMI; for the 35–49 age group, it was gender and subjective health perception; for the 50–64 age group, it was subjective health perception and the BMI; and for the 65 and older group, it was the BMI. This suggests that healthy eating habits and behaviors such as aerobic exercise are very important for preventing and managing dyslipidemia as age increases.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Cardiovascular Health and Risk Factors · Artificial Intelligence in Healthcare
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Division of Software Liberal Arts, University of Mokwon, Daejeon 35349, Republic of Korea
Well-Dying Convergence Laboratory, University of Konyang, Daejeon 35365, Republic of Korea
Funding
Ministry of Education of the Republic of the Korea
References (19)
Show source-provided references
1. (2024, October 01). World Health Organization. Available online: https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds).
2. Korea Disease Control and Prevention Agency (2024, October 01). National Health Information Portal, Dislipidemia. Available online: https://health.kdca.go.kr/healthinfo/biz/health/gnrlzHealthInfo/gnrlzHealthInfo/gnrlzHealthInfoView.do?cntnts_sn=6054.
3. Noncommunicable diseases: Current status of major modifiable risk factors in Korea10.3961/jpmph.2013.46.4.165 ↗
4. Epidemiology of dyslipidemia in Korea10.5124/jkma.2016.59.5.352 ↗
5. Korean Society of Lipid Atherosclerosis (2018). The Committee of Clinical Practice Guideline of Korean Guidelines for the Management of Dyslipidemia, The Korean Society of Lipid and Atherosclerosis. [4th ed.].
6. Risk Factors affecting Dyslipidemia according to Age Group in Women: Data from the Korea National Health and Nutrition Examination Survey, 2014~201710.7739/jkafn.2020.27.3.310 ↗
7. Effects of age, sex, and menopausal status on lood cholesterol profile in the Korean population10.4070/kcj.2015.45.2.141 ↗
8. Prevalence and risk factors for dyslipidemia among adults in rural and urban China: Findings from the China National Stroke Screening and prevention project (CNSSPP)10.1186/s12889-019-7827-5 ↗
9. Relationship between age-specific practicing of a healthy diet and dyslipidemia in Korean Adults: A study of Korean adults without diagnosed cardiovascular disease
10. Convergence Study on the Comparison of Risk Factors for Dyslipidemia by Age and Gender: Based on the Korea National Health and Nutrition Examination Survey (2013~2015 year)
11. Correlation between Physical Activity Status and Dyslipidemia in Korean Adults: The 2010 Korea National Health and Nutrition Examination Survey10.15656/kjcg.2014.15.1.35 ↗
12. The Study on the Difference of Blood Level of HDL-Cholesterol by Obesity and Health Behavior from the Seventh (2016) Korea National Health and Nutrition Examination Survey10.3746/jkfn.2020.49.12.1377 ↗
13. Korea Centers of Disease Control and Prevention (2024, April 20). Sampling Design of the 6th KNHNES and Weighting Adjustment of the 8th KNHNES. Available online: https://knhanes.kdca.go.kr/knhanes/main.do.
14. The regression analysis of binary sequences10.1111/j.2517-6161.1958.tb00292.x ↗
15. Breiman, L., Friedman, J.H., Olshen, R.A., and Stone, C.J. (1984). Classification and Regression Trees, Wadsworth International Group.
16. A comparison of logistic regression to decision-tree induction in a medical domain10.1006/cbmr.1993.1005 ↗
17. comparison of logistic regression to decision tree induction in the diagnosis of carpal tunnel syndrome10.1006/cbmr.1999.1521 ↗
18. (2024, December 15). QuantDare. Available online: https://quantdare.com/decision-trees-gini-vs-entropy.
19. (2025, February 20). Scikit-Learn Machine Learning in Python. Available online: https://scikit-learn.org/stable/modules/tree.html.