» Articles » PMID: 39526215

Factors Associated with Acquiring Exercise Habits Through Health Guidance for Metabolic Syndrome Among Middle-aged Japanese Workers: A Machine Learning Approach

Overview
Journal Prev Med Rep
Date 2024 Nov 11
PMID 39526215
Authors
Affiliations
Soon will be listed here.
Abstract

Objective: Physical inactivity increases the risk of mortality and chronic morbidity. Therefore, it is crucial to establish strategies to encourage individuals to increase their physical activity and develop exercise habits. The objective of this study was to explore factors associated with acquiring exercise habits using machine learning algorithms.

Methods: The analyzed dataset was obtained from the Specific Health Guidance for metabolic syndrome systematically implemented by the Japanese Ministry of Health, Labor, and Welfare. We selected target individuals for health guidance without exercise habits in 2017 and assessed whether the participants acquired exercise habits through health guidance in 2018. We applied ten machine learning algorithms to build prediction models for acquiring exercise habits.

Results: This study included 16,471 middle-aged Japanese workers (age, 49.5 ± 6.2 years). Among the machine learning algorithms, the Boosted Generalized Linear Model was the best for predicting the acquisition of exercise habits based on the receiver operating characteristic curve on the test set (ROC-AUC, 0.68). According to the analyses, the following factors were associated with the acquisition of exercise habits: being in the maintenance or action stage of changing exercise and eating behaviors based on the transtheoretical model; regular physical activity or walking; normal high-density lipoprotein cholesterol; and high alcohol consumption.

Conclusions: Our findings can be used to establish an efficient strategy for encouraging individuals to acquire exercise habits through Specific Health Guidance or other health guidance. However, the lower ROC-AUC suggests that additional variables are necessary to enhance the prediction model.

References
1.
Sumimoto Y, Yanagita M, Miyamatsu N, Okuda N, Nishi N, Nakamura Y . Association between socioeconomic status and physical inactivity in a general Japanese population: NIPPON DATA2010. PLoS One. 2021; 16(7):e0254706. PMC: 8282078. DOI: 10.1371/journal.pone.0254706. View

2.
Tsushita K, Hosler A, Miura K, Ito Y, Fukuda T, Kitamura A . Rationale and Descriptive Analysis of Specific Health Guidance: the Nationwide Lifestyle Intervention Program Targeting Metabolic Syndrome in Japan. J Atheroscler Thromb. 2017; 25(4):308-322. PMC: 5906184. DOI: 10.5551/jat.42010. View

3.
Koba S, Tanaka H, Maruyama C, Tada N, Birou S, Teramoto T . Physical activity in the Japan population: association with blood lipid levels and effects in reducing cardiovascular and all-cause mortality. J Atheroscler Thromb. 2011; 18(10):833-45. DOI: 10.5551/jat.8094. View

4.
Yin J, Li Y, Xue J, Zong G, Fang Z, Zou L . Explainable Machine Learning-Based Prediction Model for Diabetic Nephropathy. J Diabetes Res. 2024; 2024:8857453. PMC: 10821806. DOI: 10.1155/2024/8857453. View

5.
Liangpunsakul S, Crabb D, Qi R . Relationship among alcohol intake, body fat, and physical activity: a population-based study. Ann Epidemiol. 2010; 20(9):670-5. PMC: 2921229. DOI: 10.1016/j.annepidem.2010.05.014. View