Relative Effects of Demographic, Psychological, Behavioral, and Social Factors on the Initiation and Maintenance of Leisure-time Physical Activity: Results From a Confirmatory Path Analysis in a Longitudinal Study

Background: There is a lack of evidence of the complicated pathways of underlying determinants in the phases of physical activity. The purpose of this study was to evaluate simultaneously a set of potential determinants on the initiation and maintenance phases of leisure-time physical activity (LTPA). Methods: The longitudinal data of 54,359 Korean adults aged 40–69 years from the Health Examinees study were used. The median follow-up duration was 4.2 years. The self-reported durations per week of LTPA was repeatedly assessed. Based on previous longitudinal studies, the potential determinants were selected, and hypothetical models were constructed that consider the complex associations between the determinants. The standardized coefficients for direct and indirect effects were estimated using path analysis to differentiate contributions of mediation from the total effects. Results: In the total population, age, education, chronic diseases, smoking, depression symptoms, and self-rated health were significantly associated with both initiation and maintenance phases. Income (B = 0.025) and social supports (B = 0.019) were associated only with the initiation phase. Waist-to-hip ratio (B = −0.042) and stress (B = −0.035) were associated only with the maintenance phase. After stratifying by sex, the significant effects of education, chronic diseases, and smoking were found only in men. The initiation phase-specific effects of income and social supports and the maintenance phase-specific effects of stress were found only in women. It was estimated that indirect effects contributed approximately 15% of the total effect. Conclusion: The findings suggested that there were initiation- or maintenance-specific determinants of leisure-time physical activity according to sex.