Bidirectional Relationships Between Loneliness, Social Isolation, and Physical Inactivity in the Household, Income and Labour Dynamics in Australia Cohort Study

Background Cross-sectional studies show associations between loneliness, social isolation and physical inactivity. Cohort studies are shedding light on these relationships and further longitudinal investigations are needed. Purpose This study aimed to assess the longitudinal and bidirectional associations between loneliness, social isolation, and physical inactivity. Methods Data were drawn from five annual waves of the Household and Labour Dynamics of Australia Survey (2015–2019), providing a sample of 17,303 persons (mean age = 46.3 years [SD = 18.0], 49.4% female). Relationships between loneliness, social isolation, and physical inactivity were examined using cross-lagged panel modeling, with estimation of simultaneous cross-lagged effects across each wave. Models adjusted for sociodemographic factors, chronic disease status, psychological distress, and mutually for social isolation or loneliness. Moderation of associations by sex was explored. Results There were modest lagged effects of physical inactivity on loneliness across the survey waves (odds ratio 1.16 [95% confidence interval 1.04–1.29] to 1.20 [1.07, 1.33]). A lagged effect of physical inactivity upon social isolation was only present across three of the waves (odds ratio 1.20 [1.02–1.41] to 1.23 [1.05–1.42]). While loneliness and social isolation showed lagged effects upon physical inactivity, these did not persist with adjustment for psychological distress. Conclusions Longitudinal analysis found that physical inactivity consistently predicted loneliness, but not social isolation. After adjustment for confounding, loneliness and social isolation were not predictive of physical inactivity. While the strength of the associations was modest, further investigation is warranted of the type and dose of physical activity that is most beneficial for reducing loneliness.