Abstract
The study explored cross-sectional associations between loneliness and occupational dysfunction in community-dwelling older adults. Seventy-four older adults (12 men and 62 women; mean age 73.9 ± 8.3 years) completed a questionnaire survey that included the Japanese version of the UCLA Loneliness Scale Version 3 and the Classification and Assessment of Occupational Dysfunction (CAOD). Bayesian statistical modeling was used for a more stable estimation given the small sample. For model selection, we assumed a univariate analysis model of the CAOD (Model 1); a multivariate analysis model, including confounding factors in Model 1 (Model 2); and a multivariate analysis model, including random effects in Model 2 (Model 3). The best model was selected by comparing the widely applicable information criterion (WAIC) and the widely applicable Bayesian information criterion (WBIC) for each model. Bayesian statistics with the dependent variable as "loneliness" showed that the best model used "occupational dysfunction" as the independent variable and included confounding factors and random effects (WAIC = 474.5 and WBIC = 213.1). The best model identified an association between occupational dysfunction and loneliness (odds ratio [OR] = 2.363; 95% Bayesian confidence interval [CI] = 1.105-5.259). This study highlights the role of occupational dysfunction in addition to the risks and related factors reported to date when dealing with loneliness. Therapists could help older adults cope with loneliness by supporting their social participation and improving their occupational dysfunction.