Table_1.xlsx

Abstract

Loneliness is a major public health issue, with its prevalence rising during COVID-19 pandemic lockdowns and mandated ‘social distancing’ practices. A 2020 global study (n = 46,054) found that, in comparison to women, men experienced the greatest levels of loneliness. Although research on predictors of loneliness during COVID-19 is increasing, little is known about the characteristics of men who may be particularly vulnerable. Studies using prospective data are needed to inform preventative measures to support men at risk of loneliness. The current study draws on rare longitudinal data from an Australian cohort of men in young to mid-adulthood (n=283; aged M=34.6, SD=1.38 years) to examine 25 pre-pandemic psychosocial predictors of loneliness during COVID-19 social restrictions (March–September 2020). Adjusted linear regressions identified 22 pre-pandemic predictors of loneliness across a range of trait-based, relational, career/home and mental health variables. Given the extensive set of predictors, we then conducted penalized regression models (LASSO), a machine learning approach, allowing us to identify the best fitting multivariable set of predictors of loneliness during the pandemic. In these models men’s sense of pre-pandemic environmental mastery emerged as the strongest predictor of loneliness. Depression, neuroticism and social support also remained key predictors of pandemic loneliness (R2 = 26, including covariates). Our findings suggest that men’s loneliness can be detected prospectively and under varying levels of social restriction, presenting possible targets for prevention efforts for those most vulnerable.