Figure 1.

Natural direct effect (NDE) and natural indirect effect (NIE) of high social activities (HSA) on psychological distress as a percentage of the total causal effect (TCE). (Corresponding odds ratios [OR] and 95% confidence intervals are labelled.)

Analyses were restricted to only those participants who (1) completed both the baseline and follow-up survey and (2) were classified as having ‘none’ or ‘mild’ psychological distress (K10 score <24) at baseline (N = 57,961). Mediation analysis using a counterfactual approach (Lange et al., 2012; VanderWeele, 2015) was used to assess the TCE of level of social activity (high/low) on psychological distress (K10 score ⩾24) (adjusting for sex, age group, educational achievement, migrant status, marital status, employment or retirement status, and household income). The TCE decomposed the association into (1) the NDE of physical social activity on psychological distress and (2) the NIE through non-physical social activities (talking with friends and family on the telephone) using the imputation-based approach.