Socioeconomic status and suicidal behaviour in the UK Biobank: prospective cohort study on effect modifiers

Socioeconomic status (SES) is a risk factor of suicidal behaviour. Little is however known if such association varies across strata of loneliness, social isolation, and area-level deprivation. We examined the association between SES and suicidal behaviour, and its modification by loneliness, social isolation and neighbourhood deprivation. We used a large prospective cohort from the UK Biobank comprising participants aged 38-73 years and followed-up between December 19, 2006 and March 31, 2021. Suicidal behaviour was defined as suicide death or incident non-fatal self-harm obtained through linkage to death and hospital records. Composite individual SES was derived from latent class analysis using household income, employment status and education. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox regression models. The analytic sample comprised 316,676 participants with 1,093 suicidal events over 3,815,232 person-years of follow-up. Compared with participants in the least-exposed strata of high SES-least lonely, high SES-least isolated and high SES- residing in least deprived neighbourhoods, those in the double-exposed strata of low SES-highly lonely, low SES-highly isolated, and low SES-highly deprived neighbourhoods had 3.16 (95% CI 2.35 to 4.24), 2.72 (2.08 to 3.57) and 2.92 (2.11 to 4.04) times higher risks of suicidal behaviour, respectively. There was evidence of synergistic interactions between low SES and loneliness, and low SES and social isolation upon higher risk of suicidal behaviour. Tailor-made preventive interventions promoting social engagement and provisioning resources and services in deprived neighbourhoods may reduce burdens of suicidal behaviours among at-risk subgroups.