Research aim This study aimed to quantify how much of the adult social gradient in sick leave can be attributed to the mediating role of physical workload, while accounting for childhood and adolescent social position and neuroticism.

Methods Our sample consisted of 2099 women and 1229 men from a Norwegian birth cohort study (born 1967–1976) who participated in the Nord-Trøndelag Health Study (2006–2008) (HUNT3). Data on sick leave (defined as >16 calendar days; 2006–2009) and social position during childhood, adolescence, and adulthood were obtained from national registers. Study outcome was time-to-first sick leave spell. Physical workload and neuroticism were self-reported in HUNT3. Mediating effects through physical workload were estimated using a method based on the additive hazards survival model.

Findings A hypothetical change from highest to lowest group in adult social position was, for women, associated with 51.6 [95% confidence interval (95% CI) 24.7–78.5] additional spells per 100 000 person-days at risk, in a model adjusted for childhood and adolescent social position and neuroticism. The corresponding rate increase for men was 41.1 (95% CI: 21.4–60.8). Of these additional spells, the proportion mediated through physical workload was 24% (95% CI: 10–49) and 30% (95% CI: 10–63) for women and men, respectively. The effect of adult social position on sick leave was partly mediated through physical workload, even while accounting for earlier lifecourse factors. Our findings suggest that interventions aimed at reducing physical workload among those with lower adult social position could reduce the social gradient in sick leave.