Sick leave absence and the relationship between intra-generational social mobility and mortality: health selection in Sweden.

BACKGROUND: Poor health could influence how individuals are sorted into occupational classes. Health selection has therefore been considered a potential modifier to the mortality class gradient through differences in social mobility. Direct health selection in particular may operate in the short-term as poor health may lead to reduced work hours or achievement, downward social mobility, unemployment or restricted upward mobility, and death. In this study, the relationship between social mobility and mortality (all-cause, cancer-related, cardiovascular disease-related (CVD), and suicide) is explored when the relationship is adjusted for poor health. METHODS: Using Swedish register data (1996-2012) and discrete time event-history analysis, odds ratios and average marginal effects (AME) of social mobility and unemployment on mortality are observed before and after accounting for sickness absence in the previous year. RESULTS: After adjusting for sickness absence, all-cause mortality remained lower for men after upward mobility in comparison to not being mobile (OR 0.82, AME -0.0003, CI - 0.0003 to - 0.0002). Similarly, upward mobility continued to be associated with lower cancer-related mortality for men (OR 0.85, AME -0.00008, CI - 0.00002 to - 0.0002), CVD-related mortality for men (OR 0.76, AME -0.0001, CI - 0.00006 to - 0.0002) and suicide for women (OR 0.67, AME -0.00002, CI - 0.000002 to - 0.00003). The relationship between unemployment and mortality also persisted across most causes of death for both men and women after controlling for previous sickness absence. In contrast, adjusting for sickness absence renders the relationship between downward mobility and cancer-related mortality not statistically different from the non-mobile. CONCLUSIONS: Health selection plays a role in how downward mobility is linked to cancer related deaths. It additionally accounts for a portion of why upward mobility is associated with lower mortality. That health selection plays a role in how social mobility and mortality are related may be unexpected in a context with strong job protection. Job protection does not, however, equalize opportunities for upward mobility, which may be limited for those who have been ill. Because intra-generational upward mobility and mortality remained related after adjusting for sickness absence, other important mechanisms such as indirect selection or social causation should be explored.