Author	A.M.M. Hughes
Author	A. McMunn
Author	M. Bartley
Author	M. Kumari
Abstract	Background An extensive literature links unemployment to ill-health, but almost no research has investigated links in terms of systemic inflammation. A cardiovascular risk factor influenced by psychosocial stress, systemic inflammation may also be involved in the aetiology of depression, therefore providing a plausible pathway from the social stressor of unemployment to both psychological and physical illness. This thesis investigates associations between unemployment, systemic inflammation, and depressive symptoms in a contemporary UK context. Methods Cross-sectional associations of unemployment and inflammatory markers were investigated by pooling data from the Health Survey for England, Scottish Health Survey, National Child Development Survey (NCDS), and UK Household Longitudinal Survey (UKHLS) in an individual-participant meta-analysis. For longitudinal analyses, employment histories were constructed for NCDS and UKHLS participants spanning 34 and 10 years respectively. Total unemployment in months, number of spells, age at first unemployment, and recentness of last unemployment were calculated. Associations were investigated between these summary measures, inflammatory markers, and depressive symptoms by regression using multiply imputed data. Mediation was explored by socioeconomic position, health factors, health behaviours, and current unemployment. Results Current unemployment was robustly associated with inflammatory markers, but associations varied considerably by country (stronger outside England) and study population (no effects in UKHLS). Longitudinally, unemployment did not robustly predict inflammatory markers, and inflammatory markers did not robustly predict later depressive symptoms. Aggregated unemployment did predict depressive symptoms, explained more by socioeconomic position and long-term illness than other factors. Conclusions Results suggest associations of unemployment and inflammation may be under certain conditions substantial, but are largely transitory. Country variation remains unexplained, but stronger associations in higher-unemployment areas go against a model on which the poorer health of jobseekers is primarily explained by non-causal selection effects. Results suggest the influence of inflammation in depressive aetiology is modest compared to other factors, but that unemployment may have lasting effects on psychological health.