Countries like France have long reported large disparities in death by sex and by occupational class (OC). Moreover, studies have highlighted the double burden of manual and low-skill workers who not only live shorter lives but also spend more years in disability than high-skill professionals. However, several knowledge gaps persist, in particular regarding how these patterns and trends unfold over time, in recent periods and younger generations; the causes of death that could explain such lifespan disparities; and the contribution of different conditions and causes of death to loss in healthy life. Furthermore, moving from description to understating the underlying mechanisms that drive these inequalities requires that we integrate perspectives from occupational health, lifecourse research and social determinants of health. Indeed, not only can work cause illness and disability (through social causation), but illness and disability, be it or not work-related, can also alter individuals’ labour market success (through health selection). Empirical evidence in occupational health and social sciences support both of these notions, yet they are rarely considered in a unified framework, adopting a gendered and intersectional lens. Building on a theoretical model that accounts for the interdependent pathways between work and health throughout the lifecourse, our project will pursue three overarching goals. Firstly, we will identify gender-OC-patterns and trends of inequalities in mortality and ill-health. Secondly, we will explain the underlying mechanisms that link health and work dynamically through empirical case studies conducted on disabling work accidents and mental health. And lastly, we will explore the promises and pitfalls of medico-administrative big data to advance knowledge in these fields. We will indeed use a novel and unique source of record-linkage longitudinal data, the EDP-Santé - EDP for Permanent Demographic Sample, linked with the National Health Data System, SNDS. EDP-Santé combines demographic, socio-economic, and work characteristics from administrative data with health care consumption and causes of death on a large representative sample of more than 3.5 million people living in France, followed up over at least 13 years. EDP-Santé has just started to be accessible to researchers, that will make us among the first teams to use it in population sciences. On top of making empirical contributions to disentangle processes of selection (health-related) and social causation (work-related) in the observed inequalities, the project will make use of up-to-date methodological tools for monitoring socio-occupational gaps in life and health expectancy. This is particularly important in the context of a changing labour market, lifelong career diversification and an ageing population. A second innovation is to link work and health in a bidirectional way, thanks to a dynamic approach based on massive, longitudinal and multidimensional data. Another innovation will be to adopt an intersectional perspective by considering the social relations that can attenuate or exacerbate these processes, in particular gender and racialization. A further originality will then be to critically review these newly linked data. Although often considered a limitation in quantitative research, we contend that the identification of blind spots concerning specific risks and particular population segments is key to improving epistemic justice. This is particularly relevant in acknowledging occupational health as a component part of public health and to inform public health and social protection policies accordingly.