|| Checking for direct PDF access through Ovid
The organisation of work is undergoing rapid change making our traditional definitions and methods of exposure assessment less applicable or effective in understanding work-related risks. In a keynote address at EPICOH in 1995 I discussed concepts of measurement error and argued for use of statistical methods which explicitly linked exposure concepts with the outcomes. These concepts are now widely understood and we have moved beyond them using mixed models and a variety of more advanced statistical methods which were new at the time. However, our framework for that work is less and less relevant in current and future occupational settings. Non-standard employment arrangements and increasing disparities in income, often associated with other demographic characteristics, compel an increasing focus on the health of working populations, rather than individual diseases or conditions. However, the ways in which the work context is defined and implemented in research is highly variable. While many of the terms used have overlapping attributes, and each have been associated with changes in work exposures and health risk, their lack of clear conceptual definitions hampers our ability to explain the apparent health risks with which they are associated. This paper describes the various concepts and labels associated with precarious work and suggests that precarity requires integration of concepts of work organisation with worker vulnerability. By doing so, we can better understand the relationships between job content, working conditions and power dynamics within the workplace and its social context. Thus, we identify work as a social determinant of health, and can better assess the health implications of precarious work. These concepts are implemented using a multidimensional approach to job quality which incorporates both work organisation and workplace power dynamics, based on latent class cluster analysis, to define an integrated typology for defining contemporary employment conditions, as developed by Van Aerden, et al, in a European dataset. We adopt this approach within the US based General Social Survey, and compare the clusters identified in the EU with those observed in the US. This typological approach overcomes the limitations of characterising work organisation or social determinants on single dimensional characteristics, such as contract-type or demographic factors, and offers a new framework for understanding the implications of precarity. While the ‘exposure,’ health outcomes, and methods differ substantially than those relevant in the 1990s, we again argue that a clear conceptual definition, measurement methods, and linkage with outcomes of importance in the 21 st century, are required to continue understanding the impact of working conditions on health.