Worklessness and regional differences in the social gradient in general health: Evidence from the 2001 English census

There has been much focus on separating contextual and compositional influences on social inequalities in health. However, there has been less focus on the important role of place in shaping the distribution of risk factors. Spatial variations in worklessness are one such factor. In this paper, then we examine the extent to which between and within regional differences in the social gradient in self-rated general health are associated with differences in rates of worklessness.

Data were obtained for men and women of working age (25-59) who had ever worked from the Sample of Anonymised Records (Individual SAR)-a 3% representative sample of the 2001 English Census (349,699 women and 349,181 men). Generalised linear models were used to calculate region and age adjusted prevalence difference for not good health by education (as an indicator of socio-economic status) and employment status. The slope index of an inequality was also calculated for each region.

For both men and women, educational inequalities in worklessness and not good health are largest in those regions with the highest overall levels of worklessness. Adjusting for worklessness considerably attenuated the educational health gradient within all English regions (by over 60%) and virtually eliminated between region differences.

Macroeconomic policies, which influence the demand for labour, may have an important role in creating inequalities in general health of the working age population both within and between regions. Employment policy may therefore be one important approach to tackling spatial and socio-economic health inequalities.