Does equality legislation reduce intergroup differences? Religious affiliation, socio-economic status and mortality in Scotland and Northern Ireland: A cohort study of 400,000 people
Religion frequently indicates membership of socio-ethnic groups with distinct health behaviours and mortality risk. Determining the extent to which interactions between groups contribute to variation in mortality is often challenging. We compared socio-economic status (SES) and mortality rates of Protestants and Catholics in Scotland and Northern Ireland, regions in which interactions between groups are profoundly different. Crucially, strong equality legislation has been in place for much longer and Catholics form a larger minority in Northern Ireland. Drawing linked Census returns and mortality records of 404,703 people from the Scottish and Northern Ireland Longitudinal Studies, we used Poisson regression to compare religious groups, estimating mortality rates and incidence rate ratios. We fitted age-adjusted and fully adjusted (for education, housing tenure, car access and social class) models. Catholics had lower SES than Protestants in both countries; the differential was larger in Scotland for education, housing tenure and car access but not social class. In Scotland, Catholics had increased age-adjusted mortality risk relative to Protestants but variation among groups was attenuated following adjustment for SES. Those reporting no religious affiliation were at similar mortality risk to Protestants. In Northern Ireland, there was no mortality differential between Catholics and Protestants either before or after adjustment. Men reporting no religious affiliation were at increased mortality risk but this differential was not evident among women. In Scotland, Catholics remained at greater socio-economic disadvantage relative to Protestants than in Northern Ireland and were also at a mortality disadvantage. This may be due to a lack of explicit equality legislation that has decreased inequality by religion in Northern Ireland during recent decades.