OP94 Changing labour market, changing inequalities? Labour market participation and inequalities in self-rated health in Britain, a difference in-difference analysis using the General Household Survey. (2nd September 2014)

• Record Type: Journal Article
• Authors: Popham, F; Curnock, E
• Is Part Of: Journal of epidemiology and community health. Volume 68(2014)Supplement 1
• Page Start: A46
• Page End: A46
• Publication Date: 2014-09-02
• Subjects: health inequalities -- non-employment; Public health -- Periodicals; Epidemiology -- Periodicals; 614.4
• DOI: 10.1136/jech-2014-204726.96
• Languages: English
• ISSNs: 0143-005X

Abstract

Abstract : Background: Many have advocated for reducing health inequalities by tackling the social determinants of health, such as labour force participation. However, there is a lack of evidence of how change in such social determinants impacts health inequalities at the population level. In the last 40 years the UK has seen an end to male full employment, increasing female labour market participation, rising levels of non-employment due to sickness / disability and cyclical unemployment. The extent of these changes varies across social classes. We aimed to assess the impact long term of these changes on health inequality trends in Britain. Methods: We used data from the General Household Survey, a nationally representative annual cross-sectional study. We compared two periods with similar levels of unemployment but different levels of employment and labour market non-participation, 1977–1979 and 2002–2004. Using self-rated health categorised as 'poor', we studied differences in social class inequalities for prime working age (30 to 54) men and women between the two periods using a difference-in-difference design implemented through a regression analysis using Stata. Sensitivity analysis is planned to assess time-varying confounding. Results: Preliminary results suggest that the trend in poor self-rated health was less favourable for lower social class groups and that these differential trends were largely attenuated by controlling for social class differences in labour force changes. For example, compared to the professional and managerial class the trend for skilled manual social class men was 4 percentage points worse (95% CI 2 to 5) but reduced to little difference (0;-2 to 2) after accounting for labour force changes. Class differences in trends in self-rated health were stronger for men than women. Labour force status seems to attenuate differences because lower social class men saw a greater fall in the employment rate and a greater rise in non-work due to sickness and disability. For women, gains in employment were less for the lowest social class group, who also saw a rise in non-work due to sickness and disability. Conclusion: Although cautious about causality, trends in social class inequalities in poor self-rated health seem related to changes in labour force participation, suggesting that policies affecting labour force participation may be important for tackling health inequalities.