In this paper we used wave 1 (2002) to wave 6 (2012) of the English Longitudinal Study of Ageing (ELSA) to assess whether trajectories in the prevalence of self-reported limiting long term illness differed before and after retirement and whether any such differences in slope were found across socio-economic characteristics and the conditions of work in the final years of employment. The longitudinal analysis used a sub-sample of ELSA comprising those who retired between wave 2 (2004) and wave 6 (2012). We fitted a repeated measure logistic regression to model the trajectory in the log odds of illness before and after retirement. We found evidence of a slower increase in self-reported illness after than before retirement that was most strongly observed for those in the least favourable circumstances prior to retirement (lower social class, depressed prior to retirement and single). A similar retirement effect was observed for those in physically demanding occupations. Whilst we did not detect differences in post and pre-retirement slopes according to other aspects of working conditions we found higher rates of self-reported illness for individuals involved in unsatisfying jobs or in work where demands exceed rewards. These differentials were robust to inclusion of information on socio-economic circumstances and appeared to persist beyond retirement. Our interpretation of the results is that disadvantage across the life course makes those in the least favourable circumstances or involved in physically demanding work less able to draw on social, economic and health reserves to cope with the demands of work in the final years of employment. As a result these groups of disadvantaged individuals experienced a slower increase in rates of illness after than before retirement when the demands of work are removed. Proposals to increase retirement age without also tackling inequalities in circumstances in the final years of employment are likely to increase inequalities in self-reported health.
- self-reported illness