Aim Generally, good work is good for health but there are few objective data about the effect of permanent exit from work (either through normal retirement, or health-related job exit) on health at older ages. We aimed to explore if exit from the workforce is followed by a change in self-rated health, using longitudinal data from the Health and Employment After Fifty (HEAF) Study.

Methods A cohort of participants aged 50–64 years were recruited from 24 English general practices in 2013–14 and have been followed-up annually by postal questionnaire. At baseline, information was collected about demographic and employment circumstances, physical workload, psychosocial aspects of work and their general health. At each subsequent follow-up, participants self-rated their health and additionally reported whether they were still in paid employment or whether they had exited the workforce, and if so, whether the reason for exit was at least partly due to their health. We used logistic regression modelling to explore the effect of exit from the workforce on changes in self-rated health after adjustment for self-rated health at baseline, before and after controlling for demographic, employment, and socio-demographic factors.

Results HEAF recruited a total of 8,134 people aged 50–64 years at baseline, amongst whom 5,059 were in paid employment and were successfully followed-up. Of these, 3,617 were still working 5 years later, 947 exited work permanently not for health reasons, and 333 exited work permanently at least partly due to their health. Self-rated health remained the same for: 53% of those still in paid employment; 55% of those who exited the workforce not on health grounds; and 47% of those who exited due to their health. Self-rated health improved from baseline to 5-year of follow-up for: 21% of those still in paid employment; 25% of those who exited the workforce not on health grounds; and 18% among those who exited due to their health. Regression analysis showed that normal exit from the workforce was associated with improving health subsequently (OR: 1.32, 95%CI: 1.07,1.61), while health-related exit was associated with poorer health subsequently (OR: 2.88, 95% CI: 2.16,3.85). These effects were stronger among males than females, and were robust to adjustments for demographic, employment, and socio-demographic factors.

Conclusion This study highlights the need for more in-depth exploration of the dynamic impact of work exit on health amongst older people, aiming to develop effective policy measures for a healthy transition from work to retirement.