P65 How are occupational histories associated with self-rated health in middle-aged adults? A cross-sectional analysis of retrospective UK Biobank data

Background Employment and occupation can greatly influence one’s life course, identity, resilience, health and health inequalities. At the same time, the working environment is rapidly changing, with flexible and agile working conditions and extended working lives, often resulting in ‘non-traditional’ career trajectories. We aimed to explore and identify common occupational history patterns and assess the associations between these patterns and self-rated health. Methods We analysed baseline UK Biobank data to construct participants occupational histories using employment start/end dates to identify patterns of employment status during working life, with each year being categorised into different employment or employment gap states. We used sequence analysis, followed by optimal matching and cluster analysis methods to classify respondents’ occupational history patterns. We produced Sequence Index Plots (SIPs) by gender and age groups. Theoretically derived occupational history patterns based on the SIPs were compared to the data-driven ones, to determine final patterns. Logistic regression models were run using occupational histories as explanatory variables for the health outcome of self-rated health (dichotomised: ‘excellent/good’ versus ‘fair/poor’). Results There was good agreement between the SIPs and cluster analysis; resulting in 3–5 different occupational history patterns per age/gender group, including continuous employment, employed then retired and employed with breaks for education and caring responsibilities as examples. Women aged 40–49 had better overall self-rated health if continuously employed compared with those with a pattern that included employed/in education/caring responsibilities (OR=0.80; CI:0.69, 0.93), or those off work at some point due to ill health (OR: 0.68; CI: 0.59, 0.79). Men aged 40–49 reported worse self-rated health if employed and retired (OR=0.64; CI:0.53, 0.78)) or were off work at some point due to ill health (OR=0.42; CI: 0.31, 0.57). Women aged 50–59 and 60–69 tended to have better self-rated health if retired, even if their employment histories involved multiple gaps due to caring responsibilities or short-term employment. This was not the case for men; being continuously employed or continuously employed and then retired was generally associated with better self-rated health compared to groups with non-continuous employment. Conclusion Continuous employment appeared to be associated with better self-rated health in men but not always for women. Our study is limited by the retrospective nature of the data and the limited representativeness of the study population. The modern working environment is rapidly changing, increasingly giving rise to ‘non-traditional’ career trajectories which might result in future adverse health impacts.