Health and Engagement in Employment among Older Individuals: A Comparative Analysis Based on Transitions in Health, Employment, Social Engagement and Inter- generational Transfers in Singapore Study (THE SIGNS Study)

Abstract

Background: Given the growing relevance of post-retirement employment, the lack of significant within-individual health changes following re-employment underscores the importance of controlling for unobserved heterogeneity and cautions against over- interpreting simple associations. Research Aims: This study investigates the bidirectional relationship between health and re-employment among older adults, examining how baseline health predicts re-employment and how re-employment, in turn, affects subsequent health outcomes. Methods: Data were drawn from the Transitions in Health, Employment, Social Engagement and Inter-generational Transfers in Singapore (THE SIGNS) Study, a nationally representative longitudinal survey of adults aged over 60 (N = 2,266). Logistic regression assessed baseline health predictors of re-employment; OLS, ordinal logistic, and first-difference models examined 2019 health outcomes. Results: Baseline self-rated health (OR for excellent vs. poor = 0.689, 95% CI: –0.609 to 1.987), cognitive function (OR= 0.023, 95% CI: –0.227 to 0.273), and depressive symptoms (OR = 0.007, 95% CI: –0.007 to 0.091) were not significantly associated with re-employment in 2019. Re-employment, in turn, was linked to lower odds of better self- rated health when modeled as an ordinal outcome (OR = 0.357, 95% CI: –0.016 to 0.730), but showed a small positive association when health was treated as a continuous variable (β = 0.157, 95% CI: 0.002 to 0.312), indicating model-dependent sensitivity. A marginally positive association was observed for cognitive function (β = 0.107, 95% CI: –0.017 to 0.231), while no significant was found for depressive symptoms (β = –0.034, 95% CI: –0.334 to 0.265). These associations were not replicated in first-difference models, though effect directions remained consistent. Conclusions: Re-employment showed limited health selection effects and modest, model-sensitive associations with later-life health, underscoring the complexity of work-health dynamics in aging populations.Keywords: Aging, Re-employment, Self-rated health, Cognition, Depression, Singapore