Among 13,879 U.S. Health and Retirement Study participants, those who maintained sustained part-time employment between ages 51 and 75 had 38% lower odds of dementia after age 76 (OR: 0.62, 95% CI: 0.48–0.80) compared with people who retired from full-time work in their early-to-mid 60s. Notably, retirement timing alone — whether early or late — showed little association with dementia risk. Prolonged disability-related exits from the workforce were associated with 53% higher dementia odds (OR: 1.53, 95% CI: 1.22–1.91), suggesting health-driven labor force departure carries its own cognitive burden.
This finding reframes an ongoing debate about whether retirement itself harms cognition. The dominant hypothesis has been that cognitive disengagement post-retirement accelerates decline — but this study suggests the relevant variable is employment intensity, not the timing of full exit. Sustained part-time work likely preserves cognitive stimulation, social engagement, and structured routine — three factors independently associated with dementia resilience in prior longitudinal research. The disability trajectory finding almost certainly reflects reverse causation: pre-dementia cognitive decline may force early work cessation rather than disability causing dementia.
Limitations are significant. This is an observational cohort — causal inference is impossible, and residual confounding by education, wealth, and occupational complexity is plausible. The dementia measure is algorithmically derived, not clinically confirmed. As a preprint not yet peer-reviewed, these results require independent replication before influencing personal decisions. Still, the practical implication is meaningful: for cognitively motivated adults, a gradual, flexible wind-down from full-time work may be preferable to abrupt full retirement.