Cognition-Mortality Associations Are More Pronounced When Estimated Jointly in Longitudinal and Time-to-Event Models

Area of Science:

• Gerontology
• Cognitive Epidemiology
• Biostatistics

Background:

• Growing aging populations worldwide increase interest in cognitive decline and mortality risk.
• Existing research uses comparative or two-stage modeling for cognitive trajectories and mortality.
• Joint longitudinal-survival models offer a more integrated approach to these associations.

Purpose of the Study:

• To compare different analytic approaches for estimating cognitive decline and its association with mortality risk.
• To investigate how methodological differences influence estimates across cognitive domains.
• To highlight the utility of joint longitudinal-survival models in cognitive epidemiology.

Main Methods:

• Applied a two-stage modeling framework and joint longitudinal-survival models.
• Utilized data from a large-sample, repeated-measures study of older adults (N=5,954; ages 50-87).
• Examined cognitive trajectories and their association with mortality risk, adjusting for covariates.

Main Results:

• Worse cognitive performance observed in decedents and in joint models, attenuated by covariates.
• Better cognitive performance predicted lower mortality risk; associations were stronger in joint models.
• Crystallized abilities' association with mortality risk emerged only under joint estimation.

Conclusions:

• Joint longitudinal-survival models provide more robust estimates of cognition-mortality associations.
• These models are vital for understanding cognitive reserve, particularly for crystallized abilities.
• Joint models are underutilized but essential for advancing cognitive epidemiology research.