Bivariate joint models for survival and change of cognitive function

Area of Science:

• Gerontology
• Biostatistics
• Longitudinal Data Analysis

Background:

• Cognitive function changes are crucial in ageing research.
• Analyzing multiple cognitive test scores and accounting for participant attrition is complex.
• Existing models may not fully capture the relationship between cognitive decline and survival.

Purpose of the Study:

• To develop a novel joint statistical model for analyzing bivariate longitudinal cognitive test scores.
• To incorporate survival analysis to account for attrition due to death in ageing studies.
• To investigate the correlation between two cognitive test scores over time.

Main Methods:

• Utilized an extension of the bivariate binomial distribution to model two integer-valued cognitive test scores.
• Employed Weibull and Gompertz hazard models to address attrition from death.
• Constructed a shared random-effects model with bivariate normal distribution for random effects.
• Integrated random effects to link the bivariate longitudinal and survival models.

Main Results:

• The proposed joint model successfully integrates bivariate longitudinal cognitive data with survival data.
• Demonstrated the ability to model the correlation between two distinct cognitive measures.
• Successfully accounted for the impact of attrition on the analysis of cognitive trajectories.

Conclusions:

• The developed joint model provides a robust framework for analyzing cognitive ageing and survival.
• This approach enhances the understanding of cognitive function changes in ageing populations.
• The model is applicable to real-world longitudinal ageing datasets, such as the English Longitudinal Study of Ageing.