A joint model for multiple dynamic processes and clinical endpoints: Application to Alzheimer's disease

Area of Science:

• Neuroscience
• Biostatistics
• Gerontology

Background:

• Alzheimer's disease (AD) involves progressive brain impairments affecting cognition and function.
• Previous studies analyzed AD components independently due to a lack of joint statistical models.
• Interrelatedness of AD progression necessitates a unified analytical approach.

Purpose of the Study:

• To propose a novel joint model for analyzing multiple correlated components in Alzheimer's disease.
• To simultaneously describe the dynamics of latent processes and their relationship to diagnosis.
• To accommodate competing clinical endpoints within the model.

Main Methods:

• Developed a joint model for multivariate longitudinal data and time-to-event analysis.
• Modeled diagnosis as exceeding a covariate-specific threshold of a combined pathological process.
• Utilized a maximum likelihood estimation procedure, implemented in an R package.
• Handled competing risks for clinical endpoints.

Main Results:

• The proposed model successfully integrates multiple correlated disease components.
• The estimation procedure is computationally feasible and robust, validated by simulations.
• The method was applied to a French cohort, analyzing clinical manifestations and risks of dementia and death.

Conclusions:

• Joint modeling provides a more comprehensive understanding of Alzheimer's disease progression.
• The new statistical framework can capture the complexity of dementia diagnosis.
• This approach offers a valuable tool for analyzing longitudinal data in neurodegenerative disease research.