Developing a Dementia Platform Databank Using Multiple Existing Cohorts.

Area of Science:

• Neuroscience
• Gerontology
• Data Science

Background:

• Integrating diverse dementia cohorts is essential for large-scale research.
• Existing dementia patient data often exists in separate, non-interoperable cohorts.
• Challenges in data harmonization hinder the creation of comprehensive dementia research platforms.

Purpose of the Study:

• To evaluate the feasibility of building an integrated dementia platform.
• To converge pre-existing dementia cohorts from multiple sources.
• To identify barriers and propose solutions for cohort data integration.

Main Methods:

• Pilot project integrating four distinct dementia cohorts: CREDOS, K-BASE, EPINEF, and DPKR.
• Data from 29,916 patients were analyzed with 348 integrated variables.
• Assessment of data variability and missing information across cohorts.

Main Results:

• Successfully integrated data from 29,916 patients across four cohorts.
• Identified significant variability in study variables and measurements as a key integration barrier.
• High rates of missing information (≥50%) were noted for lifestyle and demographic variables.

Conclusions:

• Integrating pre-existing dementia cohorts presents significant challenges due to data heterogeneity.
• A standardized protocol considering data integration from the cohort establishment stage is necessary.
• Future efforts should focus on developing robust data harmonization strategies for dementia research.