This deliverable outlines the development of a streamlined, interoperable data model to support integration and reuse of clinical and biosample data across Alzheimer’s disease (AD), Lewy Body Dementia (DLB), and Parkinson’s disease (PD). Designed to underpin EPND case studies (WP5) and patient-level discovery functionality (WP1), the model captures essential information on disease progression, clinical phenotype, and biomarker status, including genetically defined targets.

Led by the University of Oxford, the development process involved regular consultation with stakeholders from many of EPND's partner organisations across WP1, WP3, WP4 and WP5. Following detailed discussions, the model was split into two core components:

• The Analysis Core, comprising approximately 50 harmonised variables per disease, supports research on disease mechanisms and was informed by data from major cohorts, WP5 case studies, and contributions from the Critical Path for Parkinson’s consortium.
• The Discovery Core, derived from the Analysis Core, includes 10–20 high-level concepts optimised for sample and patient discovery, in collaboration with WP1 and EPND cohort discovery networks.

Variables are grouped into higher-level domains (such as demographics, imaging, cognition, and biomarkers) with attention to interoperability between overlapping measures (e.g. MMSE vs MOCA) and data structure (categorical vs continuous).

EPND has produced a minimal, common-use dataset of 159 variables, covering clinical features (diagnostic, demographic, medication status, family history, functionality, vital signs), as well as imaging and biosample variables. Each domain includes both overarching variables and linked sub-variables, structured to support secondary analysis across diseases. The full variable list is available on request as an Excel file and will continue to be refined during the five-year project.

For more information, find the full “SOP Minimal Dataset for Data Glossary” here.