Expressing Diversity of Health Datasets in the European Health Data Space

Area of Science:

• Health Informatics
• Biomedical Data Science
• Health Equity Research

Background:

• Diverse health datasets are essential for equitable health interventions and outcomes.
• Current methods for assessing dataset diversity may be insufficient.
• The European Health Data Space (EHDS) aims to facilitate health data sharing.

Purpose of the Study:

• To explore methods for measuring diversity in health datasets.
• To investigate how dataset diversity information can be represented in the EHDS using metadata.
• To propose a framework for assessing the utility of multimodal datasets for research.

Main Methods:

• Review of existing literature on health data diversity metrics.
• Analysis of metadata standards for health data.
• Conceptual framework development for structured interaction with data owners.

Main Results:

• Diversity in health datasets encompasses sociodemographic, clinical, lifestyle, and genetic factors.
• Simple annotation is inadequate for capturing the full utility of diverse datasets.
• Structured interaction with data owners is necessary for accurate assessment of dataset utility.

Conclusions:

• Effective measurement and metadata representation of health dataset diversity are crucial for the EHDS.
• A more sophisticated approach than simple annotation is required to leverage diverse datasets for research.
• Facilitating structured dialogue with data owners will enhance the utility of multimodal health data.