Meet our Project Partner: Leiden University Medical Center (LUMC)

General information

The Translational Data Science (TDS) lab connects practical problems in healthcare practices to fundamental challenges in data science to address both simultaneously. The TDS lab operates from Leiden University Medical Center (LUMC), specifically through Leiden University’s Health Campus in The Hague, whose central mission is to improve health equity through a sustainable approach with a broad health perspective.

LUMC brings invaluable knowledge and experience in translating data science innovations into digital health solutions to the INSAFEDARE project. This includes crucial contributions to data integration, synthetic data generation and validation, and the development and validation of digital health tools.

Achievements/Update

LUMC is contributing major work in the following areas of the INSAFEDARE project:

• Synthetic data generation

  • Developing a method for tabular data synthesis which yields high utility, without necessarily yielding high fidelity, which in turn improves privacy-preservation. The method was published in the Cell Patterns journal (https://doi.org/10.1016/j.patter.2025.101287).

  • Developing the synthyverse, a modular Python package for synthetic data generation and evaluation (https://github.com/synthyverse/synthyverse).

• Synthetic data evaluation

  • Providing an overview of common and effective evaluation metrics for synthetic data (fidelity, utility, and privacy).

  • Presenting a discussion on privacy in synthetic health data at the Medical Informatics Europe (2025) conference in Glasgow (https://doi.org/10.3233/shti250404).

• Data integration

  • Developing software components for integrating health data, applied to the Electronic Health Record repository MIMIC-IV as use case.

• Data analysis techniques

  • Providing an overview of data analysis techniques commonly used in Digital Health studies, by leveraging a Retrieval Augmented Generation pipeline which extracts relevant information from scientific articles.

  •