A Federated Database for Obesity Research: An IMI-SOPHIA Study.

Area of Science:

• Obesity research
• Data science in healthcare
• Biomarker discovery

Background:

• Obesity is often misclassified as a lifestyle choice, hindering its perception as a progressive disease.
• Effective obesity management requires better tools for stratifying patients based on disease risk and treatment response.
• Siloed clinical data limits comprehensive analysis and biomarker discovery for obesity.

Purpose of the Study:

• To establish a federated database for stratifying obesity phenotypes.
• To overcome data silos and enable collaborative analysis across multiple cohorts.
• To develop advanced tools for personalized obesity therapy.

Main Methods:

• Implemented a federated database using open-source DataSHIELD technology.
• Integrated 16 diverse cohorts (research, clinical trials, routine health data).
• Utilized the R statistical environment for remote, privacy-preserving analyses.

Main Results:

• Successfully federated data from 16 cohorts via a central gateway.
• Performed a proof-of-concept federated linear model (BMI and systolic blood pressure).
• Achieved point estimates comparable to traditional meta-analysis, without patient data disclosure.

Conclusions:

• The SOPHIA federated database enables secure, large-scale data analysis for obesity research.
• This approach facilitates reproducible biomarker discovery and personalized treatment strategies.
• Sets a benchmark for safe, federated analysis across diverse data types and stakeholders.