Bridging Data Silos in Oncology with Modular Software for Federated Analysis on Fast Healthcare Interoperability...

Area of Science:

• Oncology
• Health Informatics
• Data Science

Background:

• Real-world data (RWD) from diverse sources like electronic health records and cancer registries offer valuable insights into patient populations beyond clinical trials.
• Six Bavarian university hospitals have established a collaborative research IT infrastructure to effectively leverage oncological RWD for advancing cancer research.

Purpose of the Study:

• To design, implement, and deploy a modular data transformation pipeline for oncological RWD.
• To convert RWD into Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR) format and subsequently into a tabular format.
• To prepare data for a federated analysis (FA) across six Bavarian Cancer Research Center university hospitals, ensuring privacy through decentralization.

Main Methods:

• Developed a pipeline to convert the oncological basic dataset (oBDS) into HL7 FHIR and tabular formats for FA.
• Ensured the pipeline's adaptability to diverse IT infrastructures and systems while maintaining data privacy via decentralization.
• Assessed pipeline functionality and validity by defining a cohort for two medical research questions and comparing FA results with the Bavarian Cancer Registry and local tumor documentation systems.

Main Results:

• Conducted an FA of 17,885 cancer cases from 2021/2022, identifying breast, prostate, and malignant melanoma as prevalent diagnoses.
• Observed gender-specific trends: larynx and esophagus cancers more common in males; breast and thyroid cancers more frequent in females.
• Noted discrepancies with the Bavarian Cancer Registry, likely due to differing time periods and data source scope, with the registry reporting approximately three times more cases.

Conclusions:

• The modular pipeline successfully transformed oncological RWD across six hospitals, enabling privacy-preserving federated analysis.
• The federated approach facilitated comprehensive data analysis while maintaining patient privacy.
• Future work includes supporting newer oBDS versions, automating data quality checks, and integrating additional clinical data to enhance federated health data networks for cancer research.