Structured, Harmonized, and Interoperable Integration of Clinical Routine Data to Compute Heart Failure Risk Scores

Area of Science:

• Medical Informatics
• Clinical Decision Support
• Cardiology

Background:

• Accurate risk prediction in heart failure (HF) is vital for personalized patient management and resource allocation.
• Existing HF risk scores are underutilized in clinical practice due to data accessibility and interoperability challenges.
• Routine clinical data presents an opportunity for more accessible risk assessment.

Purpose of the Study:

• To demonstrate the feasibility of a multi-site solution for deriving and calculating HF risk scores from routine clinical data.
• To implement an interoperable data collection framework for harmonized HF phenotypic data.
• To overcome practical barriers in implementing risk scores into daily healthcare.

Main Methods:

• Developed an interoperable solution within the openEHR framework to collect a harmonized HF phenotypic core data set (CDS).
• Implemented a multi-site data retrieval and processing pipeline across five medical centers.
• Automated data extraction from primary systems to minimize manual data entry for calculating MAGGIC and Barcelona Bio-HF scores.

Main Results:

• Successfully demonstrated the feasibility of using clinical routine data across multiple sites for HF risk score computation.
• Validated the interoperable data structures and processing pipeline in a real-world, multi-center setting.
• Showcased the practical application of deriving two specific HF risk scores (MAGGIC and Barcelona Bio-HF) from harmonized data.

Conclusions:

• A multi-site, interoperable approach enables the feasible computation of HF risk scores using routine clinical data.
• This solution can be extended to various clinical applications beyond HF risk prediction.
• Facilitating the use of routine data for risk scores can improve clinical decision-making and healthcare resource management.