Semantic Harmonization of Clinical and mHealth Data for Heart Failure Diagnosis & Risk Stratification
Year: 2026
Type: conference-paper
Abstract: The volume and diversity of data originating from clinical information systems, mobile health applications, and wearable devices in heart failure (HF) care are rapidly increasing. These heterogeneous data streams offer new opportunities for early detection and personalized risk stratification, yet their effective use is hindered by the lack of semantic harmonization and interoperable representations. To date, few ontology-based approaches explicitly address the unified modeling of clinical, patient-reported, and wearable-derived HF data. This paper presents the STRATIFYHF Heart Failure Ontology (SHFO), a FAIR-compliant semantic framework designed to harmonize multimodal HF data and support interoperable diagnosis and risk prediction. Grounded in a clinician-defined reference data model, SHFO provides a schema-level ontology that captures conventional clinical measurements alongside patient-reported outcomes and continuous wearable-derived physiological signals. The ontology aligns with established clinical standards, including SNOMED CT, LOINC, and UCUM, while accommodating novel digital health constructs through project-specific extensions. Quantitative characterization demonstrates that SHFO offers a compact yet expressive semantic structure, enabling consistent data representation and validation across distributed systems. By addressing the growing need for semantic harmonization, SHFO establishes a robust foundation for scalable, transparent, and interpretable HF decision support and risk stratification. Clinical Relevance — SHFO integrates wearable-derived and patient-reported data by semantically aligning them with clinician-defined variables and established clinical standards, ensuring consistent representation alongside conventional clinical measurements. Transparency is achieved through explicit, standardized definitions of all data elements, enabling interoperable decision support without modifying existing clinical workflows.
Authors: Boucharas, Dimitrios, Grigorios, Kotoulas, Milovancev, Aleksandra, Kyprakis, Ioannis, Manousos, Dimitrios, Preveden, Andrej, Velicki, Lazar, Okwose, Nduka, Charman, Sarah, Fuller, Amy, Geroski, Tijana, Milicevic, Bogdan, Lazar, Dasic, Ognjen, Pavic, Atanasijevic, Aleksandar, Jakovljevic, Djordje, Filipovic, Nenad, Tsiknakis, Manolis, Fotiadis, Dimitrios
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7169626554
FWCI:
Cites:
Cited by:
Related to: