Work

A User-Centered, Requirement-Driven Reference Architecture for AI-Enabled Mobile Heart Failure Management

Year: 2026

Type: conference-paper

Abstract: Heart failure management increasingly relies on mobile health technologies; however, existing solutions often lack modularity, transparent integration of heterogeneous data sources, and coordinated support for both patient self-management and clinical workflows within a unified framework. In addition, many systems are insufficiently described from an architectural and methodological perspective, limiting reproducibility and scalability. This work presents the co-design and co-development of mAIHeart, a mobile application developed through an iterative, user-centered system development lifecycle. System requirements were elicited through stakeholder engagement with patients, caregivers and healthcare professionals, complemented by literature review and analysis of existing digital health solutions. These requirements were systematically translated into modular, role-aware reference architecture supporting actor workflows. The application integrates retrospectively collected multicentre data within a client–server architecture and is designed to support validation using prospectively collected data through an AI-enabled decision support system. The proposed architecture enables multimodal data acquisition, including wearable-derived physiological signals and novel diagnostic modalities such as cardiac output response to stress tests and voice biomarkers. Dual user interfaces address distinct patient and healthcare professional needs, enabling risk monitoring, goal management, and secure data exchange. This paper focuses on the design methodology, architectural rationale, and functional capabilities of the mAIHeart system. Clinical relevance— The mAIHeart framework is intended to support heart failure self-management and clinical decision support through continuous, multimodal patient monitoring and structured information exchange between patients and healthcare professionals. By facilitating earlier risk identification and personalized care planning, the system may contribute to improved care coordination and more timely clinical interventions, subject to future clinical validation. Keywords—Mobile Health; Remote Monitoring; Artificial Intelligence; Heart Failure; Wearables

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.21902788