Front Public Health. 2025 Dec 4;13:1713882. doi: 10.3389/fpubh.2025.1713882. eCollection 2025.
ABSTRACT
INTRODUCTION: Within the Horizon Europe-funded AI-POD (AI-based tools for the Prediction of Obesity-related vascular Diseases) project, a clinical decision support system and citizen-facing mobile health application are being developed to enable personalized cardiovascular risk prediction in individuals living with obesity, through the integration of clinical, imaging, laboratory and lifestyle data. To inform the responsible development and implementation of these innovations, this study explored stakeholder perspectives on anticipated benefits, concerns, and challenges across four European countries.
METHODS: Semi-structured interviews were conducted with 21 stakeholders between February and July 2025. Participants represented diverse (professional) backgrounds including radiology (n = 5), artificial intelligence (n = 4), medical informatics and healthcare innovation (n = 2), dietetics (n = 2), endocrinology (n = 2), and general practice (n = 1). In addition, our sample included two patient representatives (n = 2), as well as individuals with expertise in social sciences and ethics (n = 1), law and policy (n = 1), and public health (n = 1). Most were based in Belgium (n = 16), with others from Austria (n = 3), the United Kingdom (n = 1), and Sweden (n = 1). Seven participants were affiliated with the AI-POD consortium, while 14 were external experts. All interviews were audio-recorded, transcribed verbatim, and analyzed using inductive content analysis.
RESULTS: Participants identified several benefits of the AI-POD tools, including the integration of multimodal data, improved risk stratification, and enhanced patient engagement and health literacy. However, concerns were raised about potential anxiety stemming from risk scores, the reinforcement of weight stigma, limited evidence supporting personalized lifestyle recommendations, and equitable access to the tools. Key challenges included data heterogeneity, algorithmic bias, small sample sizes, and technological barriers such as device incompatibility and varying levels of digital literacy. Participants anticipated that implementation would be further complicated by difficulties in engaging patients and by healthcare professionals' reluctance to adopt solutions that fall outside established guidelines.
CONCLUSION: While stakeholders acknowledged the promise of the AI-POD tools for advancing personalized cardiovascular risk prediction in individuals living with obesity, they also identified critical challenges related to equitable access, sustained user engagement, and effective integration into clinical practice. Addressing these challenges will be essential for the successful implementation, adoption, and uptake of the tools envisioned within the AI-POD project.
PMID:41426664 | PMC:PMC12711769 | DOI:10.3389/fpubh.2025.1713882