Patient and clinician voices in shaping AI for equitable heart failure management

Abstract Background Artificial intelligence (AI) tools for healthcare are often developed in high-income countries, based on data and care models specific to those settings. This limits their applicability in low- and middle-income settings, where infrastructures, patient populations, and clinical workflows differ significantly. The AI4HF project promotes equitable innovation by engaging patients and healthcare professionals from 3 continents in co-designing trustworthy AI tools for tailored heart failure management. Methods Local Working Groups (WGs) were set up at 5 clinical sites (Netherlands, Spain, Czechia, Tanzania, and Peru) to assess the cultural and organizational appropriateness of the proposed AI tools and identify real-world implementation pathways. Each WG includes 6-8 patients and 6-8 healthcare professionals. Three rounds of workshops were conducted to date using a Design Thinking approach: 1. focused on empathy via persona creation; 2. mapped interactions among patients, professionals, and technology; and 3. explored sources of bias and tested the low-fidelity prototypes of the tools. Results Across sites, key contextual barriers emerged such as long travel to care, low health literacy, mistrust in technology, and limited digital literacy among patients and professionals. This evidence is guiding the tools development and informing higher-level multi-stakeholder discussions. Patients felt empowered by their involvement; professionals valued the opportunity to engage beyond clinical appointments. Many expressed interest in continuing the WGs beyond the project's lifetime. Conclusions Local WGs are a feasible and impactful approach to co-design AI tools that are culturally and contextually adapted. This participatory method fosters trust, empowers patients, and builds awareness of local barriers. Continued biannual sessions will promote equitable, real-world integration of AI in healthcare, contributing to more inclusive innovation and better health outcomes. Key messages • Engaging local stakeholders ensures AI tools reflect real-world needs and promote equity in heart failure care across diverse global contexts. • Participatory dialogue empowers patients and enables healthcare professionals to engage beyond the clinic, driving more human-centred AI solutions in healthcare.