Document resource
Introduction Access to nailfold capillaroscopy is limited by a shortage of trained specialists and the complexity of image acquisition and interpretation. Artificial intelligence (AI) offers the potential to facilitate access to capillaroscopy by enabling non-specialists to acquire and interpret images with automated image capture and reporting. We developed an AI-guided system with automated image acquisition, real-time feedback and reporting functions. Our aim was to explore user perspectives of healthcare professionals and patients to ensure the tool was acceptable, usable and sustainable.Material and Methods Online focus groups were conducted with participants from a pilot study of AI-guided image acquisition and interpretation, including six patients and eight healthcare professionals from three hospitals. Sessions were audio-recorded, transcribed, and analysed using framework analysis. Responses were mapped against predetermined themes including training, image acquisition, reporting, clinical practice integration and patient experience. Questionnaires collected feedback about usability, tolerability and system improvements (n=13 staff; N=87 patients). Involvement from patients, carers and clinicians shaped the research design and delivery. Further focus groups are scheduled for Autumn 2025 to assess iterative improvements.Results Initial findings indicated widespread support for the system’s potential to save time, travel and costs. Patients described diverse diagnostic journeys, often delayed by limited awareness of systemic sclerosis in primary care. Capillaroscopy was well-tolerated (98.8% agreed/strongly agreed they felt physically comfortable) and viewed as a ‘step forward’ compared to older diagnostic methods. Both patients and clinicians supported AI assistance, provided clinicians remained responsible for interpretation. Overall, 69% of staff agreed or strongly agreed the software task was easy enough to complete. Nonetheless, inexperienced staff reported longer image acquisition times while they were gaining initial experience (53% agreed/strongly agreed they completed tasks quickly enough). The automated clinical reports were well received, though adjustments to layout and language were suggested to improve interpretation for non-specialists, emphasising overall disease assessment, image quality confidence and population-specific reference ranges. Logistical factors (e.g. Wi-Fi connectivity and clinic space) also influenced practical implementation.Conclusions AI-guided capillaroscopy has the potential to modernise and broaden access to systemic sclerosis diagnostics, improving efficiency and equity in care. User research identified areas for improvement in both software and reporting. Findings also highlighted the importance of considering wider service redesign to support adoption, addressing not only technical usability but also patient experience and diagnostic pathways. Future user research will evaluate the second iteration of the system and inform implementation strategies.