Advancing Patient-Centred AI with Adaptive Machine Learning

Paper by Dimitrios Kalogeropoulos, Paul Barach, Andrea Downing, Stefaan Verhulst and Maryam B. Lustberg: “Healthcare services and data ecosystems remain fragmented, inequitable, and misaligned with the real-world needs of patients, clinicians, and public health systems. Existing pathways to patient-centred AI often lack contextual sensitivity and perpetuate disparities, limiting the transformative potential of AI to create personalised and inclusive care. This non-systematic narrative review examines three suitable pathways for integrating Artificial Intelligence (AI) into healthcare and identifies their limitations in realising patient-centred care. We propose a fourth pathway: Adaptive Machine Learning (AML). AML strategically integrates AI into learning health systems, allowing continuous model updates using population-level, context-sensitive real-world data. This quintuple aim based approach enhances personalisation, promotes quality and equity, and strengthens system resilience. We identify three critical enablers of AML: integrative data governance, adaptive study designs, and regulatory evidence sandbox facilities. Taken together these elements can advance the goal of sustainable digital health autonomy and responsible, collaborative data use. The aim of this study is to define a practical and ethically grounded framework for operationalising AML as a fourth pathway to patient-centred AI that aligns with international standards for responsible healthcare innovation, equitable governance, and digital transformation. Realising the full potential of AI in patient-centred healthcare requires urgent and coordinated actions across three priority areas to: (1) develop high-priority clinical use cases that demonstrate how AI can safely learn from real-world data and improve patient outcomes; (2) advance adaptive evaluation frameworks that reflect the lived experiences of diverse and underserved populations; and (3) establish regulatory evidence sandboxes to foster transparent, participatory, and multistakeholder innovation. Future research should prioritise integration of collective consent models and alignment of AI and medical device regulations with international governance toolkits to promote safe, patient-centred, inclusive, and trusted AI adoption in health ecosystems…(More)”.