Key points are not available for this paper at this time.
Abstract Background: Artificial intelligence (AI) in healthcare has grown rapidly in recent years. The United Kingdom government recognises AI’s potential to enhance National Health Services with increased funding available. Despite this, AI acceptance presents unique challenges in primary care (PC), characterised by fragmented structures and multiple government departments. The organisational levels within PC are categorised as macro, meso, and micro levels. Many existing studies focus on micro-level stakeholders. Methods: This study investigates the factors affecting trust in artificial intelligence (AI) within PC settings through a mixed-methods approach. An online survey addressed this research gap by encompassing stakeholder perspectives at all organisational levels. To validate the results in-depth semi-structured interviews were conducted with some survey participants enabling triangulation of the data. Results: The results demonstrate the crucial role of meso-level stakeholders in facilitating trust in and acceptance of AI. Thematic analysis identified key barriers which include: a requirement for skills development, concerns about job displacement and factors associated with resistance to change. The study also highlighted disparities in care and potential inequities arising from varied AI usage rates. Public perception, leadership approval and media influence were identified as critical factors needing careful management to ensure successful AI integration in healthcare. Conclusion: The key contribution to the research field is the data from all stakeholder levels on the perceptions of AI for PC. Despite the study's robustness, limitations such as self-selection bias and low interview participation were noted. The findings underscore the necessity for ethical AI systems, targeted stakeholder engagement, and strategies to ensure equitable and effective AI implementation in PC. Further research in the relationship between trust and equity of care would be beneficial to the important research in the field of AI for PC.
Sides et al. (Fri,) studied this question.