The study systematically reviews stakeholder perspectives on the implementation of artificial intelligence (AI) in diagnostics, a crucial area given the global increase in demand for diagnostic services and existing workforce shortages. Through a comprehensive search across multiple databases, 44 articles were analyzed, involving 689 interviewees and 402 focus group participants, spanning four main stakeholder groups: patients, clinicians, researchers, and healthcare leaders. A notable finding is the under-representation of patients, researchers, and leaders in the literature.

A key contribution of this research is the development of a conceptual model that underscores the importance of trust, engagement, and collaboration among stakeholders. Additionally, the study introduces a modified Non-adoption, Abandonment, Scale-up, Spread, Sustainability (NASSS) framework, specifically tailored for diagnostic AI, termed NASSS-AI. This adaptation provides a structured approach to understanding and addressing the complexities involved in implementing diagnostic AI technologies.

The findings emphasize the need for inclusive representation of all stakeholder groups in future research and implementation strategies. The study suggests that successful implementation of diagnostic AI should consider how the technology fits within the broader healthcare ecosystem, as outlined in the extended NASSS-AI framework, and how it addresses stakeholder priorities and concerns. This research is supported by an NIHR Doctoral Research Fellowship grant, highlighting its significance in guiding future diagnostic AI initiatives.