Innovation in healthcare: leadership perceptions about the innovation characteristics of artificial intelligence- a qualitative interview study with healthcare leaders in Sweden

Abstract

Abstract Background There are substantial challenges in the implementation of intelligence (AI) applications in healthcare. This study aimed to provide an insight into implementation preconditions by exploring the perceptions of leaders and managers in Swedish healthcare concerning the intervention characteristics of AI as an innovation to be implemented into their organization. Methods The study had a deductive qualitative design, using constructs from the domain of intervention characteristics in the Consolidated Framework for Implementation Research (CFIR). Interviews were conducted with 26 leaders in healthcare. Results The participants perceived that AI could provide relative advantages in solutions for the management of care, for clinical decision-support and for early detection of disease and disease risk. The development of AI in the organization itself was perceived as the main current intervention source. The evidence strength behind AI-technology was questioned by the participants, who highlighted a lack of transparency and potential quality and safety risks. Although the participants perceived AI to be superior for humans in terms of effectiveness and precision in the analysis of medical imaging, they expressed uncertainty about the adaptability and trialability of AI in other clinical environments. The participants perceived that user and end-user views on design quality and packaging would impact implementation at all levels. Complexities such as the characteristics of the technology, the lack of consensus about AI as a concept, and the need for many implementation strategies to achieve potentially transformative practice change were spoken of, and the participants also expressed uncertainty about the costs involved in AI-implementation. Conclusion The leaders saw the potential of the technology and its use in practice, but also perceived that AI’s opacity limits its evidence strength, and that there was a high level of complexity both in AI itself and in introducing it in healthcare practice. More research is needed about the perceptions of AI implementation in other stakeholder groups and about outcomes from the implementation of AI in real-world situations. New theories, models and frameworks may need to be developed to meet the challenges related to the implementation of AI.