| dc.description.abstract | Clinical radiology in the Netherlands is being flooded with digital data, mainly in the form of medical images. Software applications that perform computerized automated image analysis, so-called artificial intelligence (AI) applications, are becoming increasingly accurate, showing better performance than trained radiologists for certain tasks. Yet, very few AI applications are currently implemented in routine clinical use in radiology departments of Dutch hospitals. The practical implementation of new technologies in the medical field, especially in hospital settings, depends on a range of different factors, such as the large variety of stakeholders involved, the rigid routines and strong professional identities, as well as the strict legal and regulatory standards to be abided. These factors hinder or facilitate the implementation process and often interact in dynamic ways, as demonstrated by the recently published nonadoption, abandonment, scale-up, spread and sustainability (NASSS) framework, which focuses explicitly on determinants of unsuccessful adoption. This research aimed at identifying facilitating and hindering factors to the successful implementation of AI applications in Dutch radiology departments and how the hindering factors could be overcome. Due to the early stage of adoption of AI applications in radiology, an exploratory, qualitative research design was followed, based on an embedded multiple case study. In a first deductive step, guiding propositions were derived from the existing NASSS framework. In a second inductive step, the framework was refined for the case of AI applications in radiology. The results showed a wide array of facilitating and hindering factors to successful implementation of AI applications in Dutch radiology departments. Among the most important facilitating factors is the presence of a ‘local champion’, an individual with a strong personal interest in AI applications, which most often initiated and actively pushed forward the implementation of AI applications in their respective organization. Among the most prominent hindering factors are the uncertain added-value for clinical practice of AI applications, which causes low acceptance of AI applications among adopters and complicates the mobilization of funds to acquire AI applications. Furthermore, the failure to include all relevant stakeholders in the planning and execution phase of the implementation of AI applications was found a major hindering factor. To increase low acceptance among adopters, more evidence of the added-benefit of their AI applications in the clinical setting is needed. Also, all affected stakeholders (most notably radiologists and referring clinicians) should be included in the decisions and the design of implementation processes of AI applications. | |
