|dc.description.abstract||Unravelling cell characteristics is key for the accurate diagnosis and treatment of haemato-oncological diseases.
Major advancements in the field have led to high-throughput single-cell technologies, providing large and multidimensional datasets. Detecting the relevant information in these datasets remains a challenge, however. Artificial intelligence (AI) can support the human capability and can thereby prove to be the next step in improving patient survival rates. However, although the academic interest in AI developments is substantial, the real-world deployment remains behind. This was referred to as the invention-innovation gap, invention being an idea, and innovation being the commercialization of that idea.
Starting from the idea that innovation does not happen in isolation but is part of a system, both the institutional and technological environment are crucial for an invention to become an innovation. In that line of thought, institutional entrepreneurs are actors that actively transform the institutional environment in favour of an emerging technology. In this research, it was examined whether the haemato-oncological field offers an enabling environment for institutional entrepreneurship to occur in the context of AI technologies.
The institutional entrepreneurship approach was specified for the haemato-oncological field with domains of the non-adoption, abandonment, scale-up, spread and sustainability (NASSS) approach, a framework designed to study failure of innovation in the medical field. An explorative, qualitative research design was followed, using a single case study. Data was gathered through 18 semi-structured interviews and triangulated by 10 expert events. An abductive approach was followed, starting from prior theoretical knowledge, subsequently translating empirical findings into theoretical contributions.
This paper presents an extensive description and in-depth analysis of the key findings. This resulted in four technological prerequisites that should be fulfilled to provide an environment in which institutional entrepreneurship can help bridge the invention-innovation gap: a seamless fit with the clinical workflow, an initial focus on simple disease patterns, validity proof and a financial injection. Institutionally, the field-level characteristics on an individual level and wider system level are enabling for institutional entrepreneurship to occur. At an organizational level (inter-hospitals), more standardization is needed. Furthermore, establishing a favourable social position is key, by connecting to key opinion leaders and central players (manufacturers), and multiple embeddedness in both the data sciences field and medical field. Concluding, institutional entrepreneurship is likely to occur for tailor-made AI solutions. Future standardization on an inter-hospital level could provide opportunities for field-wide AI innovations.||