Development of a Determinant Framework to Guide the Translation of AI Systems in Clinical Care

Artificial Intelligence (AI) tools are increasingly being implemented in clinical settings, but there are inadequate studies evaluating the human, social, and organizational aspects of their integration, performance, and usage. Additionally, there is a lack of holistic frameworks for the implementation of AI systems in clinical spaces. This study aims to assess the socio-technical and systems factors that influence the implementation of AI tools in clinical settings and adapt it to create a predictive evaluation framework. Surveys and in-depth interviews will be conducted with staff at the University of Illinois Health center who interact with the Epic Sepsis tool. The data will be analyzed using Dedoose software. Based on the results, a framework and checklist will be developed to guide the predictive evaluation of AI tools in clinical settings. The checklist will be evaluated during the development of the DPI-diabetes AI tool for completeness and usefulness. This study will help in developing a new determinant framework.