Reengineering Clinical Decision Support Systems for Artificial Intelligence

There has been a tremendous growth of digital health applications within healthcare. Clinicians are inundated with new types of data to synthesize in a timely manner to make a clinical decisions about the care of their patients. The overwhelming abundance of data leads to physician burnout, which is a major problem within healthcare. In parallel, there is an immense hype building up about implementation of Artificial Intelligence (AI) technologies, such as Machine Learning or Deep Learning, to augment clinician decision making processes. Healthcare is a highly regulated environment so it's imperative to involve clinicians and data scientists in the entire model development, validation and implementation lifecycle. There ought to be a mechanism in place for integration of human feedback to build trust in the AI model, through human in the loop implementation models and participatory design approaches. The overarching aim of this stury is to formulate a problem statement and propose the development a system dynamics model highlighting the feedback loops within clinical decision-making workflows leading to diffusion of innovation of AI within healthcare. We propose system dynamics modelling as a mechanism to articulate the problems that are best suited for AI models, and conceptualize how the models would fit into the current workflow.