Artificial intelligence for building learning Health Care Organizations

About thirty years of research in Artificial Intelligence in Medicine (AIM), together with a partial failure to disseminate AIM systems despite the significant progress in developing the underlying methodologies, has taught that AIM is not a field that can be separated from the rest of medical informatics and health economics. Since medicine is inherently an information-management task, effective decision-support systems are dependent on the development of integrated environments for communication and computing that allow merging of those systems with other patient data and resource management applications. The explosion of communication networks raised more recently another goal for AIM researchers: the full exploitation of those facilities require to model the organization where health care providers and managers work. This means that the AIM community has to acquire knowledge from other fundamental disciplines, as organization theory, sociology, ethnography, in order to exploit its modeling methodologies to represent behavior within an organization. It will allow the development of systems able to support collaborative work among everybody involved in patient care and organization management. A most promising approach is the exploitation of workflow management systems. They support the modeling and execution of workflows, which focus on the behavioral aspects of personnel involved in clinical processes. Workflow management systems provide tools for the design and implementation of innovative workflow-based Hospital Information Systems. This represents a great challenge for AIM researchers to prove that their theoretical background is essential to build those innovative systems.