A Distributed Decision Support Architecture for the Diagnosis and Treatment of Breast Cancer

Clinical decision support for the diagnosis and treatment of breast cancer needs to be provided for a multidisciplinary team to improve the care. The execution of clinical knowledge in an appropriate representation to support decisions, however, is typically centrally orchestrated and inconsistent with the nature and environment that specialists work together. The use of guideline language of PROforma for breast cancer has been examined with the issues raised, and an agent-oriented distributed decision support architecture is put forward. The key components of this architecture include a goal-decomposition structure (shaping the architecture), agent planning rules (individual decision-making), and agent argumentation rules (reasoning among decision options). The shift from a centralised decision support solution to a distributed one is illustrated using the breast cancer scenario and this generic approach will be applied to a wider range of clinical problems in future.