The Multiagent Systems paradigm offers expressively rich and natural fit mechanisms for modeling and negotiation for solving distributed problems. Solving complex and distributed real world problems in dynamic domains however presents a significant challenge and requires the integration of technology innovation and domain expertise to create intelligent solutions. Scheduling of patients, staff, and resources for elective surgery in an under-resourced and overburdened public health system presents an excellent example of this class of problems. In this paper, we discuss the research challenges presented by the problem and outline our efforts of applying distributed constraint optimization, intelligent decision support, and prediction based theater allocation to address these challenges. We also discuss how these technologies can be used to drive better planning and change management in the context of surgery scheduling.