Ensuring optimum use of scarce resources is one of the largest challenges facing health providers today. However it is not easy to generate an optimised schedule, as the health system is unusually and highly dynamic. Scheduling systems must be extremely flexible while still producing an efficient, acceptable schedule. Furthermore the scheduling system should be able to cross health boundaries inside and outside hospitals to perform load sharing. To solve this problem we propose an encoding of the patient scheduling problem as a dynamic distributed constraint optimisation problem and show how it can be solved using Support Based Distributed Optimisation. The resulting system will be able to generate good schedules and update them in real time. It is also able to maintain privacy across hospital boundaries to enable load balancing.