In this paper we discuss different architectures for reasoning under uncertainty related to our ongoing research into building a medical decision support system. The uncertainty in the medical domain can be divided into a well understood part and a less understood part. This motivates the use of a hybrid decision support system, and in particular, we argue that a Bayesian network should be used for those parts of the domain that are well understood and can be explicitly modeled, whereas a case-based reasoning system should be employed to reason in parts of the domain where no such model is available. Four architectures that combine Bayesian networks and case-based reasoning are proposed, and our working hypothesis is that these hybrid systems each will perform better than either framework will do on its own.