We consider the partitioning of care types into wings from the perspective of a hospital administrator, who wishes to optimize the use of a fixed number of beds that provide services for heterogeneous care types. Patients of each care type request admission to the hospital stochastically, and if admitted, require a stochastic length-of-stay and provide the hospital some utility. If a bed is not immediately available, patients wait a random amount of time for admission, which may take place in a physical queue (as in an emergency room) or in a virtual queue (as in scheduled procedures). If their wait becomes excessive, they abandon the queue to seek care elsewhere. The hospital administrator decides on the number of wings to form, the number of beds to allocate to each wing, and the set of care types to assign to each wing in order to maximize the total utility to the hospital. The administrator faces an inherent trade-off between forming large wings to pool demand and bed capacity, and forming specialized wings to focus on narrow ranges of care types. Specialized wings not only provide advantages from focused care, but also allow the protection of beds for high utility care types. We provide an optimization model for the wing formation decision and address the advantages of focus endogenously in our model. Using data from a large urban teaching hospital in the United States along with a national database, we report on a number of managerial insights. In particular, as the overall demand increases across all care types, wings are formed to reserve more beds for higher utility care types, which leads to higher overall hospital utility but also some disparity across care types, as hospital access increases for some care types and decreases for others. Furthermore, as more wings are formed, overall bed occupancy in the hospital decreases. However, if sufficient focus is attained, shorter lengths-of-stay associated with focused care may increase overall patient throughput. We also observe that when patients are willing to wait longer for admission, the hospital tends to form more wings. This implies that hospitals that garner longer waits can form more specialized wings, and thereby benefit from focused care, whereas hospitals that cannot will tend to form fewer, if any, wings, choosing to pool demand and bed capacity
