Bed planning using queuing theory models of hospital occupancy: a sensitivity analysis.

Long-range planning for hospital beds is be? coming an increasingly common phenomenon. However, the tools available for carrying out such planning are not entirely satisfactory. This paper focuses on models for determining how many acute beds are needed to give adequate service to a specified series of patients without undue delay. Although this sidesteps the issue of appropriateness of utilization (by requiring that the issue be settled before specifying the forecasted demands), the problem is still diffi? cult. Involved is a trade-off between low average bed occupancy and shortages of beds due to unusually high demand. Stimson and Stimson1 describe numerous modeling efforts that are directed toward both management and planning of hospitals. The studies cited vary from Markovian models of patient progress to queuing and simulation models of the hospital or units thereof, includ? ing varying degrees of administrative control over census. However, in spite of these and many other studies, we find bed planning being done by much simpler methods?for example, on the basis of published desirable occupancy rates, or on rules based on theory but modified in an heuristic fashion to more accurately re? flect the local situation2. In describing the difficulties encountered when attempting to use the more complex models, Martin3 suggests that the use of simple techniques may actually be more appropriate. In order to develop a method that can be widely used, one must understand what the