Rehabilitation stands between acute care and long-term care. As such, it is difficult to define and harder still to measure. Put simply, it is a process, an attitude of mind, a philosophy; something that patients need but acute hospitals would rather do without. This paper discusses the problems of measuring and forecasting rehabilitative activity before explaining how behavioural models of flow demonstrate the interaction between acute, rehabilitative and long-stay care. Managers, directors of social service departments, and clinicians need scientifically valid methods that enable them to measure performance and to optimize decision-making. Inherent flaws in the current methods used to measure hospital activity must be overcome if rehabilitation is to occupy its rightful place in health care management.