Why are we trying to reduce length of stay? Evaluation of the costs and benefits of reducing time in hospital must start from the objectives that govern change.

Introduction In many different countries, those responsible for the care of patients in hospital are under some degree of political and managerial pressure to keep length of stay to a minimum. In the United Kingdom, for example, length of stay is one of the Health Service indicators on which some managers' performance related pay depends. It is known that patients admitted for treatment with the same condition may experience considerable variation in length of stay. And it is also widely held that reductions of time spent in hospital will reduce costs without compromising patient outcomes. What are the implications of this, and what exactly is the basis for it? It is not difficult to see why reduction in stay is a frequent focus for managerial attention. The concept is an accessible one; and progress in such a reduction can be measured relatively easily. Furthermore, the alternatives are fraught with difficulty. Most managers are aware that action on their part to encourage other improvements in a service-for example, through increasing clinical effectiveness-may be seen as an attempt to infringe clinical freedom. Clinicians have their own reasons for holding the view that it is desirable to reduce time spent in hospital. There is a tradition of rather dire warnings about the dangers of a long stay in hospital. Often length of stay is seen to be synonymous with bed rest. For example, Asher as early as 19481 said, "Beneath the comfort of the blanket there lurk a host of formidable dangers." An editorial in the BMY a year later put forward much the same view, and in 1960 in an article in the Lancet entitled "Hotels or hospitals?" Stallworthy' berated London teaching hospitals for their variability in length of stay and pointed out the reduced throughput that this implied. Similar articles have appeared since that time, and a recent newspaper article described the substantial reduction in time spent in hospital as a result of realisation by the medical profession that bed rest "lowers morale and causes insomnia and constipation".4 The aims of this review are to describe variation in length of hospital stay (particularly in relation to surgical procedures), along with currently available explanations for that variation; and to examine its relation to health outcomes and costs. In particular, the review considers four questions: * What determines length of stay? * Does length of stay make a difference to patients' health outcomes? * Do reductions in length of stay really save money? * Should we be trying to reduce length of stay?