RIS ID
106574
Abstract
Admitted rehabilitation activity accounts for an increasing proportion of health care expenditure in Australia. This presentation describes the development of a tool to predict demand for rehabilitation care generated by acute inpatient episodes provided in Australian public sector facilities. Previous work by Dr Lynette Lee identified a set of "rehabilitation-sensitive AR-DRGs" , or AR-DRGs from which patients are more likely to require subsequent rehabilitation. In developing the predictive tool, this earlier work was extended, primarily by quantifying the degree of sensitivity and by incorporating the patient's age. The model uses variables, both clinical and demographic, found within routinely collected administrative data sets and could be used in routine service planning. Rehabilitation activity can be predicted as the number of episodes or the expected number of bed days. A strategy for "finding" additional bed days was proposed for situations where the prediction exceeds the current level of activity.
Publication Details
J. Green, J. McNamee, C. Kobel, H. Seraji & S. Lawrence "Development of a model to predict demand for rehabilitation from activity in acute care AR-DRGs", The Hague, Netherlands, 14-17 October 2015, (2015)