Towards data-driven acute in-patient classification schemes: a hospital management perspective

In this paper, a case study describing the application of both supervised and unsupervised data mining techniques to grouping of acute in-patients in one of Melbourne's hospitals is presented. It is argued that the Diagnosis Related Groups currently used as a basis for the casemix funding system both in Australia and overseas cannot accurately and adequately describe patients' resource consumption patterns either for funding or for hospital resource management purposes. It is further suggested that supplementing the existing system with a data mining-based counterpart offers a number of gains as far as adequate capturing of patients' resource consumption is concerned.