Knowledge about resource consumption and utilisation is vital in modern healthcare environments. In order to manage both human and material resources efficiently, a typical approach is to group the patients based on common characteristics. The most widely used approach is driven by the Case Mix funding formula, namely to classify patients according to diagnostic related groups (DRGs). Although it is clinically meaningful, our experience suggests that DRG groupings do not necessarily present a sound basis for relevant knowledge generation. In this paper, we propose an alternative grouping of the patients based on a neural clustering approach, which generates homogeneous groups of patients with similar resource utilization characteristics. Demographic information is used to generate the clusters, which reveal interesting differences in resource utilisation patterns. A detailed case study is presented to demonstrate the quality of knowledge generated by this process. The proposed approach can therefore be seen as an evidence-based predictive tool with high-knowledge generation capabilities.