Reverse Engineering Hospital Processes Out of Visited Nodes

During hospital planning, "flow-chart"-like depictions of processes (nodes and edges) are often overlaid over the floor plan of a clinic, in order to model patient pathways. However, such static definitions may not be accurate, as patients freely cross from one node into another, depending on individual treatment. Modelling each possible sequence of nodes is neither practical nor intelligible. This means graphical complexity due to overlapping edges. In our work, we thus present an edgeless process model in which each node acts as "dispatcher": It carries the logic for choosing a next node, thus enabling us to omit edges and introduce dynamic distribution of simulated patients. "Processes" (i.e. often-perceived sequences of activities) can nevertheless be inferred in an a posteriori step with an algorithm that will be presented, utilizing the actual patient pathways as indicator. Furthermore, the visualization of how processes are generated out of activity chains by reverse engineering will be discussed too.