Modified Needleman–Wunsch algorithm for clinical pathway clustering

Clinical pathways are used to guide clinicians provide a standardised delivery of care. Because their standardisation, the aim clinical is reduce variation in both care process and patient outcomes. When learning from data through mining, it common practice represent each pathway as string corresponding movements activities. Clustering techniques popular methods for therefore this paper focuses on distance metrics applied k-medoids clustering. The two main aims firstly, develop technique that seamlessly integrates expert information with secondly, metric purpose data. overall goal was allow more meaningful clustering results be found by adding context into similarity calculation. Eight applicability discussed. These prove give an arbitrary distance, without consideration context, produce different results. As result, describes development new metric, modified Needleman–Wunsch algorithm, allows interaction calculation assigning groupings rankings activities, which strings. This algorithm has been developed partnership UK’s National Health Service (NHS) focus lung cancer pathway, however handling application any disease type. method contained within Sim.Pro.Flow, publicly available decision support tool.