Page - 20 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

a system, a common mistake made by those who advocate big data in healthcare. We have however modelled urgent care as a part of the larger system of a hospital and the wider health system, for example across a district and a city region, using a systems- within-systems approach that does fit well with GST. We recognize health systems as GST Level 8 open systems; the relationship between system and environment is complex and evolving. In this context, process mining is useful in looking for those patterns and structures that emerge from a holistic view of the system. In the UK there have been national targets for at least 95% of patients attending Accident and Emergency departments to be admitted, transferred or discharged within four hours. A pattern that emerges from process mining many such departments is that a median of 3.9 hours is common. Root cause analysis discussions with domain experts suggests this is game theory at work. The national target leads to the perverse behavior that the staff wait until, and then respond to, the impending deadline perhaps also believing that a full waiting room and a long wait will discourage less seriously ill patients. We also found evidence that the patterns and sequences of processes change during the day. Standard process conformance metrics were noticeably at their worst around early evening when routine processes give way to a period of apparent chaos with, for example, beds being requested for patients that had not yet been seen by a clinical specialist. We traced this flurry of activity to the time when the overnight shift starts work and a new allocation of beds become available; our discussions suggest that the new shift prioritize operational concerns such as booking beds over the routine updating of the computer system. In both cases these are very human activity patterns that can be explained through GST and only revealed by systematic enquiry. The other contribution from GST has been the construction of models to represent systems of study. Simple models such as process maps and mathematical formulae can be seen as GST Level 2 or 3 and therefore inadequate for explaining the behavior of a GST Level 8 organization. In the ClearPath method we use a care pathway simulation tool called NETIMIS (www.netimis.com) to present dynamic, runnable models back to multidisciplinary teams as part of a facilitated discussion about care pathway improvement. Simulation modelling might be seen as GST Level 4 and therefore inadequate in capturing the complexity of real-life healthcare. However, the real learning in LHS is still done by people so the discussion and the interactions and ideas it sparks are the real outputs of process mining of care pathways. 4. Discussion 4.1. Is GST relevant to modern health informatics? The enduring strength of GST is that it opens a window into a powerful way of viewing the world. At its most general, it sees the world as made of systems many of which are dynamic, complex and ever changing _ a melting pot of complexity where structures still emerge and have permanence while the relationships that hold them together are maintained, a wave crashing on a beach, a flight of birds forming a characteristic “V” shape. In our healthcare contexts, a patient’s body fighting serious infection and a surgical team at the end of a tough shift while also perhaps battling with a stubborn computer system. Or the cash-strapped health provider organization that spent too much procuring that computer system because it lacked the internal competencies to appreciate the importance of health informatics. O.Johnson /GeneralSystemTheoryand theUseofProcessMining to ImproveCarePathways20