Analysing the Impact of Changes in User Interface of e-Health Record Systems on Clinical Pathways using Process Mining

The provision of care in a hospital includes a series of activities that are often recorded in the electronic health record (EHR) systems. Analysing the data in these EHRs has the potential to support the understanding of care processes and exploring the opportunities for process improvement.
One of the emerging data analytics approaches for such analyses is process mining, and one critical challenge in working with EHR data is that processes might change over time. This thesis uses a process mining approach to detect process change over time and analyse the impact of those changes on the EHR data. The overall aim is to
summarise the attributable change in the data due to the process so that clinicians can better analyse the data.
Three datasets were used in this study to understand the variability of the EHR systems. The first dataset is a publicly available EHR data that was used for developing the methods and supporting the reproducibility of the research. The second dataset is a de-identified subset of the database of cancer patients from the Leeds Cancer Centre. The second dataset was used in the experiments to improve on the results of a previous study using the same dataset. The third dataset was the full Leeds Cancer Centre EHR database after more comprehensive ethics was approved. In the third dataset, experiments were done to analyse the impact of a known system change on clinical pathways and to explore process change over time without a known system change. All three datasets were analysed using process mining.
Process mining was shown to be useful for analysing clinical pathways and exploring process changes over time. It can be used to visualise the process before and after a known change. When the system change is unknown, process mining can be used to explore the process execution over time and identify the potential period where the system was changed. This thesis explores some aspects of the complex interrelatedness of process and user interface (UI) of the EHR system.