Leveraging Data-Driven Process Simulation to Support Capacity Management Decisions in Healthcare

Abstract

Healthcare expenditures are increasing worldwide. The increasing demand for healthcare services due to population growth, ageing, lifestyle factors, and raising prevalence of chronic diseases, coupled with the high costs of healthcare technologies, equipment, and treatments, have led to significant increases in healthcare spending over the years. In 2020, global spending on health reached $9 trillion, corresponding to 10.8% of the global Gross Domestic Product (GDP). Moreover, the total spending on health is growing faster than GDP. At the same time, healthcare budgets are under pressure due to economic austerity. Furthermore, the COVID-19 pandemic posed significant challenges to healthcare systems struggling to manage the surge in demand for hospital beds, ventilators, and other critical resources, as well as a vicious cycle of impaired job performance, burnouts, and shortage of healthcare professionals due to extreme work pressure and morally challenging decisions. Against this background, healthcare managers are confronted with various challenges to manage their scarce resources efficiently in order to safeguard high-quality healthcare services. Healthcare processes, which encompass the sequence of activities involved in providing healthcare services, play a crucial role in addressing these challenges. Proper management of healthcare processes and services is essential for optimising resource allocation, improving quality and patient safety, ensuring continuity of care, and promoting innovation in healthcare. Business Process Simulation (BPS), or simulation in short, can offer valuable decision-support information for healthcare managers confronted with Capacity Management (CM) decisions. In BPS, a computer model of a process is simulated to imitate the behaviour of a process in a virtual setting. This has the advantage of being able to test process modifications without having to implement these changes in practice. Using simulation, one could, e.g., safely test whether removing an underutilised medical devicewould not result in significantly higherwaiting times without having to try this in reality, and thereby unnecessarily endangering patients’ health. This thesis focuses on using the information stored in process execution data, e.g., from Hospital Information Systems (HIS) to develop Business Process Simulation models supporting effective Capacity Management decision-making in healthcare. Despite the intrinsic potential of Process Mining to support the development of BPS models, several prevailing challenges impede the path towards fully deploying Data-Driven Process Simulation (DDPS). A significant research gap persists as literature on the operationalisation of DDPS is characterised by a proof-of-concept depiction. This implies that the simulation models generated using DDPS typically exhibit unrealistic and oversimplified behaviour to be effectively applied for supporting CM decisions in healthcare. The practical limitations and shortcomings in existing DDPS approaches emphasise the compelling need to bridge the research gaps. By addressing these challenges, the objective of this thesis is to leverage the full potential of DDPS as a valuable tool for supporting CM decision-making in the complex healthcare environment.