Healthcare organisations have the characteristics of complex adaptive systems (CAS) and their behaviours are difficult to predict, because small perturbations in inputs have non-linear effects. Furthermore, healthcare organisations are required to simultaneously balance multiple performance measures, such as the risks to patients, staff workloads and costs. Decisions made to improve one measure may affect the others. The literature that describes healthcare as a CAS, has provided qualitative explanations, but has not advanced the discussion into estimating quantitative behaviours. This thesis provides an approach to estimate the complex dynamic aspects of the performance using a tertiary hospital Intensive Care Unit (ICU) and the complexity involved in the management of the decisions related to patient throughput.
Models of the components of the ICU and their behaviours and interactions between components were developed to simulate patient flow and the impact of management decisions on ICU performance. The simulation is validated using historical patient records and by an experienced clinical manager. The simulation is designed to study the impact of the organisational decisions that affect patient throughput. The decisions considered within this study relate to strategic-planning, tactical-planning and the actions taken on a day-to-day basis. This thesis demonstrates that the simulation developed to mimic the components of the ICU and their interactions can generate insights into the ICU performance that cannot be directly extracted from historical data. Such insights can be useful for estimating the quantitative effects of proposed decisions and their impacts on resource allocations.
The simulation models developed provide mechanics of the way a healthcare system works and thus offers a logical approach for improving the system. One such improvement explored is the implementation of a feedback control scheme based on a cumulative-sum control chart to detect small changes in mean performance measures that exhibit high levels of variability. Research provides data based evidence to conclude that simulation is useful for exploring the quantitative impacts of management decisions within a healthcare system that exhibits complex adaptive behaviours.