Analysis methods for performance & availability in critical care medicine

Operations of critical care departments in health systems are increasingly reliant on the availability of interoperable medical devices. Many large health care systems have fully transitioned in recent years to uniform electronic health record platforms, introducing the need for a computing infrastructure at each point of care. As demand on national health infrastructures increases, the management of finite resources such as capacity to treat in a ward requires thorough consideration. We produce a model for evaluating technology-reliant intensive care units (ICU) using the extensive Stochastic Reward Net formalisms. The technology-reliant model evaluates ICU performance in the context of reliance on fallible computer systems and medical devices. Technical enhancements to the model specification are discussed in relation to the problem context. The final structured model is contrasted to the currently employed queuing models, and solved using discrete event simulation. We then demonstrate the model's ability to quantify several planning decision trade-offs in capacity, the probability of meeting varying service demands, and estimating the utilization of medical devices. These results motivate continued development of the outlined approach, as well as discussion of the technical feasibility of SRN solution in relatively large systems of practical interest.