This chapter examines the paradigm that a health care system’s behavior may be examined using an agent simulation in order to illuminate its macroscopic characteristics and the effects of policy on its over all operation. Further, if the individual components are well articulated, the component behavior may be also studied. Health care systems in North America are generally regulated by various processes and mechanisms in order to provide orderly access to, and control of, the health care system. While all processes are designed to be fair and equitable, in many ways the system can not be examined or optimized because the risk, that making changes to the system might result in degraded services, is too great to permit making even simple changes. In this context we propose the development of a health care system model in which agents mimic the behavior of the key components of the system. These components interact and engage each other in a manor analogous to the operation of the health care system. The formulation of such a system is, by its very nature, an extremely complex process, and necessitates development in components or units. In this chapter we present the first components of such a system. Each component has unique and complex behaviors. These components will, with additional development, form the basic structure of a health care system model. Specifically we present results from the development of a diabetic patient agent model, the development of an agent-based neurosurgery ward bed allocation system, and the development of an agent-based scheduling system that may be used to allocate resources within the health care system.
TopIntroduction
The modeling of health care actors, components and systems in order to develop a complete understanding of component interactions and system reaction is one of the more challenging simulation and modeling problems for software agent systems (Moreno, Isern & Sanchez, 2001). There are many benefits to this approach, which may be summarized by simply stating that through development of a good understanding of the mechanics of the system, optimal, or more realistically closer to optimal, operation of the system may be possible, with reduced cost, greater efficiency, reduced pain and suffering, and a healthier population. For example, better allocation of resources in health care management may be possible by modeling each component in the system with the goal to clarify inefficiencies in the system. By testing and evaluating the various scheduling techniques and policies used, enhancing patient outcomes are possible. A second example may be the modeling of disease in the human population. If the evolution of disease can be modeled, the expected impacts on the health care system over time and as population’s move, age, etc. may be come more predictable, allowing for better planning.
We propose to build a working model of the healthcare system, which will accurately depict its current operation and function, and in addition will enable physicians, hospital administrators, and government officials to test various new policies to characterize how effectively they deploy the available resources or to see if the implementation of these policies will result in better patient outcomes. The goal being to move the health care system closer to an optimal operation so that, the use of hospital resources will be efficient, there will be a decrease in waiting periods for patients for services, there will be a reduction in the cost of delivering health care services, and to use existing systems will be at maximal efficiencies.
This type of work attempts to answer a number of important questions regarding the use of health care resources: First, can we construct a faithful macro-model of healthcare system behavior? Second, can we implement this model using distributed software and agent systems? Third, can this model be used to evaluate the present operation of the healthcare system particularly with respect to resources allocation? Fourth, can these models be used to propose new and improved ways to allocate scarce resources in order to improve efficiency, reduce costs, and most significantly reduce human suffering?
An agent model of the health care system requires the formulation of agents that will be programmed to mimic the behaviors of key components of the health care system. The process of applying health care policy to these agent components, which form the system model, is manifested in the behavior and function of the agents, and in the interactions between agents. Mechanisms to monitor and observe the behavior of the system will then allow us to examine the performance of the system.
In this work we propose the use of a uniquely symmetric modeling structure in which we identify four foci around which the model may be centric. These four foci are: (1) the patient, (2) the physician, (3) the nurse, and (4) the hospital infrastructure. Our structure embeds a high degree of connectivity between each of these foci, and in fact allows the model to be seen as centered about any one of these foci depending on which focus is specifically being tracked or analyzed. The foci themselves are typically more than a single agent and may represent be a set of agents with similar goals, objectives or expectation of the operation of the health care system,