2016 Conference Presentation
Objective: Present a computational support system for decision making in management of chronic patients.
Methods: The new system uses a simulation tool based on the Jason multi-agent systems platform to model the chronic care strategy in an entire health department –department 11- in the Valencian Region (Spain). Multi-agent systems allow the tool to include human relationships, preferences and social abilities that take place between chronic patients and healthcare system professionals.
To design the computational system it is necessary to analyze the interactions of patients within the healthcare system (with the possibility of including social resources), based on clinical variables, variables on the use of the system (emergency, previous hospital admissions...) and non-professional support.
The system includes three kinds of agents: patients, healthcare system professionals and a case management team (CMT). Patients are people over 65 years old with one or more of the following conditions: heart failure, COPD, diabetes and hypertension. The patient’s personal and medical information will be used by professionals to screen the patient and decide whether she/he is under the CMT or not. Patients not eligible are sent to the conventional healthcare system and, thus, they do not enter the simulated system. When they are selected, though, the CMT acts as the committee in charge of designing a pathway for each of these patients. The system is able to simulate these decisions and successive interactions of patients within the healthcare system.
Results: The computational model to be used (i) to refer patients and (ii) to assess the effectiveness of the pathways for specific patients. Examples of indicators that can be geographically represented (in the chosen cartographic unit) are the percentage of patients who require a specific resource (prevent the collapse of services), their age distribution or the total number of referred patients between services. In the pilot the system can simulate a population of up to ten thousand agents in a few seconds (1.13 sec), and a population of fifty thousand agents in around five minutes.
Policy implications: The impossibility of testing a LTC model in large scale, and the wide range and complexity of the variables it manages, makes it necessary the use of ICT solutions that can generate virtual scenarios and analyze their repercussions in health and social policy at both local and regional scale. Therefore, the outputs of the simulator will provide policy makers with the long-term effects of different policies on the population considered, as well as on the health system. And, an overall evaluation of the social and economic sustainability of changes in healthcare policies (improving the coordination between health and social system) and effects of these policies on the citizen’s quality of life. The system will enable stakeholders to discuss the issue at hand and shape LTC policies to be applied in different welfare contexts.