Modern hospitals prioritize the needs and experiences of patients. Consequently, hospital design is becoming increasingly complex, focusing on integrating the diverse needs of patients with hospital functions. This complexity presents a significant challenge for architects in analyzing interactions between buildings and users. Currently, architects mainly rely on traditional and deductive methods, such as previous case studies or post-occupancy evaluations. However, these approaches can lead to uncertainties in performance, causing discrepancies between expected and actual outcomes.
In this context, although various pre-occupancy evaluation tools have been usefully employed for assessing hospital performance, they have limitations in systematically analyzing and evaluating the performance of hospital designs based on human factors and needs. To overcome these limitations, agent-based simulation has been proposed. This simulation, composed of autonomous and rational humanoid computer agents, or Virtual Users (VUsers), helps understand how hospital designs impact interactions between the environment and users.
However, most simulation models focus on modeling the economic aspects of hospitals related to medical staff and healthcare systems, rather than the individual experiences of patients and caregivers, the primary users of hospitals. Consequently, there is still a lack of tools for architects to evaluate hospital performance related to patient satisfaction.
In response, this study developed an agent behavior model that simulates the overall physical and social interactions that patients and caregivers can experience during outpatient visits. This model allows architects to analyze hospital performance related to patient satisfaction, such as waiting times and personal space interference.
The study developed three types of agent models, distinguished by their computable interactions. Agent Model [1] is a basic model that simulates the use of hospital facilities according to outpatient procedures and interactions with the hospital environment and medical staff. Agent Models [2] and [3] are advanced models based on the basic model, further reflecting the social elements among hospital visitors. These models calculate behaviors for maintaining patient-caregiver relationships and order among hospital visitors by observing and responding to the states of other agents. The difference between Models [2] and [3] lies in how they recognize other agents and their relationships. Model [2] computes visual perception based on tags, while Model [3] is based on behavior. These differences in mechanisms impact the observable behaviors of patients and caregivers and the overall hospital performance evaluation.
To compare the practical applicability and usefulness of the three developed models, an expert evaluation was conducted. Seven hospital design experts analyzed six hospital projects using the three simulation models and then evaluated each model in terms of (1) the reliability of hospital performance evaluation and analysis results, (2) the applicability and usefulness in practical design, (3) the usefulness in managing hospital infrastructure and medical services, (4) the reliability of patient-caregiver agent behavior, and (5) unexpected findings related to hospital performance regarding patient satisfaction.
The results, analyzed through one-way ANOVA, showed that the use of Agent Model [2] better facilitated reliable hospital performance evaluation and analysis, as well as practical applicability and usefulness, compared to the other two models. It also demonstrated reliability in measuring five aspects of hospital performance, including space utilization density, interference with personal and intimate areas, maintenance of patient-caregiver relationships, and maintenance of order among visitors.
This research will contribute to the development of agent-based simulation models that help architects effectively evaluate the overall performance of hospitals related to patient satisfaction and understand various patterns of social interactions within hospitals, including patient-caregiver relationships.