Two-Tier Hospital System Redesign Optimization Based on Consumer Choice and Queueing Network Modeling(CEIBS Working Paper, No. 036/2020/POM, 2020)

The hospital system in many countries experiences capacity shortage, which results in poor access to healthcare in certain parts of the country. Moreover, rapid urbanization, which creates many new population centers, has exacerbated the imbalance between care need and service capacity. One way to address the challenge is to conduct capacity expansion and spatial redistribution. In this research, we studied the problem of optimal decision making for locating new hospitals in a two-tier hospital system comprising both central and district hospitals, and upgrading existing district hospitals to central hospitals, with incorporation of patient preferences on seeking care. We first formulated the problem with a discrete location optimization model to minimize the total cost (i.e., a weighted sum of travel cost, waiting cost, and government spending). Then we constructed a multinomial logit model with real-world data to characterize hospital choice behaviors, and quantify patient arrival rates at each hospital accordingly. We also developed a multi-hospital queueing network model to analyze the impact of hospital locations on patient flows. By solving the resultant nonlinear combinatorial optimization problem via a genetic algorithm, we verified the effectiveness of hospital location reconfiguration and confirmed the influence of individual-specific attributes (e.g., insurance type and balking tendency).