Optimizing healthcare queues: a case study on chronic respiratory illness

Abstract Accepted by: M. Zied BABAI This study employs a data-driven approach to assess the evolving resource needs of chronic obstructive pulmonary disease (COPD) patients, exploring the impact on the hospital system. It integrates segmentation, operational queuing theory and parameter recovery from incomplete data to overcome limitations in fine-grained data availability, yielding operational insights using only administrative data. Initiating with a population clustering from granular data, the paper utilizes a multi-class $M/M/c$ model, extracting parameters through parameterization and Wasserstein distance. This model facilitates an informative analysis of the queuing system and population needs through various what-if scenarios. The comprehensive analyses encompass all patient arrival types, revealing that addressing the impact of COPD patients on the system necessitates more than just expanding capacity. Our work demonstrates the potential for specific improvement in clinical performance in respect of COPD patients.