Chronic care management in times of capacity shortages: an integrated patient assignment and treatment scheduling problem for post-disaster hemodialysis planning

Abstract

Individuals with chronic kidney diseases rely on hemodialysis for survival and face significant vulnerabilities and healthcare challenges during and after disasters. Ensuring continuity of treatments and access to care for these patients becomes difficult due to a substantial reduction in hemodialysis capacity in the affected region, often caused by damage to facilities and infrastructure. In collaboration with the Renal Disaster Relief Task Force (RDRTF), a key actor in coordinating emergency response efforts for these patients, we aim to support post-disaster decision-making regarding patient assignments to available hemodialysis centers and scheduling of patients’ treatments, considering alternative treatment options with different durations and duration-induced treatment frequency requirements. We present an integer programming model for the proposed integrated patient assignment and treatment scheduling problem to maximize the number of patients served in affected regions while minimizing reliance on shortened treatments. To efficiently solve this problem, we develop two decomposition-based solution methods: a Logic-Based Benders Decomposition algorithm and an Iterative Constructive Heuristic. Furthermore, we demonstrate how these methods can be implemented within a rolling horizon framework to enable periodic updates and re-optimization as post-disaster conditions evolve. We present numerical results that demonstrate the benefits of the proposed system and solution algorithms through a case study focusing on earthquake response in Istanbul. Our findings across different disaster scenarios highlight the importance of providing alternative treatments and pooling hemodialysis capacity through centralized planning. Based on feedback from the RDRTF, we also present a prototype of a decision support system dedicated to renal disaster coordination.