Fair online hospital diagnostic service scheduling: Helping both patients and providers
Abstract
The management of diagnostic services is a challenging and important task for hospital systems. Suboptimal scheduling of diagnostic services increases patient wait times and reduces the availability of overnight beds that could otherwise be utilized by incoming patients. Patients are served by a limited number of diagnostic technologists, who frequently experience uneven workload distributions and disproportionate workload demands. We propose a framework for analyzing and addressing workload imbalances by developing algorithmic solutions to target contributing factors. In collaboration with a large academic medical center in the Northeastern US, we identify key factors related to total workload, load difficulty, perception of load fairness, and workload variability, and develop an online algorithm that optimizes daily diagnostic technologist schedules to mitigate these factors. Our proposed algorithm, Patient-Provider Load-Balance (PPLB), balances technologist workload considerations with patient and hospital needs, simultaneously incorporating considerations for patient urgency levels and characteristics. We provide analytical guarantees and analyze the algorithm’s performance using echocardiogram order data from our hospital partner. Computational studies show that our online load-balancing algorithm effectively mitigates all identified workload imbalance contributors by improving load fairness for technologists by up to 46%, limiting difficult scan load, and reducing daily workload variability compared to current hospital policy - all without compromising operational performance in terms of patient throughput and wait time. We also demonstrate PPLB’s overall superior performance across both patient and provider factors compared to several benchmarks. Our algorithm’s policy shortens patient wait time and improves throughput, especially for high demand scenarios.
