Predicting Patient No-Shows Using Machine Learning to Optimize Clinic Scheduling

Authors

Lucas De Smet, Marie Dubois, Johan Vermeulen, Sofie Claes, Thomas Lambert Hasan KU Leuven, Belgium; Ghent University, Belgium; Université catholique de Louvain, Belgium; Vrije Universiteit Brussel, Belgium; University of Liège, Belgium. Author

Patient no-shows for scheduled appointments represent a critical operational and financial challenge for healthcare

providers, leading to wasted resources, reduced access to care, and increased wait times. This paper investigates the

application of machine learning (ML) to predict the risk of a patient missing their appointment. By analyzing

historical electronic health record (EHR) and scheduling data, predictive models can identify patients with a high

probability of no-show, enabling clinics to implement targeted interventions such as reminder calls, overbooking

strategies, or offering slots to waitlisted patients. This research details the development of several classification

algorithms, including Logistic Regression, Random Forest, and Gradient Boosting, to forecast no-show events. The

results demonstrate that ensemble methods can accurately identify high-risk patients, providing a data-driven

foundation for improving clinic efficiency, maximizing resource utilization, and enhancing patient access to care