CalMD: A Web Application for Automated Optimal Doctor Scheduling

Abstract

Currently, most medical residency programs in the U.S. create their annual rotation schedules through a manual, error-prone process that takes up to months to complete. These schedules are not only important for meeting hospital requirements, but also have a profound effect on doctors’ happiness. In this thesis, we provide a novel optimization model of this scheduling problem, and present a web application called CalMD that allows residency programs to automatically create utility-maximizing schedules. We evaluate CalMD using Barnes-Jewish Hospital’s 2015-2016 scheduling data, and assess the optimization model’s effectiveness in satisfying objective factors in schedule subproblems. We find that optimal solutions to 6- to 8-week subproblems maximize residents’ clinic frequency, couple matching, and vacation and elective request satisfaction, and can be found with little computation time. We also present doctors’ positive responses to CalMD, and describe potential avenues for future work.