Collaborative Research: Models and Algorithms for Family Medicine Residency Scheduling with Clinic First Objectives

This award will contribute to the advancement of national health and welfare by developing a comprehensive modeling and solution framework for generating annual rotation schedules for family medicine residents, in alignment with the objectives of the Clinic First model. The Clinic First model has emerged in academic family medicine as an innovative educational approach that prioritizes ambulatory (out-of-hospital) training and continuity of care while incorporating physicians' specialized interests. The United States faces an acute shortage of primary care physicians (PCPs), a trend projected to worsen over the next decade, potentially jeopardizing the effectiveness of the healthcare system. The shortage is exacerbated by many family medicine residents opting for sub-specialization or hospitalist roles instead of becoming family medicine physicians. A major contributing factor to this trend is the prevailing emphasis within family medicine residency training programs on hospital-based experiences rather than outpatient primary care. This project is focused on creating specialized methods for designing annual rotation schedules that support the Clinic First model, with the goal of better preparing residents and improving retention rates in family medicine. These schedules are designed to enhance residents' clinic experiences, motivating them to pursue PCP careers and mitigating the PCP shortage nationwide. This research aims to develop innovative analytical methods to enhance residents' clinic experiences through a new decision-support framework. At the core of this framework is a novel Dantzig-Wolfe formulation designed to create annual rotation schedules that optimize two objectives consistent with the Clinic First model. Methodologically, the project seeks to advance the solution of large-scale bi-objective integer programs using column generation-based decomposition algorithms by introducing new branching and cutting plane strategies to improve computational efficiency. Additionally, new methods based on Branch-and-Price will be developed to efficiently identify a diverse set of such solutions for large-scale bi-objective integer programs. The research will also contribute new solution algorithms and insights into Parametric Integer Linear Programs. The benefits of the new schedules will be analyzed and compared with those created manually at a partnering family medicine residency program. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.