Optimal physician assignment and patient demand allocation in an outpatient care network

Due to significant supply-demand imbalance in outpatient specialty care, many patients are either denied of access to care or imposed with serious travel burden. An important reason for poor access is the fact that many outpatient specialty consultation appointments are scheduled at central hospitals, which could be far away from patient residence. To improve access, physicians working for integrated multi-site care networks are asked to hold their consultation sessions in outreach clinics that are closer to patient׳s residence. We develop a new type of integer programming model, which is characterized as a multi-commodity traveling capacitated facility location problem, to determine travel assignments for physicians and appointment locations for patients. Given the large spatial variability in patient demand, solving the resultant integer program is computationally demanding. We thus propose several column generation based primal heuristic algorithms to construct physicians' assignments and develop several local search algorithms to further improve the assignments. We conduct case studies based on the mental health care delivery in Veteran Affairs Care Networks. Our numerical studies demonstrate the solution efficiency of the developed algorithms and the potential of achieving significant improvement in patient access to care.