Dynamic appointment scheduling for outpatient clinics with multiple physicians and patient choice

Mitigating the adverse effects of uncertainty in appointment systems, arising from heterogeneous patient needs and preferences, is critical to the effective use of scarce medical resources and patient satisfaction. This study addresses an online scheduling problem with multiple servers and consideration of patient preference for physicians and their appointment times. The receptionist immediately determines whether a request should be accommodated and offers an appointment time slot for each accepted patient request. The patient may reject an undesirable appointment time slot with a certain probability, or may accept it, but the no-show probability will be higher. A stochastic overbooking model is formulated to maximize the expected profit, which is defined as the revenue generated from accepted requests minus the cost incurred by patients waiting and physicians' overtime. A myopic scheduling policy is developed based on certain structural properties of the objective function. This study advances the study of appointment systems by generating a non-unimodal profit evolution. Moreover, both the decision of accommodating more requests for certain slots and the scheduling of appointments depend on the patient choice rather than the patient type. Further, computational experiments and analysis offer valuable insights into performance improvement in outpatient clinics.