Capacity and Yield Management in Outpatient Clinics

• Main Author: Aras, Gokce Akin
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01.01.2014
• Subjects: Health care management; Industrial engineering; Operations research
• ISBN: 9781321582673; 1321582676

Outpatient clinics receive appointment requests from a variety of patient types who exhibit different cancellation, reschedule and no-show behaviors. It is important to provide timely appointments to these patients. In fact, our analysis reveals that the later the appointment date, the less likely the patient will be seen. Therefore it is imperative that we minimize delay in order to reduce patient loss and provide patient access. This is challenging and requires significant effort due to the number of uncertainties in the demand structure and appointment calendars in outpatient clinics. Given the increasing demand and limited capacity of outpatient clinics, it is particularly important to utilize capacity effectively and efficiently. Clinics strive to avoid appointment schedules that result in physician idle time or delay patient appointment dates due to improper capacity management. In this study, we consider outpatient clinics, in which there are multiple patient classes with different demand rates; revenues; and behavioral functions associated with reschedules, cancels, and no-shows, each of which depends on the appointment delay (i.e., the number of days between the appointment request and the actual appointment). To address this problem we first develop discrete-event simulation models to evaluate different appointment policies. The simulation models show that for some patient classes it is especially important to provide timely appointments in order to increase the net profit as well as the seen-patient percentages. We use these insights to develop mathematical programming models to identify the optimal appointment assignment policy for each patient class, so that the net revenue is maximized subject to the clinic’s daily capacity constraints. The initial optimization model assumes stationary demand and focuses on the stochasticity in the delay-based patient behaviors, i.e., cancellation, rescheduling and no-shows. We show that under certain conditions, this model reduces to a Multiple Choice Knapsack Problem. Using this structure of the problem, we derive optimal policy properties and show that the optimal policy is either to assign next-day appointments or the latest-allowed-day (based on maximum allowed delay) appointments to patients considering the available capacity. In the last part of this dissertation, we develop a time-varying optimization model, to incorporate time-varying demand and the timing of the reschedule requests. We also modify this model and obtain a restricted which allows only next day or latest-allowed day appointments. We compare the policies obtained from the stationary optimization model, the time-varying optimization model, and the restricted time-varying optimization model with the simulation. We observe that all the optimal policies perform significantly better than the current system in the outpatient clinic (that motivates this research), in terms of net profit and the seen-patient percentages. The time-varying optimization model solutions perform much better than the stationary optimization model solution, due to additional components, such as time-varying demand, that it takes into account.