953. Appointment Scheduling across Patient Groups with Varied No-Show Rates
Invited abstract in session TA-13: Appointment scheduling, stream OR in Healthcare (ORAHS).
Tuesday, 8:30-10:00Room: Clarendon SR 1.01
Authors (first author is the speaker)
| 1. | Qihao Wu
|
| Department of of Data and Systems Engineering, The University of Hong Kong | |
| 2. | Yong-Hong Kuo
|
| Department of of Data and Systems Engineering, The University of Hong Kong | |
| 3. | Haolin Feng
|
| School of Business, Sun Yat-sen University | |
| 4. | Yiwu Jia
|
| School of Management, North Sichuan Medical College |
Abstract
The disparity of equity in access to healthcare services could be attributed to scheduling preferences for avoiding appointment no-shows, which are found to be relevant to ethnic, racial, and socioeconomic characteristics. Our study aims to design appointment scheduling (AS) schemes that address equity issues and evaluate how patient-group information affects both schedule efficiency and equity in access to medical appointment services (i.e., fairness). In our problem, appointment requests come in gradually, requiring the scheduler to account for stochasticity and patient group-dependent no-show rates. We propose AS schemes that come with different levels of procedural fairness, and each proposed scheme leads to a Markov Decision Process (MDP). By deploying dynamic programming for a manageable-scale problem and reinforcement learning for the large-scale problem in practice, our experiments reveal consistent managerial insights. The Group-Aware Scheme (GAS) that fully leverages patient category information achieves the highest utility. The Semi-Aware Scheme (SAS), which utilizes only the group information from the existing schedule when making decisions regarding upcoming requests without needing the group identity of those requests, performs exceptionally well under specific operational environments characterized by patients’ show-up rates. This finding suggests that SAS can offer a competitive alternative to GAS and simultaneously achieve efficiency and perceived fairness.
Keywords
- Health Care
- Scheduling
- Programming, Dynamic
Status: accepted
Back to the list of papers