4 Presentations

• 
  10:30 AM - 10:55 AM

  Improving Access to Stroke Prevention Consults through Enhanced Capacity Planning and Patient Appointment Scheduling

  • Shahryar Moradi, presenter, University of Ottawa
  • Antoine Saure, University of Ottawa
  • Jonathan Patrick, Telfer School of Management, University of Ottawa

  Motivated by the problem faced by The Ottawa Hospital (TOH), this presentation focuses on the study of patient appointment scheduling practices at a Stroke Prevention Clinic (SPC). A patient who is referred to a SPC is typically scheduled for an initial consult with a neurologist. Prior to that initial consult several tests, depending on the patient's condition, may need to be performed. Some SPCs face several challenges in reducing patient wait times for consults and in having all test results available before consultation. To address these challenges, we propose a dynamic multi-priority, multi-resource appointment scheduling model which we approximately solve using Approximate Dynamic Programming (ADP) techniques. The main purpose of this model is to identify good policies for allocating available testing and consultation capacity to incoming patients, while reducing patient wait times and increasing the number of test results available before consultation in a cost-effective manner. The potential benefits from the proposed approach are evaluated through simulation for a practical example based on data provided by TOH. We also investigate the quality and practical implications of the resulting appointment scheduling policy by comparing its performance to that of other heuristic policies available in the literature.
• 
  10:55 AM - 11:20 AM

  Dynamic ambulatory care appointment scheduling

  • Amirhossein Moosavi, presenter,
  • Onur Ozturk, Telfer School of Management, University of Ottawa
  • Jonathan Patrick, Telfer School of Management, University of Ottawa

  We investigate an ambulatory care scheduling problem derived from a real case in Canada that offers multi-appointment multi-class multi-priority treatments in geographically distributed campuses with multiple resources. We consider a dynamic setting with uncertain patient arrival and use of the emergency department. This problem is formulated as an infinite-horizon Markov decision process model. We hybridize this model with a neural network to simplify feasibility constraints while respecting all assumptions. Due to the curse of dimensionality, we use an affine approximation architecture to approximate the value function. Then, an equivalent linear programming model is solved through column generation in order to compute approximate optimal policies. Simulation results demonstrate that the approximate optimal policy considerably outperforms well-known incumbent scheduling policies. Finally, we demonstrate that the application of our methodology can enhance performance metrics in a large ambulatory care center in Canada. A static scheduling rule may not be (near-) optimal for real-world ambulatory care scheduling problems. For example, a myopic scheduling rule can result in high resource utilization but poor scheduling decisions. An approximate optimal policy, however, equips a booking clerk with intelligent scheduling rules that are difficult for her to predict in real-time and work well in comparison to scheduling templates.
• 
  11:20 AM - 11:45 AM

  An Adaptive Scheduling Policy for Hospital Elective Patient Admissions During Pandemics

  • Peyman Varshoei, presenter, University of Ottawa, Telfer School of Management
  • Jonathan Patrick, Telfer School of Management, University of Ottawa
  • Onur Ozturk, Telfer School of Management, University of Ottawa

  The COVID-19 pandemic has presented hospitals with the challenge of managing the sudden increase in demand while continuing to provide essential services to patients. This research focuses on developing an adaptive scheduling policy to optimize patient throughput while also ensuring that the hospital can quickly vacate a predetermined number of beds within a short notice period (e.g. 5 days) during a pandemic. This flexibility is referred to as "nimbleness". We proposed a heuristic approach that utilizes two mixed-integer linear programming models and a simulation model. The first model generates an initial schedule for patient admissions while maximizing throughput. The second model considers uncertainty in emergency arrivals – including those related to the pandemic – as well as patient length of stay, and maximizes scheduled admissions while preserving nimbleness. A simulation model is developed to simulate daily random arrivals and discharges. These models are integrated through an automated feedback loop until the heuristic approach achieves a solution. The numerical results show that the approach is effective in maintaining high throughput while responding to pandemic surges, without having to reserve capacity for pandemic patients or experiencing significant cancellations.
• 
  11:45 AM - 12:10 PM

  A Branch-and-Price Approach to Intraday Scheduling of Chemotherapy Patients

  • Gustavo Angulo, Pontificia Universidad Católica de Chile
  • Alejandro Cataldo, Universidad Católica de Chile
  • Alejandro Cifuentes, Pontificia Universidad Católica de Chile
  • Pablo A. Rey, presenter, Universidad Tecnológica Metropolitana
  • Antoine Sauré, University of Ottawa

  Chemotherapy scheduling at cancer treatment centres is a complex problem due to high and growing demand, diversity of treatment protocols, limitations on resources and the need to coordinate treatment session times with laboratory preparation of medication. Over a given planning horizon, treatment centres assign patients first to specific days (interday scheduling) and then to specific times within each day (intraday scheduling), the latter process including the definition of medication preparation time. In this presentation, we address the intraday scheduling problem using an integer programming model that attempts to schedule all patients assigned to the horizon, and the preparation of the medication to be administered, simultaneously. The model, which is based on treatment patterns, is solved using a branch-and-price algorithm. The proposed approach allows for medication preparation on the day of treatment or a previous day subject to time slot availability. A case study is conducted using actual data from a Chilean cancer centre to compare through simulation the schedules generated by the proposed approach and the centre's manual method.