Women’s College Hospital Uses Operations Research to Create an Ambulatory Clinic Schedule

Brendan Eagen, Timothy C. Y. Chan, Michael CarterService 2018 · DOI

Abstract: Women’s College Hospital (WCH) in Toronto, Canada offers roughly 300 outpatient clinics every week. In this article, the authors describe a project started in April 2011 with WCH to design a new schedule for their clinics to accommodate a move to a new hospital building, which was completed in May 2013. They developed an integer programming model to optimize the assignment of clinics to timeslots and locations, based on the desire to minimize changes from the historical schedule. In cooperation with senior leadership of WCH, the authors tested multiple scenarios that explored changes to space utilization policies at WCH and ultimately generated a new clinic schedule, which WCH implemented in May 2013. In this paper the authors highlight the value the work has created for WCH and present lessons learned in development of the model and through collaboration with the WCH team.

Introduction

Women’s College Hospital (WCH) is a premier ambulatory care facility in Toronto. Historically, its scheduling was decentralized, with individual departments managing their own spaces and timeslots. This "organic" growth led to a fragmented system with no central oversight. When the hospital planned a move to a new facility in 2013, it faced a significant challenge: the new building would have a 22% smaller footprint for examination rooms. To manage this transition, WCH turned to Operations Research (OR) to design a centralized, high-efficiency scheduling system.

The Scheduling Challenge

The move necessitated a shift from department-owned rooms to a "hoteling" model, where rooms are shared across specialties rather than assigned to specific departments. The primary objective was to fit approximately 300 weekly clinics into the new, smaller space while minimizing disruption to clinicians. Specifically, the hospital leadership wanted to maintain clinics in their historical timeslots and on preferred floors whenever possible to ensure continuity of care and staff satisfaction.

Methodology: Integer Programming

The researchers developed an Integer Programming (IP) model to optimize clinic assignments. The objective function was designed to minimize "penalties" associated with changes. Moving a clinic from its historical timeslot carried a high penalty (weighted 4x), while moving it to a non-preferred floor carried a lower penalty (weighted 1x).

Hard constraints ensured the feasibility of the schedule:

  1. Clinician Availability: A clinician could not be scheduled in two places simultaneously.
  2. Floor Capacity: Total room capacity per floor was never exceeded.
  3. Spatial Continuity: All rooms required for a single clinic session had to be located on the same floor.

The model, implemented in AMPL and solved using CPLEX 12.1, typically reached an optimal solution in under 30 minutes. This speed allowed leadership to run multiple "what-if" scenarios, such as evaluating the impact of converting administrative offices or consult rooms into exam rooms.

Data and Spatial Constraints

The study relied on detailed primary data collected in 2011. Initially, the team identified 278 clinics across seven major groups, involving 105 clinicians. The distribution of these clinics and their requirements is shown in Table 1.

Table 1: Initial clinic data by specialty group, showing the distribution of exam rooms and clinicians.

The new building's capacity was strictly defined across three floors, totaling 84 exam rooms and 20 consult rooms. The distribution of these spaces, which formed the core constraints of the IP model, is detailed in Table 2.

Table 2: Spatial capacity of the new hospital building, detailing exam rooms, consult rooms, and offices per floor.

Key Results and Insights

The implementation of the optimized schedule (specifically "Scenario 2B") yielded impressive quantitative improvements:

  • Space Efficiency: The hospital successfully transitioned to a facility with at least 22% less examination space.
  • Utilization Boost: Average exam room utilization rose from 76% to 97%.
  • Workload Balancing: The standard deviation of utilization across the week dropped from 20% to 4%, indicating a much more balanced and predictable workload.
  • Off-Peak Recovery: Utilization on Friday afternoons—historically a "dead" time—jumped from 32% to 88%.

Beyond the numbers, the project highlighted several critical "soft" factors. The use of a mathematical model acted as a "political buffer," providing a transparent, rule-based rationale for assignments that removed perceptions of internal politics. Furthermore, the project underscored the importance of data validation; the team discovered many "hidden" clinics—meetings occurring in private offices that clinicians didn't formally recognize as clinics—which were essential to include for accurate space planning.

Conclusion

By applying Integer Programming, WCH transformed its scheduling from a decentralized, "organic" process into a centralized, data-driven system. This transition not only accommodated a smaller physical footprint but also significantly enhanced the efficiency and predictability of hospital operations. The success of the project was largely attributed to the "hoteling" policy and a dedicated hospital liaison who bridged the gap between clinical workflows and mathematical modeling.

Summary updated

Abstract

Notes. Women’s College Hospital (WCH) in Toronto, Canada offers roughly 300 outpatient clinics every week. In this article, the authors describe a project started in April 2011 with WCH to design a new schedule for their clinics to accommodate a move to a new hospital building, which was completed in May 2013. They developed an integer programming model to optimize the assignment of clinics to timeslots and locations, based on the desire to minimize changes from the historical schedule. In cooperation with senior leadership of WCH, the authors tested multiple scenarios that explored changes to space utilization policies at WCH and ultimately generated a new clinic schedule, which WCH implemented in May 2013. In this paper the authors highlight the value the work has created for WCH and present lessons learned in development of the model and through collaboration with the WCH team.
BibTeX
@inproceedings{eagen2018womens,
  author = {Brendan Eagen and Timothy C. Y. Chan and Michael Carter},
  title = {{Women’s College Hospital Uses Operations Research to Create an Ambulatory Clinic Schedule}},
  booktitle = {Service},
  year = {2018},
  doi = {10.1287/serv.2018.0221},
}
Figure