Optimizing operations in cancer care

Cancer care providers are navigating an increasingly complex landscape, marked by rising global cancer incidence, and rapidly evolving treatments. Healthcare systems are facing an array of challenges as they strive to adapt to these changes and deliver effective care.

At the core of these challenges lies the intricate task of patient scheduling. Cancer treatments typically require a series of coordinated appointments across various departments—radiology, chemotherapy, surgery, radiation therapy, and follow-up consultations, to name a few. The efficiency of this scheduling process is critical to the overall performance of the health system; both financially and clinically. It directly influences patient experience through minimized wait times and tailored scheduling to meet individual preferences. For staff, it ensures balanced workloads, reducing burnout and improving job satisfaction. Financially, effective scheduling optimizes patient throughput and resource utilization, which are key determinants of a center's economic stability and performance.

This use case depicts how the CHUM, one of Canada's leading cancer centers, leveraged AI-powered software to streamline their operations, reduce costs, improve access to care, and improve staff experience.

The problem

Like many cancer centers, the CHUM is facing major challenges: amidst a labor shortage, treatments are becoming more complex, and patient volumes are increasing. The center operates at capacity offering no margins for the increasing volumes. In radiation oncology as well as in the infusion center, coordinating and scheduling care trajectories has become a labor-intensive, highly manual process, burdening staff that are already overworked.

The calendar functionality within legacy software is used to schedule patients but the process is not automated nor optimized, remaining highly time-consuming and inefficient. It requires a human to manually select slots while mentally taking into account a set of sometimes ambiguous clinical rules. These rules are mostly tacit, within the mind of the booking clerks and knowledge is at risk of being lost as staff turnover.

Care plan functionality within legacy software is used but not in a way that removes the ambiguity from the physician’s request (who frequently amend the care plans on a case-by-case basis). Translating treatment requests into a set of appointments remains highly manual and requires several back-and-forth communications between clerks and physicians. In the best case, it wastes time, and in the worst case, it leads to scheduling errors.

In medical oncology,
The creation of the “daily roster” is a complex and manual process performed by highly qualified clerks. It takes hours and does not allow the flexibility to adjust to the last-minute cancellations. As a consequence, the workload is unbalanced across the day and among nurses. Overtime is frequent.
The process of selecting a new patient’s treatment start date also causes inefficiencies. As senior clerks turnover, newly hired clerks struggle with the complexity of these tasks, frequently seeking support. An assistant nurse is mobilized full-time to support the team. Errors occur, and patients are not scheduled or show up without being fully prepared to receive treatment.

In radiation oncology,
For each patient, the clerk manually selects the set of appointments, their duration and their priority, while managing the availability of required resources and treatment due date. The ambiguity of requests requires clerks to perform multiple back-and-forth communications.
The department operates at maximum capacity while patient delays are increasing. In this context, holding “emergency slots” is not possible. Last-minute urgent patients put the entire schedule at risk and often lead to overtime. It creates huge stress for the scheduling team, unable to easily identify scheduled patients that can be postponed to make space for more urgent ones. For this reason, lower-priority patients are typically scheduled last and can have lengthy delays.

The objectives

The CHUM deploys GrayOS, an AI-powered software aimed at streamlining operations in cancer care, in both its medical oncology and radiation oncology departments with the following objectives:

  • Reduce inefficiencies
  • Maximize capacity
  • Reduce administrative burden
  • Reduce wait times
  • Standardize practice
The CHUM implements GrayOS, a software that automates and orchestrates the logistics of cancer care operations. The software leverages decades of operations research and predictive analytics to optimize the scheduling process across radiation therapy and infusion therapy. It has 3 key functionalities:

  • Multi-disciplinary care trajectory templates
  • Automated patient scheduling
  • Predictive analytics and dashboards

The platform is used by scheduling clerks, head nurses, and oncology directors to perform both operational tasks (e.g. patient scheduling) and strategic tasks (e.g. resource planning or the monitoring of care pathways). Due to the tool’s user-friendly interface and major time-savings, GrayOS reaches 100% adoption in just a few weeks even though the clerks are still able to continue using the previous system.

Key results

$500k saved each year

80% reduction of administrative burden

50% reduction of time required to train new clerks

+11h of additional capacity unlocked every day

Through a 5% efficiency gain, the tool unlocks additional treatment capacity (+11 hours/day) despite the center already operating at high resource utilization rates, resulting in major cost savings ($500k /year).

Scheduling patients, which used to require several days of work from the most senior clerks, is now performed 80% faster.

The RadOnc department operates at maximum capacity, and machine idle time is virtually nonexistent. However, GrayOS’ improves management of patients priorities and avoides up to 1000 days of patient delays.

Time required to train new clerks is reduced by 50% thanks to the standardization of care plans.

Being deployed in both departments, the coordination between the two units is streamlined (specifically for concomitant treatments).

Nurse workload balance is optimized, resulting in an overall increase in staff satisfaction across booking clerks, therapists, nurses and administrators. 

Department managers report that they are able to better anticipate and adjust their resources.


IT & Software, Innovation, Operations
Better Customer Experience, Better Quality, More Efficient, Saving Cost
Optimization, Prediction