A data-driven approach to solve the RT scheduling problem.

Area of Science:

• Oncology
• Medical Physics
• Health Informatics

Background:

• Radiotherapy (RT) demand is increasing, presenting complex scheduling challenges due to its time-critical and interdependent workflow.
• Scheduling delays are a prevalent global issue in RT departments, impacting patient care and operational efficiency.

Purpose of the Study:

• To develop and evaluate an automated strategy for generating patient allocation lists to improve radiotherapy scheduling.
• To assist scheduling staff in creating a more efficient and timely radiotherapy treatment process.

Main Methods:

• Utilized historical data from a large Swedish RT department (2022) with 11-13 linear accelerators.
• Developed a C# algorithm incorporating patient-specific factors: timeline, booking category, diagnosis, and treatment intent for individual priority assignment.

Main Results:

• The algorithm generated a prioritized weekly patient scheduling list.
• Compared to manual allocation, the algorithm reduced patient delays by 10% and decreased average delays by 2 weeks.
• Algorithm-driven prioritization improved the balance across different diagnosis groups.

Conclusions:

• Automated algorithms can significantly reduce scheduling time for RT staff by rapidly assessing patient profiles.
• Implementing such algorithms can accelerate RT scheduling decisions and enhance overall departmental performance, potentially avoiding major organizational changes.