Patient scheduling based on a service-time prediction model: a data-driven study for a radiotherapy center

Area of Science:

• Oncology
• Operations Research
• Health Informatics

Background:

• Increasing patient demand strains healthcare systems, particularly for cancer care.
• Inefficient patient management, not resource scarcity, exacerbates long queues.
• Cancer treatment scheduling requires optimized patient flow for timely care.

Purpose of the Study:

• To develop a data-driven patient appointment scheduling model for a cancer center.
• To predict radiotherapy treatment duration accurately using data mining and regression.
• To improve daily patient throughput and reduce waiting times and technologist overtime.

Main Methods:

• Collaborated with Centre Intégré de Cancérologie de Laval (CICL).
• Employed a non-traditional approach to patient appointment scheduling.
• Utilized data mining and regression techniques, including classification and regression trees, to build a predictive model for treatment duration.

Main Results:

• Developed a predictive model for radiotherapy treatment duration with 84% accuracy.
• Designed optimized workday divisions based on predicted treatment durations.
• Evaluated new scheduling strategies using various patient sequencing rules.

Conclusions:

• The data-driven scheduling approach significantly enhances daily patient capacity, treating 40 additional patients daily.
• Substantial improvements observed in patient waiting times and reduction in technologist overtime.
• Highlights the potential of predictive modeling in optimizing cancer care logistics.