Analysing workflows in radiation oncology: A process mining approach to identify treatment process challenges

Background:

Process mining (PM) is a powerful approach for analysing and optimising complex workflows. Radiation therapy (RT) involves multiple steps and resources, making it well suited for PM analysis.

This study aims to apply a PM approach to characterise the real-world workflow of a high-volume RT department, with the goal of identifying critical transitions and bottlenecks, quantifying their impact on treatment timelines, and exploring the main factors associated with treatment suspensions and cancellations.

Material and Methods:

Patients treated with RT at our centre between January 2017 and December 2021 were included. All patient-related events were extracted from the institutional database. A first-order Markov model was used to analyse event sequences and identify anomalies, while the Kruskal–Wallis test compared median completion times across stratified sub-cohorts.

Results:

The study analysed 43,183 events associated with 8608 treatments. The pathway diagram depicted key events, such as Prescription , Scheduling , CT simulation , RT start , RT cancelled , RT suspension , and RT end . Statistical analysis revealed a significant difference in terms of median time in case of suspension event during the path, leading to a delay in the start of RT treatment (p<0.05). The analysis of the employed time from Prescription to RT start demonstrated a significant impact of suspension on the time interval for breast, genitourinary, gastrointestinal, metastatic, head and neck, gynaecological, thoracic, and skin cancer (p<0.05).

Conclusions:

This study illustrates the applicability of PM as a methodological framework for analysing care pathways. By demonstrating how PM can identify delays, interruptions, and workflow inefficiencies, it highlights its potential to support process evaluation and optimisation.