Process mining to optimize palliative patient flow in a high-volume radiotherapy department.

ABSTRACT

INTRODUCTION: In radiotherapy, palliative patients are often suboptimal managed and patients experience long waiting times. Event-logs (recorded local files) of palliative patients, could provide a continuative decision-making system by means of shared guidelines to improve patient flow. Based on an event-log analysis, we aimed to accurately understand how to successively optimize patient flow in palliative care. METHODS: A process mining methodology was applied on palliative patient flow in a high-volume radiotherapy department. Five hundred palliative radiation treatment plans of patients with bone and brain metastases were included in the study, corresponding to 290 patients treated in our department in 2018. Event-logs and the relative attributes were extracted and organized. A process discovery algorithm was applied to describe the real process model, which produced the event-log. Finally, conformance checking was performed to analyze how the acquired event-log database works in a predefined theoretical process model. RESULTS: Based on the process discovery algorithm, 53 (10%) plans had a dose prescription of 8 Gy, 249 (49.8%) plans had a dose prescription of 20 Gy and 159 (31.8%) plans had a dose prescription of 30 Gy. The remaining 39 (7.8%) plans had different dose prescriptions. Considering a median value, conformance checking demonstrated that event-logs work in the theoretical model. CONCLUSIONS: The obtained results partially validate and support the palliative patient care guideline implemented in our department. Process mining can be used to provide new insights, which facilitate the improvement of existing palliative patient care flows.