An automated scheduling system for radiotherapy physicist on-call using Monte Carlo simulation

Abstract

An automated physicist on-call program was developed to support emergency radiotherapy in a cancer centre. A computer program was created to generate an on-call schedule according to the credit score approach. Monte Carlo method was used to simulate the number of treatment cases per shift of on-call physicists (total 32) based on 8 years of data (2010–2017), and the “Most Credit First” criteria was used to justify the order of physicists in the schedule. Evaluation of the old schedule, in which the physicists were randomly assigned, with the new one was carried out. The deviations of mean for the number of shifts and treatment cases for every physicist were determined between the new and old schedule. By considering the on-call physicists who contributed more than or equal to 10 shifts of treatment cases in 2010–2017, in the old schedule there were 6 physicists having shifts and treatment cases greater than 30% of the mean values. While in the new schedule, nobody has similar workloads over 30% of the mean during the same year range. Using the new scheduling method, the mean number of shifts was reduced from 16.5 to 11.8 per physicist, and the mean number of treatment cases was reduced from 25.6 to 19.7 per physicist, as compared to the old schedule. It is concluded that our new method based on Monte Carlo simulation and credit score approach can produce a more equitable physicist on-call schedule for a list of physicists in the emergency radiotherapy program. The workload balance using our new method is better than our old method that assigned physicists randomly.