An Australasian assessment of the basic treatment equivalent model derived from New South Wales data

The current method of assessment of radiation oncology linear accelerator throughput is either by patients per unit time or fields per unit time. This, however, does not take into consideration the complexity of different treatment techniques or of casemix. A model has been developed in an earlier study, called 'basic treatment equivalent' (BTE), to measure patient throughput of a linear accelerator, which includes consideration of the complexity of treatment techniques. The present study compared the BTE model with the current best measure of patient throughput of fields per hour. All 37 departments in Australia and New Zealand were invited to participate in testing the model, and 36 agreed to participate. The study period for each department was a consecutive 4 weeks between August and December, 1996. The prospective data collected were the total BTE units treated per linear accelerator per day, the total number of patients and fields treated per linear accelerator per day, and the total linear accelerator hours of operation per day excluding calibration time and significant breaks of linear accelerator time such as planned meal breaks. The treatment breaks between consecutive treatment fractions were not excluded from the linear accelerator treatment time. The throughput data for 36 departments (92 linear accelerators) were collected over the 4-week study period. The average throughput for the departments was 10.8 fields per hour and 4.2 patients per hour. The average BTE per department was 5.7 BTE per hour. The average BTE per episode per department was 1.38. The BTE model was found to be a more sensitive measure of productivity compared with fields per hour (P < 0.001). Some treatment techniques were thought to be not well represented by the BTE formula, particularly those techniques where junctions were present. The BTE model is a more sensitive measure than fields per hour and better reflects the variations in complexity in techniques. Despite this result there is further refinement to be performed to make the model even more sensitive.