Can automation in radiotherapy reduce costs?

BACKGROUND: Computerized automation is likely to play an increasingly important role in radiotherapy. The objective of this study was to report the results of the first part of a program to implement a model for economical evaluation based on micro-costing method. To test the efficacy of the model, the financial impact of the introduction of an automation tool was estimated. A single- and multi-center validation of the model by a prospective collection of data is planned as the second step of the program. MATERIAL AND METHODS: The model was implemented by using an interactive spreadsheet (Microsoft Excel, 2010). The variables to be included were identified across three components: productivity, staff, and equipment. To calculate staff requirements, the workflow of Gemelli ART center was mapped out and relevant workload measures were defined. Profit and loss, productivity and staffing were identified as significant outcomes. Results were presented in terms of earnings before interest and taxes (EBIT). Three different scenarios were hypothesized: baseline situation at Gemelli ART (scenario 1); reduction by 2 minutes of the average duration of treatment fractions (scenario 2); and increased incidence of advanced treatment modalities (scenario 3). By using the model, predicted EBIT values for each scenario were calculated across a period of eight years (from 2015 to 2022). RESULTS: For both scenarios 2 and 3 costs are expected to slightly increase as compared to baseline situation that is particularly due to a little increase in clinical personnel costs. However, in both cases EBIT values are more favorable than baseline situation (EBIT values: scenario 1, 27%, scenario 2, 30%, scenario 3, 28% of revenues). CONCLUSION: A model based on a micro-costing method was able to estimate the financial consequences of the introduction of an automation tool in our radiotherapy department. A prospective collection of data at Gemelli ART and in a consortium of centers is currently under way to prospectively validate the model.