An Analysis of the Impact on Trends in Automation on Human Error Potential in Brachytherapy

Abstract

Technologies under development in the field of radiotherapy offer the possibility of automating many of the remaining functions that are currently carried out by radiotherapy staff. If adopted, they are likely to significantly change the roles of health care professionals and modify the potential error profile of specific radiotherapy procedures. In this paper, a combination of allocation of functions methods, including Levels of Automation (LOA) [1] and Meister's prescriptive method [2], and the Human Error Analysis and Reduction Technique (HEART) [3] were used as the basis for an evaluation of the impact of increasing automation on the potential for human error in Low-Dose Rate Prostate Brachytherapy (LDRPB). An IDEFØ model of the brachytherapy treatment process previously developed by the authors was the starting point for the analysis. The process “Perform Implant” which currently has sub-processes that are under direct human control and others that utilize advanced technology was selected for analysis. The LOA model was applied to “Perform Implant” for three scenarios: current set-up, mixed-automaton and full automation. A modified set of risk criteria specified by the Irish Health Service Executive (HSE) [4] were used in conjunction with Meister's method to select the most appropriate mix of automation for further analysis. HEART was then applied to both the current set-up scenario and the selected scenario and the outcomes were compared. The HEART analysis provided one justification for the selection of a particular human-automation mix, though caution should be exercised as HEART requires further validation in the context of healthcare systems.