2.2. Error classification

The occurred errors were documented in an Excel (Microsoft Office 2010, Microsoft Corporation, Seattle, WA) spreadsheet including the following data: date, short description of the error, anatomical treatment site, treatment technique, and type of treatment machine. After short feedback discussion with the planner errors were corrected and the treatment plan was run through the review process for a second time. These second review processes were not considered in this investigation but no errors/deviations were observed in this step. The number of reviews per month varied due to alterations in patient load and holiday periods. The investigated physics review herein is only a measure within the risk management strategy in our clinic. Other errors might occur earlier in the process, as in contouring, or later e.g. during patient set-up.

We investigated the kind of errors that occurred and analyzed if the occurrence correlated with type of treatment machine, radiation technique, or anatomical treatment site. Also, the frequency of errors was analyzed in regard to shortage of staff, e.g. during holiday periods and, furthermore, what kind of errors could be trapped in the TPS.

Errors were divided into two groups: g-errors and s-errors. g-errors have the potential to harm the patient, s-errors have not, but they can delay the treatment process or can result in inconsistencies in treatment reporting. This distinction between g- and s-errors was made in an interdisciplinary discussion in our clinic, in which all errors that occurred were considered. Ford et al. [4] followed a similar method. They established baselines for high and low severity incidents in a face-to-face meeting using the ANS score [8]. Also, Siochi et al. [9] made a distinction of error severity levels. It should be emphasized that no treatment failure or incidence resulted from the errors in this investigation or occurred during the time of this study in our clinic. Statistical comparisons were performed using Chi-square tests in Excel. Statistical significance was assumed with p < 0.05. All patients in this study gave informed consent for data processing for scientific purposes for our clinic.