Detection of systematic errors in external radiotherapy before treatment delivery

The execution of an independent control of monitor units (MU) and dose distribution calculation, together with a check of the data reported in the treatment chart is an effective tool in strongly reducing the occurrence of systematic errors before treatment delivery. In this paper we report the results of the analysis of 6272 controls (about 5000 patients) registered over more than 5 years; 70 serious errors (producing a deviation larger than 5% from the prescribed daily dose) and 147 minor errors were detected and corrected before the start of the treatment. The error rate was found to be strongly operator-dependent (serious error rate ranging from 0.3 to 2.5% when considering different operators). A time-trend analysis showed a significant reduction of serious errors, i.e. 1.5% in the period from September 1991 to April 1994 compared to 0.9% in the period from April 1994 to November 1996. However, even if the double check was highly effective in revealing human errors, three serious systematic errors (errors occurring during the calculation/planning/transcription phases) escaped the control and were detected by diode in vivo dosimetry during the period October 1994 to November 1996 (in 650 patients controlled). (C) 1997 Elsevier Science Ireland Ltd.