PURPOSE: Estimating the risks of radiotherapy (RT) toxicity is important for informed consent; however, the consistency in estimates has not been studied. This study aimed to explore the variability and factors affecting risk estimates (REs). METHODS AND MATERIALS: A survey was mailed to Australian radiation oncologists, who were asked to estimate risks of RT complications given 49 clinical scenarios. The REs were assessed for association with oncologist experience, subspecialization, and private practice. RESULTS: The REs were extremely variable, with a 50-fold median variability. The least variability (sevenfold) was for estimates of late, small intestinal perforation/obstruction after a one-third volume received 50 Gy with concurrent 5-fluorouracil (RE range 5-35%). The variation between the smallest and largest REs in 17 scenarios was >or=100-fold. The years of experience was significantly associated with REs of soft/connective-tissue toxicity (p = 0.01) but inversely associated with estimates of neurologic/central nervous system toxicity (p = 0.08). Ninety-six percent of respondents believed REs were important to RT practice; only 24% rated evidence to support their estimates as good. Sixty-seven percent believed national/international groups should pursue the issue further. CONCLUSION: Enormous variability exists in REs for normal tissue complications due to RT that is influenced by the years of experience. Risk estimation is perceived as an important issue without a good evidence base. Additional studies are strongly recommended.
PURPOSE: Estimating the risks of radiotherapy (RT) toxicity is important for informed consent; however, the consistency in estimates has not been studied. This study aimed to explore the variability and factors affecting risk estimates (REs). METHODS AND MATERIALS: A survey was mailed to Australian radiation oncologists, who were asked to estimate risks of RT complications given 49 clinical scenarios. The REs were assessed for association with oncologist experience, subspecialization, and private practice. RESULTS: The REs were extremely variable, with a 50-fold median variability. The least variability (sevenfold) was for estimates of late, small intestinal perforation/obstruction after a one-third volume received 50 Gy with concurrent 5-fluorouracil (RE range 5-35%). The variation between the smallest and largest REs in 17 scenarios was >or=100-fold. The years of experience was significantly associated with REs of soft/connective-tissue toxicity (p = 0.01) but inversely associated with estimates of neurologic/central nervous system toxicity (p = 0.08). Ninety-six percent of respondents believed REs were important to RT practice; only 24% rated evidence to support their estimates as good. Sixty-seven percent believed national/international groups should pursue the issue further. CONCLUSION: Enormous variability exists in REs for normal tissue complications due to RT that is influenced by the years of experience. Risk estimation is perceived as an important issue without a good evidence base. Additional studies are strongly recommended.