for imaging requests were unknown and plausible explanations for high request rates could exist. Serious contemplation of the feedback was likely to benefit, not harm, patients. Furthermore, our study described the use of routinely collected administrative data on volume of imaging requests to generate feedback and study outcomes. We acknowledged that high request rates may not represent inappropriate use in all instances.The team carefully considered the ethical issues of conducting this trial. 4,5 We judged the trial interventions (ie, frequency of feedback [once vs twice] and visual display [standard vs enhanced highlighting of highly requested tests]) as low risk because they involved no more than minimal risk of discomfort or inconvenience compared with standard feedback. Any additional risk of harm to GP participants would be outweighed by the anticipated benefits for patients and for society more broadly from the knowledge gained from the research. The requirement to seek GP consent would make the study unfeasible, and voluntary participation would have limited the generalizability of the findings.The Human Research Ethics Committee at Bond University approved our study and was satisfied that it met requirements for a waiver of consent (clause 2.3.10) of the National Statement on Ethical Conduct in Human Research.