We are grateful to Dr Echevarria and colleagues for their interest in and comments on our study. We accept their criticism that the provision of information sheets only on the day of bronchoscopy and the inclusion for all patients of data on complications relevant only to those likely to undergo transbronchial biopsy may have affected our results. However, apart from the constraints imposed by the artificial environment of a randomised trial, we suspect that these practices are not far removed from what happens in real life. While the verbal discussion of the procedure should always put risks in a personal context, it will be difficult—given the workload involved in developing and agreeing even a single information sheet for a procedure—for any individual centre to develop multiple personally tailored information sheets.

We agree that rulings such as Chester v Afshar have changed the legal requirements for doctors so that even very small risks of serious outcomes such as death need to be discussed.1 The authors' survey of consent practices for bronchoscopy among respiratory physicians in the north-east of England confirms for bronchoscopy the same startling variations found in the amount of risk disclosure by doctors for other procedures. Standardisation of risk information, even allowing for individual tailoring, would eliminate much of this variation. However, there is an inherent tension in the fact that informed consent does “double duty” as protection both for doctors and for patients. It is worth noting O'Neill's criticism that, while greater emphasis on patient autonomy in medicine is supposed to make doctors more responsible to patients' needs and wishes, it might have the opposite effect of encouraging a culture of back covering.2 Thus, a legalistic approach to consent may lead to the doctor passing too much information and responsibility onto the patient, “muttering ‘caveat emptor’ under his breath”, as Cowley put it.3