In recent years there has been a major change on the part of funders, particularly in North America, so that data sharing is now considered to be the norm rather than the exception. We believe that data sharing is a good idea. However, we also believe that it is inappropriate to prescribe exactly when or how researchers should preserve and share data, since these issues are highly specific to each study, the nature of the data collected, who is requesting it, and what they intend to do with it. The level of ethical concern will vary according to the nature of the information, and the way in which it is collected - analyses of anonymised hospital admission records may carry a quite different ethical burden than analyses of potentially identifiable health information collected directly from the study participants. It is striking that most discussions about data sharing focus almost exclusively on issues of ownership (by the researchers or the funders) and efficiency (on the part of the funders). There is usually little discussion of the ethical issues involved in data sharing, and its implications for the study participants. Obtaining prior informed consent from the participants does not solve this problem, unless the informed consent process makes it completely clear what is being proposed, in which case most study participants would not agree. Thus, the undoubted benefits of data sharing does not remove the obligations and responsibilities that the original investigators hold for the people they invited to participate in the study.