Readers will not be surprised at the level of ignorance about radiation doses displayed by clinicians in the paper by Jacob et al., but they may be more concerned to find that radiologists did not fare all that much better, at least in their junior years. The authors’ findings have implications for the way in which radiologists and other clinicians discharge their responsibilities under the Ionising Radiation (Medical Exposures) Regulations [IR(ME)R], but they also have a wider significance in the field of radiation protection. The roles and responsibilities of referrer and practitioner under IR(ME)R are, in theory, quite different and easily separated. The referrer needs to know enough about medicine and his or her patient to give the practitioner the clinical information necessary to carry out the justification process. The practitioner is the repository of knowledge concerning radiation hazards, and will balance these when deciding whether the examination should go ahead. You could therefore argue that, although it would be a good thing for referrers to understand something about the risks associated with ionizing radiation, detailed knowledge is not a necessary prerequisite as far as their IR(ME)R duties are concerned. However, it is not quite as simple as that. The authors make the valid point that the referrer controls the wording of the request and can be selective concerning the clinical details divulged to the practitioner, and we all recognize that one of the less welcome results of implementing imaging guidelines is that referrers learn what to write on request cards in order to get the examination they want. Jacob et al. suggest that this ability to “spin” the clinical information effectively transfers part of the justification process into the hands of the referrer, and that this is a reason to ensure that they know more about the risks of radiation than is currently the case. I accept the logic of this argument, but I do wonder whether referrers who are prepared to be economical with the truth in order to obtain the practitioner’s signature on the request card would allow a short course in radiation protection to influence their behaviour very much, or for very long. Also relevant to any discussion of the justification process is the question of who is to be regarded as a practitioner. The authors include in their definition any clinician who carries out fluoroscopic procedures, and many hospitals take this position. That means that you are committed to giving radiation protection training to a large group of clinical staff on a recurring basis, training that will involve considerably more than the old unlamented half-day “POPUMET course”. Alternatively, the training could be included in the postgraduate courses of the relevant specialities, and this has recently been achieved for cardiologists. Following an approach from the Radiation Protection Committee of the British Institute of Radiology (BIR), the Royal College of Physicians (RCP) agreed to include radiation protection training in the cardiology specialist registrar syllabus, and a number of training courses, jointly validated by the BIR and RCP, have been set up around the country. The reason for targeting cardiologists was the fact that many of them spend a large proportion of their time carrying out complex interventional procedures involving radiation doses that can approach levels where deterministic effects become a possibility. There is thus good reason to educate them in the principles of radiation protection, whether or not we believe they are acting as IR(ME)R practitioners or operators. This type of training, given during the Clinical Radiology (2004) 59, 926–927