Author's reply

EDITOR,--The aim of our study was to propose a method of reinforcing clinical guidelines. The outcome that was measured showed a significant change in referral patterns and supports the conclusion that carefully designed posters are effectve in reinforcing guidelines. Reductions of 75% for abdominal radiographs and 50% for skull radiographs were achieved. Data on the overall change in referral patterns (that is, for all four areas combined) were not presented as we regarded this statistic as meaningless.

Several other possible reasons for the reduction in referral for radiography were considered, including the experience of the doctors and changes in staff. No other guidelines, including the Royal College of Radiologist's booklet, were used during the period studied. No request was refused because it fell outside the guidelines. Editorial limitations on the length of our paper precluded a full discussion of these issues.

We agree with R D Hardern and D W Hamer that this should not necessarily be regarded as a study of outcome. Clearly, if guidelines are poor and based on dubious evidence then the outcome in patients may well be adversely affected, particularly if the method of reinforcing the guidelines is as successful as in this study. Our guidelines were devised, after an extensive analysis of the literature, in conjunction with local clinicians.

We agree with Mr Dudley that the presence of a vault fracture is associated with an increased risk of intracranial injury. What Dudley does not make clear is that the figure of 1 in 4 refers to adults with both a skull fracture and an altered level of consciousness. In patients with skull fracture who are oriented the incidence of intracranial abnormality is eight times less.1 Furthermore, as many as 60% of patients with intracranial abnormality after head injury will not have a fracture.2 The important finding, the intracranial lesion, may be overlooked if focus is put on skull radiographs. For these reasons, there is now an extensive literature arguing against routine skull radiography after admission in favour of a more open policy of using computed tomography.3 4 The advanced trauma life support guidelines, mentioned by Dudley, also suport this view. Used correctly, this policy can be cost effective,5 but, clearly, local modifications will be necessary where computed tomography and neurosurgical advice are not available.

Patients who cannot be adequately assessed, whether because of unconsciousness or diminished consciousness or after severe head injury, should be assumed to have a cervical spine injury until proved otherwise. This point is adequately covered in our guidelines.

Devising guidelines alone is insufficient; the real challenge is making them work. Dudley expresses concern regarding the increasing trend towards use of clinical guidelines. We trust that Dudley does not regard the challenge of devising and testing methods for ensuring that policies for good practice are implemented as one to be resisted and use this as an excuse for doing nothing.

Senior registrar in diagnostic radiology Consultant radiologist Registrar in accident and emergency Consultant in charge of accident and emergency Department of Radiology, Nuffield Orthopaedic Centre NHS Trust Headington, Oxford OX4 4LD