Send letter to journal:
Re: Author's reply re: Cranial computed tomography in trauma: the accuracy of interpretation by staff in

Dear Editor,

We thank Dr Hynes and colleagues for their interest in our Paper (Emerg Med J 2005;22:538-540). To have selected participants would indeed have introduced a bias. The five permanent members of staff who read the images constituted the only five permanent members of staff at that time, and between them saw all out of hours CT head scans done from A&E for trauma.

We agree that only persons deemed competent should interpret images and that historically in the case of computed tomography (CT) this has been done by radiologists. Our paper seeks to show that in many ways cranial CT is no different from other radiographic images. We would point out that many other images are interpreted by on call clinicians in the acute situation and reported later by radiologists. Many of these are more difficult to interpret, we would cite the chest radiograph as an example.

In our study the images were read by a stable group of experienced medical staff. We would suggest that this may be no worse than rotating trainees in radiology or neurosurgery who are viewing these images in the acute situation on a regular basis, particularly as these staff have undergone varying amounts of training in the interpretation of cranial CT images.

Send letter to journal:
Re: Initial interpretation of head injury CT scans by A&E staff – the way forward?

Dear Editor,

Mucci et al.[1] in their study have re-explored a possibility of scan interpretation by A&E staff, that is worth following up given the rising number of CT scans done for head injuries. The overall agreement and false negative rates demonstrated by the authors would be generally acceptable especially in the light of the fact that nothing that required a transfer to a neurosurgical unit was missed.

This system would work perfectly well if ‘permanent’ A&E staff were physically present in the department at all times to interpret such scans when they are done. However this is not always the case and if staff has to be called in for these purposes one would rather have a radiologist interpreting the scans rather than A&E staff primarily because the quality of neurosurgical response to emergencies depends on the reliability and completeness of the information received from referral hospitals.[2] The present study was retrospective; I suspect a prospective study of a similar nature where the primary decision is taken by the A&E clinician out of hours and the scans later reviewed by a radiologist, would increase the false positives, decrease false negatives and increase both hospital admission rates and the amount of neurosurgical referrals, because A&E staff will be under pressure to be safe in their decision, which is not the case in a retrospective study where patients have been dealt with already. Also the aspect studied here is only trauma and the authors have rightly commented that skills in diagnosing medical conditions which might warrant a scan out of hours need further evaluation before this system can be implemented.[1]

Teleradiology links to a neurosurgical centre is perhaps the best option, but these need to be rather robust and functional at all times to be effective. It results in more work for the neurosurgeons on duty, but unnecessary transfers can be avoided a better quality service can be provided for the whole area.[2] Emergency image transfer system through a mobile telephone might be consideration for the future.[3] What is more cost effective is something that needs some deliberation. I have no doubt that the authors idea is an excellent one and might be the way forward provided all A&E staff medical staff are properly trained.

References:

1. Mucci B, Brett C, Huntley LS, Greene MK. Cranial computed tomography in trauma: the accuracy of interpretation by staff in the emergency department. Emerg Med J 2005;22:538-540.

2. Servadei F, Antonelli V, Mastrilli A, Cultrera F et al. Integration of image transmission into a protocol for head injury management: a preliminary report. Br J Neurosurg 2002 ;16:36-42.

Send letter to journal:
Re: Response to Mucci's study of "ED staff’s interpretation of cranial CTs in trauma"

Dear Editor,

We read with interest Mucci’s[1] study of the accuracy of interpretation by ED staff of cranial CTs in trauma. It is a topical subject that needs exploring, but we have questions with their design.

Firstly, their study was underpowered with only 100 scans examined. The quoted sensitivity of 86.6% has too low a 95% confidence interval (83.4% to 89.9%) to propose trusting the reliability of their ED staff interpretation of the CT scans. If these figures were translated into one scan been examined by one only reader, (which is more comparable to real-life practice) the corrected 95% CI would be 68.7%-94%. Is an error rate of more than 15% really acceptable?

We are concerned that multiple readers interpreted the same cases. They allude to Robinson’s[2] findings about inter-observer variation, but they only studied the variability between radiologists interpreting plain radiographs. Can Robinson’s findings be extrapolated to inter-observer variation of ED staff interpreting CTs?

We are intrigued at the high proportion of ‘abnormal’ scans. Does the fact that skull radiographs are routinely done at their hospital suggest that only more injured patients are scanned, increasing the chance of having serious pathology, hence easier to identify on CT? If they scanned more patients would increasingly subtle abnormalities have been harder to detect?

Finally, we chuckled at the lack of conflicting interests. Would it be churlish to suggest radiologists would welcome any study that would reduce their out of hours workload?

Overall, we welcome Mucci’s paper, but suggest that they have placed too much significance on an underpowered study with considerable inter-observer variability. Further, larger studies are required (and are being performed) to answer this question more thoroughly.

References:

1. Mucci B, Brett C, Huntley LS, Greene MK. Cranial computed tomography in trauma: the accuracy of interpretation by staff in the emergency department. Emerg Med J 2005;22(8):538-40.

2. Robinson PJ, Wilson D, Coral A, Murphy A, Verow P. Variation between experienced observers in the interpretation of accident and emergency radiographs. Br J Radiol 1999;72(856):323-30.

Andrew Hugman, Emergency Medicine SpR

Adrian Boyle, Emergency Medicine Consultant

Send letter to journal:
Re: Cranial computed tomography in trauma: The accuracy of interpretation by staff in the emergency dept

Dear Editor,

We read with interest the article Emerg Med J 2005;22:538-540 by Mucci, Brett, Huntley and Greene. In the methods section it mentions that the CT scans were reviewed retrospectively by five permanent members of the emergency department medical staff. We would be interested to know how these 5 members were chosen. Were there only five members of permanent medical staff in the acute Trust or were there other Consultants Associate Specialists or Staff Grades and why they were excluded from the study. Was it done randomly or was there a method. We suggest that the validity of the conclusion may be biased because individuals with greater knowledge than average might be more enthusiastic to participate in this study. We suggest that to demonstrate any validity all emergency department permanent medical staff in North Cumbria Acute Hospitals NHS Trust should have participated and preferably some other Trust also. We suggest that this study only demonstrates that five selected members of permanent medical staff in West Cumberland Hospital Emergency Department can safely interpret CT scans of the head in trauma patients.

In these days of heightened medico-legal awareness it is not just necessary to be able to do something, one has also to have been trained and updated and to prove one can read CT scans which has been the domain of the Consultant Radiologist. We suspect any mistake could prove very costly and the safe interpretation of a hundred images done in a study would not be considered adequate evidence of the knowledge to read CT scans.

Correspondence to:
Kilian Hynes MRCP FFAEM
Consultant in Accident & Emergency Medicine
Barnet Hospital, Barnet & Chase Farm Hospitals Trust
Barnet, EN5 3DJ,
kilian.hynes{at}bcf.nhs.uk

Brijendra Shravat FRCSEd FFAEM
Lead Clinician and Consultant in Accident & Emergency Medicine
Barnet Hospital, Barnet & Chase Farm Hospitals Trust
Barnet, EN5 3DJ