| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Correspondence |
We are grateful to be given the opportunity to reply to Dr Halligan's comments relating to our recent article [1].
The thrust of Dr Halligan's criticism is based on our use of the term reproducibility. Whilst we are grateful for his British Standards reference, our application of the term is in fact statistically correct; the only variable that we truly knew had changed between the CT studies was time. The International Standards Organisation (ISO:5725) [2] definitions of repeatability and reproducibility are as follows.
As the time interval of our experiment may have played an important role in the study, statistical advice suggested that it was probably better, however, to use the term reproducibility than repeatability. Moreover, although the equipment used for the two tests was the same, the operators were not. The ISO definition of reproducibility is expressed in terms of the reproducibility coefficient, which has the interpretation that it is the value below which we would expect 95% of the "absolute differences" to lie. This coefficient is calculated as 2.77 times the standard deviation of the measurement error or, in the case of two measurements per person, 1.96 times the square root of the average of the square differences. In our study, the reproducibility coefficient turned out to be 0.139 ml min-1ml-1 (see Figure 1
). We do not believe that such an error would have unacceptable or catastrophic consequences. Space precluded our being able to explore the analysis further.
|
The implication of Dr Halligan's last statement is that our study has potentially harmed the volunteers, which is unfounded. A complete CT perfusion study has an effective dose of 2.4 mSv, and as such the patients received considerably less than this as part of the study. The therapeutic radiation dose that patients received was of the order of 60 Gy, i.e. over 20 000 times the study dose. We are particularly indebted to the brave volunteers in this study, as most of them were very aware that they were unlikely to be alive when the study was published. We stand by our statement that "we have demonstrated little variability in absolute values for cerebral blood flow measured with CT perfusion imaging in individuals studied approximately 24 h apart".
Yours etc.
Departments of 1Radiology, 2Oncology and 3Neurosurgery, Addenbrooke's Hospital, Cambridge CB2 2QQ, UK
Acknowledgments
The authors acknowledge Brian DM Tom PhD at the Centre for Applied Medical Statistics, Institute of Public Health, University of Cambridge.
Received for publication October 23, 2001. Revision received November 26, 2001. Accepted for publication November 29, 2001.
References
Related articles in BJR:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
