In Reply: Dr Eliasziw and Ms Paddock-Eliasziw
suggest that a κ statistic rather than a McNemar test for proportions
could have been used to summarize the results of our study. We chose to use
the McNemar test because our primary goal was to show the equivalence of MRI
for detecting hemorrhage compared to CT. Had we used the κ statistic,
a point estimate of 0.84 would still suggest a strong correlation. Perhaps
more important than the choice of statistical approach are the actual data.
Acute hemorrhage was missed in 4 cases with each modality. With MRI, the hemorrhages
were interpreted as chronic rather than acute in 3 patients. With CT, the
hemorrhages were completely missed.
Kidwell CS, Chalela J, Saver JL, Hill MD, Demchuk A, Butman J, Warach S. Comparison of MRI and CT for Detection of Acute Intracerebral Hemorrhage—Reply. JAMA. 2005;293(5):550–551. doi:10.1001/jama.293.5.550-b
Customize your JAMA Network experience by selecting one or more topics from the list below.