Sir,
We read this article with interest. In the methods section the authors mention that the measurements on
the operated hip were compared with those on the contralateral hip. They assumed
that since all the measurements were taken from the same radiograph the
magnification was constant. This assumption would only be correct if a
ratio of the measurements was used and not the actual measured lengths.
The difference between the two offsets (contralateral - operated) is
altered by the magnification of the radiograph: the higher the
magnification, the larger the calculated difference. As a mean of these
figures is used in the formula for the paired t-test, the result can be
significantly different as the t-value is equal to the difference between
the group means divided by the variability of the groups.1 Hence the p values shown in Table II of the article are incorrect.
The overall decrease in offset is actually less than the authors state.
This decrease will be proportional to the magnification of the
radiographs.
Another point of confusion is found in the difference between the
text and the tables in the article. In the methods section the authors state
that the measurements on the operated hip were compared with those on the
contralateral hip, and yet in Table II they claim to be comparing pre-operative
and post-operative measurements.
P.B. Moonot, Clinical Research Fellow,
D.J. Matthews,
M.D. Cronin,
Epsom & St. Helier NHS Trust,
Surrey, UK.
1. http://www.socialresearchmethods.net/kb/stat_t.htm. (Accessed on
08.03.2006)