Sir,
I read with interest the article by McQueen et al1 in the January 2008 issue entitled 'Percutaneous screw fixation versus conservative treatment for fractures of the waist of the scaphoid: a prospective randomised study.'
Although this study has considerable merit in demonstrating more
rapid return to function with few complications in the operatively treated
group, I am uncertain how the study methodology allows such apparently
precise calculation of time to union. Clinical and radiographic review
was carried out at 8, 12, 26 and 52 weeks only, and yet table V alludes
to union at 18 and 36 weeks in the two groups. I also question the use of
the mean as an appropriate indicator of average time to union when
assessment was performed on few occasions (albeit for very understandable
and practical reasons), and when the progressively longer intervals
between reviews would be expected heavily to skew the data, especially for
any that were not united by 12 weeks.
I am particularly interested in their findings in relation to nonunion. This was defined based on clinical and radiographic findings at 16
weeks, and yet the patients were not reviewed at that interval so it is unclear how the nonunion rate was determined. I also question the use of
a standard Chi squared test to determine the difference in the rate of nonunion between the two groups when observed frequencies are so small, with
fewer than five occurrences of nonunion in both groups. It is usual
practice to use a continuity correction under such circumstances, or to
use an alternative test, such as Fisher’s Exact Test; these tests yield p
values of 0.350 and 0.353 respectively, considerably higher than the 0.19
quoted, and therefore I question whether the data presented support even a
trend towards a higher rate of nonunion in the conservatively treated
group.
R.A. BROOKS,
Consultant Orthopaedic Surgeon,
Great Western Hospital,
Swindon, UK.
eLetters: ![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
