I appreciated the article by Brattebø et al but have some
concerns about the reporting of results. The authors
allude to changes in mortality and ICU length of stay,
but do not highlight that the changes were not
statistically significant, and in fact do not report p-values.
For example the length of stay in ICU is presented as
"Mean stay decreased by 1.0 day (-0.9 to 2.9 days) from
9.3 days to 8.3 days". The confidence interval crosses
zero and as such is not considered a statistically
significant change. The clearer reporting of the findings
will ensure that readers are not mislead by the findings.
Competing interests:
No competing interests
10 June 2002
George Ntoumenopoulos
senior Clinician Physiotherapist
Physiotherapy Department, The Alfred, Melbourne, Australia
Rapid Response:
Are we confident of the results?
I appreciated the article by Brattebø et al but have some
concerns about the reporting of results. The authors
allude to changes in mortality and ICU length of stay,
but do not highlight that the changes were not
statistically significant, and in fact do not report p-values.
For example the length of stay in ICU is presented as
"Mean stay decreased by 1.0 day (-0.9 to 2.9 days) from
9.3 days to 8.3 days". The confidence interval crosses
zero and as such is not considered a statistically
significant change. The clearer reporting of the findings
will ensure that readers are not mislead by the findings.
Competing interests: No competing interests