Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on thebmj.com. Although a selection of rapid responses will be included online and in print as readers' letters, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window. Letters are indexed in PubMed.
We do not ignore the results of statistical analysis - in fact we
very clearly state in the article whether or not results were
statistically significant at the 5% level. We agree with the view, also
held by others, that the interpretation of results of data analysis does
not solely rely on the classification of results into statistically
significant or not according to the arbitrarily chosen but widely accepted
p-value of 0.05 (1). We used the word trend where p-values were between
0.05 and 0.10. We highlighted results which, while not statistically
significant, suggest that clinically important results should not be ruled
out and in turn discussed when these may suggest directions for the
further studies required to provide the definitive results necessary to
guide practice and potentially improve health outcomes. In the context of
our study, where power is limited by the number and size of studies
available, we are of the opinion this is a particularly important
contribution to the interpretation of the study's findings.
1. Rothman KJ, Greenland S. Modern epidemiology. Philadelphia:
Lipncott-Raven; 1998.
Competing interests:
No competing interests
15 February 2011
Tania Winzenberg
Senior Research Fellow
Graeme Jones
Menzies Research Institute Tasmania, University of Tasmania
Winzenberg et al repeatedly use the phrase "There was a trend..." to
describe a result which they have shown to be statistically
nonsignificant. "There was no trend..." would be more accurate. This is not
pedantry: why bother with statistical analysis if the results are to be
ignored?
Editor, I read the recent publication by Winzenberg et al. with a
great interest. I agree with the conclusion that "It is unlikely that
vitamin D supplements are beneficial in children and adolescents with
normal vitamin D levels [1]." Indeed, this is a simple thing that both
"too many" and "too few" are not appropriate for healthy life. Indeed,
there is no report concerning the adverse effeect due to over intake of
vitamin D in pediatric population. Finally, I would like to raise a
question how the appropriate level can be judged for each individual. Have
we to wait until there are overt symptoms?
References
1. Tania Winzenberg, Sandi Powell, Kelly Anne Shaw, and Graeme Jones.
Effects of vitamin D supplementation on bone density in healthy children:
systematic review and meta-analysis. BMJ 2011 342:c7254;
doi:10.1136/bmj.c7254
Response to Statistics, Rapid response on Effects of vitamin D supplementation on bone density in healthy children: systematic review and meta-analysis
We do not ignore the results of statistical analysis - in fact we
very clearly state in the article whether or not results were
statistically significant at the 5% level. We agree with the view, also
held by others, that the interpretation of results of data analysis does
not solely rely on the classification of results into statistically
significant or not according to the arbitrarily chosen but widely accepted
p-value of 0.05 (1). We used the word trend where p-values were between
0.05 and 0.10. We highlighted results which, while not statistically
significant, suggest that clinically important results should not be ruled
out and in turn discussed when these may suggest directions for the
further studies required to provide the definitive results necessary to
guide practice and potentially improve health outcomes. In the context of
our study, where power is limited by the number and size of studies
available, we are of the opinion this is a particularly important
contribution to the interpretation of the study's findings.
1. Rothman KJ, Greenland S. Modern epidemiology. Philadelphia:
Lipncott-Raven; 1998.
Competing interests: No competing interests