Intended for healthcare professionals


Association between competing interests and authors' conclusions: epidemiological study of randomised clinical trials published in the BMJ

BMJ 2002; 325 doi: (Published 03 August 2002) Cite this as: BMJ 2002;325:249

An opportunity for BMJ editors: tackling the "denominator problem"

An opportunity for BMJ editors: tackling the "denominator problem"

In my personal exchange of the material with the authors of this
paper (such as providing them with the equipoise scale, our data
extraction forms, etc), I also made a comment that their results could
have been explained by at least two additional types of bias:

1. pervasive authors’ self-selection bias (the authors tend to send
their best pieces to the high-impact journal such as BMJ)

2. the BMJ editors’ bias

The major problem with this type of research is that its results can
never be completely reliable until the “denominator problem” is addressed.
The journal editors’ decision regarding the aceptance or rejection of a
given paper for publication can potentially seriously skew distribution of
negative and positive studies. I believe that the BMJ editors have a
unique opportunity to inform this important debate by publishing data on
the number of studies that they reject (and accept) for publication
(ideally as a function of the funding source, quality of the studies and
comparator intervention). Only in this way we can learn what really
affects published body of knowledge.

I hope that the BMJ editors who, otherwise, have been very vocal on
the issue of publication bias and conflict of interest will accept this

Benjamin Djulbegovic, MD, PhD

H. Lee Moffitt Cancer & Research Institute,
University of South Florida

Competing interests: No competing interests

13 August 2002
Benjamin Djulbegovic
University of South Florida
H. Lee Moffitt Cancer Center and Research Institute, USF, Tampa, FL 33612, USA